# fix com_zhipuglm.py illegal temperature problem (#1687)

* Update com_zhipuglm.py

# fix 用户在使用 zhipuai 界面时遇到了关于温度参数的非法参数错误

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -11,6 +11,8 @@ body:
         - Please choose | 请选择
         - Pip Install (I ignored requirements.txt)
         - Pip Install (I used latest requirements.txt)
+        - OneKeyInstall (一键安装脚本-windows)
+        - OneKeyInstall (一键安装脚本-mac)
         - Anaconda (I ignored requirements.txt)
         - Anaconda (I used latest requirements.txt)
         - Docker（Windows/Mac）
@@ -32,7 +34,7 @@ body:
         - Others | 非最新版
     validations:
       required: true
-      
+
   - type: dropdown
     id: os
     attributes:
@@ -45,7 +47,7 @@ body:
         - Docker
     validations:
       required: true
-      
+
   - type: textarea
     id: describe
     attributes:
@@ -53,7 +55,7 @@ body:
       description: Describe the bug | 简述
     validations:
       required: true
-      
+
   - type: textarea
     id: screenshot
     attributes:
@@ -61,15 +63,9 @@ body:
       description: Screen Shot | 有帮助的截图
     validations:
       required: true
-      
+
   - type: textarea
     id: traceback
     attributes:
       label: Terminal Traceback & Material to Help Reproduce Bugs | 终端traceback（如有） + 帮助我们复现的测试材料样本（如有）
       description: Terminal Traceback & Material to Help Reproduce Bugs | 终端traceback（如有） + 帮助我们复现的测试材料样本（如有）
-
-
-
-
-
-

.github/ISSUE_TEMPLATE/feature_request.yml

@@ -21,8 +21,3 @@ body:
     attributes:
       label: Feature Request | 功能请求
       description: Feature Request | 功能请求
-
-
-
-
-

.github/workflows/build-with-all-capacity-beta.yml

@@ -0,0 +1,44 @@
+# https://docs.github.com/en/actions/publishing-packages/publishing-docker-images#publishing-images-to-github-packages
+name: build-with-all-capacity-beta
+
+on:
+  push:
+    branches:
+      - 'master'
+
+env:
+  REGISTRY: ghcr.io
+  IMAGE_NAME: ${{ github.repository }}_with_all_capacity_beta
+
+jobs:
+  build-and-push-image:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      packages: write
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+
+      - name: Log in to the Container registry
+        uses: docker/login-action@v2
+        with:
+          registry: ${{ env.REGISTRY }}
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Extract metadata (tags, labels) for Docker
+        id: meta
+        uses: docker/metadata-action@v4
+        with:
+          images: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}
+
+      - name: Build and push Docker image
+        uses: docker/build-push-action@v4
+        with:
+          context: .
+          push: true
+          file: docs/GithubAction+AllCapacityBeta
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}

.github/workflows/build-with-all-capacity.yml

@@ -0,0 +1,44 @@
+# https://docs.github.com/en/actions/publishing-packages/publishing-docker-images#publishing-images-to-github-packages
+name: build-with-all-capacity
+
+on:
+  push:
+    branches:
+      - 'master'
+
+env:
+  REGISTRY: ghcr.io
+  IMAGE_NAME: ${{ github.repository }}_with_all_capacity
+
+jobs:
+  build-and-push-image:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      packages: write
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+
+      - name: Log in to the Container registry
+        uses: docker/login-action@v2
+        with:
+          registry: ${{ env.REGISTRY }}
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Extract metadata (tags, labels) for Docker
+        id: meta
+        uses: docker/metadata-action@v4
+        with:
+          images: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}
+
+      - name: Build and push Docker image
+        uses: docker/build-push-action@v4
+        with:
+          context: .
+          push: true
+          file: docs/GithubAction+AllCapacity
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}

.github/workflows/build-with-audio-assistant.yml

@@ -0,0 +1,44 @@
+# https://docs.github.com/en/actions/publishing-packages/publishing-docker-images#publishing-images-to-github-packages
+name: build-with-audio-assistant
+
+on:
+  push:
+    branches:
+      - 'master'
+
+env:
+  REGISTRY: ghcr.io
+  IMAGE_NAME: ${{ github.repository }}_audio_assistant
+
+jobs:
+  build-and-push-image:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      packages: write
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+
+      - name: Log in to the Container registry
+        uses: docker/login-action@v2
+        with:
+          registry: ${{ env.REGISTRY }}
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Extract metadata (tags, labels) for Docker
+        id: meta
+        uses: docker/metadata-action@v4
+        with:
+          images: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}
+
+      - name: Build and push Docker image
+        uses: docker/build-push-action@v4
+        with:
+          context: .
+          push: true
+          file: docs/GithubAction+NoLocal+AudioAssistant
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}

.github/workflows/build-with-chatglm.yml

@@ -1,5 +1,5 @@
 # https://docs.github.com/en/actions/publishing-packages/publishing-docker-images#publishing-images-to-github-packages
-name: Create and publish a Docker image for ChatGLM support
+name: build-with-chatglm
 
 on:
   push:

.github/workflows/build-with-jittorllms.yml

@@ -1,5 +1,5 @@
 # https://docs.github.com/en/actions/publishing-packages/publishing-docker-images#publishing-images-to-github-packages
-name: Create and publish a Docker image for ChatGLM support
+name: build-with-jittorllms
 
 on:
   push:

.github/workflows/build-with-latex.yml

@@ -1,5 +1,5 @@
 # https://docs.github.com/en/actions/publishing-packages/publishing-docker-images#publishing-images-to-github-packages
-name: Create and publish a Docker image for Latex support
+name: build-with-latex
 
 on:
   push:

.github/workflows/build-without-local-llms.yml

@@ -1,5 +1,5 @@
 # https://docs.github.com/en/actions/publishing-packages/publishing-docker-images#publishing-images-to-github-packages
-name: Create and publish a Docker image
+name: build-without-local-llms
 
 on:
   push:

.github/workflows/stale.yml

@@ -0,0 +1,25 @@
+# This workflow warns and then closes issues and PRs that have had no activity for a specified amount of time.
+#
+# You can adjust the behavior by modifying this file.
+# For more information, see:
+# https://github.com/actions/stale
+
+name: 'Close stale issues and PRs'
+on:
+  schedule:
+    - cron: '*/5 * * * *'
+
+jobs:
+  stale:
+    runs-on: ubuntu-latest
+    permissions:
+      issues: write
+      pull-requests: read
+
+    steps:
+      - uses: actions/stale@v8
+        with:
+          stale-issue-message: 'This issue is stale because it has been open 100 days with no activity. Remove stale label or comment or this will be closed in 1 days.'
+          days-before-stale: 100
+          days-before-close: 1
+          debug-only: true

.gitignore

@@ -146,7 +146,10 @@ debug*
 private*
 crazy_functions/test_project/pdf_and_word
 crazy_functions/test_samples
-request_llm/jittorllms
+request_llms/jittorllms
 multi-language
-request_llm/moss
+request_llms/moss
 media
+flagged
+request_llms/ChatGLM-6b-onnx-u8s8
+.pre-commit-config.yaml

.pre-commit-config.yaml

@@ -1,32 +0,0 @@
-default_language_version:
-  python: python3
-exclude: 'dotnet'
-ci:
-  autofix_prs: true
-  autoupdate_commit_msg: '[pre-commit.ci] pre-commit suggestions'
-  autoupdate_schedule: 'quarterly'
-
-repos:
-  - repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v4.4.0
-    hooks:
-    - id: check-ast
-    # - id: check-yaml
-    - id: check-toml
-    - id: check-json
-    - id: check-byte-order-marker
-      exclude: .gitignore
-    - id: check-merge-conflict
-    - id: detect-private-key
-    - id: trailing-whitespace
-    - id: end-of-file-fixer
-    - id: no-commit-to-branch
-  - repo: https://github.com/psf/black
-    rev: 23.3.0
-    hooks:
-    - id: black
-  # - repo: https://github.com/charliermarsh/ruff-pre-commit
-  #   rev: v0.0.261
-  #   hooks:
-  #     - id: ruff
-  #       args: ["--fix"]

README.md

@@ -1,20 +1,7 @@
----
-title: GPT-Academic
-emoji: 😻
-colorFrom: blue
-colorTo: blue
-sdk: gradio
-sdk_version: 3.32.0
-app_file: app.py
-pinned: false
----
-
-# ChatGPT 学术优化
-> **Note**
-> 
-> 2023.11.12: 某些依赖包尚不兼容python 3.12，推荐python 3.11。
-> 
-> 2023.12.26: 安装依赖时，请选择`requirements.txt`中**指定的版本**。 安装命令：`pip install -r requirements.txt`。本项目完全开源免费，您可通过订阅[在线服务](https://github.com/binary-husky/gpt_academic/wiki/online)的方式鼓励本项目的发展。
+> [!IMPORTANT]
+> 2024.3.11: 恭迎Claude3和Moonshot，全力支持Qwen、GLM、DeepseekCoder等中文大语言模型！  
+> 2024.1.18: 更新3.70版本，支持Mermaid绘图库（让大模型绘制脑图）  
+> 2024.1.17: 安装依赖时，请选择`requirements.txt`中**指定的版本**。 安装命令：`pip install -r requirements.txt`。本项目完全开源免费，您可通过订阅[在线服务](https://github.com/binary-husky/gpt_academic/wiki/online)的方式鼓励本项目的发展。
 
 <br>
 
@@ -265,8 +252,7 @@ P.S. 如果需要依赖Latex的插件功能，请见Wiki。另外，您也可以
 # Advanced Usage
 ### I：自定义新的便捷按钮（学术快捷键）
 
-任意文本编辑器打开`core_functional.py`，添加如下条目，然后重启程序。（如果按钮已存在，那么可以直接修改（前缀、后缀都已支持热修改），无需重启程序即可生效。）
-例如
+现在已可以通过UI中的`界面外观`菜单中的`自定义菜单`添加新的便捷按钮。如果需要在代码中定义，请使用任意文本编辑器打开`core_functional.py`，添加如下条目即可：
 
 ```python
 "超级英译中": {

check_proxy.py

@@ -47,7 +47,7 @@ def backup_and_download(current_version, remote_version):
     shutil.copytree('./', backup_dir, ignore=lambda x, y: ['history'])
     proxies = get_conf('proxies')
     try:    r = requests.get('https://github.com/binary-husky/chatgpt_academic/archive/refs/heads/master.zip', proxies=proxies, stream=True)
-    except: r = requests.get('https://public.gpt-academic.top/publish/master.zip', proxies=proxies, stream=True)
+    except: r = requests.get('https://public.agent-matrix.com/publish/master.zip', proxies=proxies, stream=True)
     zip_file_path = backup_dir+'/master.zip'
     with open(zip_file_path, 'wb+') as f:
         f.write(r.content)
@@ -81,7 +81,7 @@ def patch_and_restart(path):
     dir_util.copy_tree(path_new_version, './')
     print亮绿('代码已经更新，即将更新pip包依赖……')
     for i in reversed(range(5)): time.sleep(1); print(i)
-    try: 
+    try:
         import subprocess
         subprocess.check_call([sys.executable, '-m', 'pip', 'install', '-r', 'requirements.txt'])
     except:
@@ -113,7 +113,7 @@ def auto_update(raise_error=False):
         import json
         proxies = get_conf('proxies')
         try:    response = requests.get("https://raw.githubusercontent.com/binary-husky/chatgpt_academic/master/version", proxies=proxies, timeout=5)
-        except: response = requests.get("https://public.gpt-academic.top/publish/version", proxies=proxies, timeout=5)
+        except: response = requests.get("https://public.agent-matrix.com/publish/version", proxies=proxies, timeout=5)
         remote_json_data = json.loads(response.text)
         remote_version = remote_json_data['version']
         if remote_json_data["show_feature"]:
@@ -159,7 +159,7 @@ def warm_up_modules():
         enc.encode("模块预热", disallowed_special=())
         enc = model_info["gpt-4"]['tokenizer']
         enc.encode("模块预热", disallowed_special=())
-        
+
 def warm_up_vectordb():
     print('正在执行一些模块的预热 ...')
     from toolbox import ProxyNetworkActivate
@@ -167,7 +167,7 @@ def warm_up_vectordb():
         import nltk
         with ProxyNetworkActivate("Warmup_Modules"): nltk.download("punkt")
 
-        
+
 if __name__ == '__main__':
     import os
     os.environ['no_proxy'] = '*'  # 避免代理网络产生意外污染

colorful.py

@@ -3,7 +3,7 @@ from sys import stdout
 
 if platform.system()=="Linux":
     pass
-else: 
+else:
     from colorama import init
     init()
 

config.py

@@ -11,10 +11,6 @@
 API_KEY = "此处填API密钥"    # 可同时填写多个API-KEY，用英文逗号分割，例如API_KEY = "sk-openaikey1,sk-openaikey2,fkxxxx-api2dkey3,azure-apikey4"
 
 
-# [step 1]>> API_KEY = "sk-123456789xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx123456789"。极少数情况下，还需要填写组织（格式如org-123456789abcdefghijklmno的），请向下翻，找 API_ORG 设置项
-API_KEY = "此处填API密钥"    # 可同时填写多个API-KEY，用英文逗号分割，例如API_KEY = "sk-openaikey1,sk-openaikey2,fkxxxx-api2dkey3,azure-apikey4"
-
-
 # [step 2]>> 改为True应用代理，如果直接在海外服务器部署，此处不修改；如果使用本地或无地域限制的大模型时，此处也不需要修改
 USE_PROXY = False
 if USE_PROXY:
@@ -34,7 +30,33 @@ if USE_PROXY:
 else:
     proxies = None
 
-# ------------------------------------ 以下配置可以优化体验, 但大部分场合下并不需要修改 ------------------------------------
+# [step 3]>> 模型选择是 (注意: LLM_MODEL是默认选中的模型, 它*必须*被包含在AVAIL_LLM_MODELS列表中 )
+LLM_MODEL = "gpt-3.5-turbo-16k" # 可选 ↓↓↓
+AVAIL_LLM_MODELS = ["gpt-4-1106-preview", "gpt-4-turbo-preview", "gpt-4-vision-preview",
+                    "gpt-3.5-turbo-1106", "gpt-3.5-turbo-16k", "gpt-3.5-turbo", "azure-gpt-3.5",
+                    "gpt-4", "gpt-4-32k", "azure-gpt-4", "glm-4", "glm-3-turbo",
+                    "gemini-pro", "chatglm3"
+                    ]
+# --- --- --- ---
+# P.S. 其他可用的模型还包括
+# AVAIL_LLM_MODELS = [
+#   "qianfan", "deepseekcoder",
+#   "spark", "sparkv2", "sparkv3", "sparkv3.5",
+#   "qwen-turbo", "qwen-plus", "qwen-max", "qwen-local",
+#   "moonshot-v1-128k", "moonshot-v1-32k", "moonshot-v1-8k",
+#   "gpt-3.5-turbo-0613", "gpt-3.5-turbo-16k-0613", "gpt-3.5-turbo-0125"
+#   "claude-3-haiku-20240307","claude-3-sonnet-20240229","claude-3-opus-20240229", "claude-2.1", "claude-instant-1.2",
+#   "moss", "llama2", "chatglm_onnx", "internlm", "jittorllms_pangualpha", "jittorllms_llama",
+#   "yi-34b-chat-0205", "yi-34b-chat-200k"
+# ]
+# --- --- --- ---
+# 此外，为了更灵活地接入one-api多模型管理界面，您还可以在接入one-api时，
+# 使用"one-api-*"前缀直接使用非标准方式接入的模型，例如
+# AVAIL_LLM_MODELS = ["one-api-claude-3-sonnet-20240229(max_token=100000)"]
+# --- --- --- ---
+
+
+# --------------- 以下配置可以优化体验 ---------------
 
 # 重新URL重新定向，实现更换API_URL的作用（高危设置! 常规情况下不要修改! 通过修改此设置，您将把您的API-KEY和对话隐私完全暴露给您设定的中间人！）
 # 格式: API_URL_REDIRECT = {"https://api.openai.com/v1/chat/completions": "在这里填写重定向的api.openai.com的URL"}
@@ -49,7 +71,7 @@ DEFAULT_WORKER_NUM = 3
 
 # 色彩主题, 可选 ["Default", "Chuanhu-Small-and-Beautiful", "High-Contrast"]
 # 更多主题, 请查阅Gradio主题商店: https://huggingface.co/spaces/gradio/theme-gallery 可选 ["Gstaff/Xkcd", "NoCrypt/Miku", ...]
-THEME = "Chuanhu-Small-and-Beautiful"
+THEME = "Default"
 AVAIL_THEMES = ["Default", "Chuanhu-Small-and-Beautiful", "High-Contrast", "Gstaff/Xkcd", "NoCrypt/Miku"]
 
 
@@ -70,7 +92,7 @@ LAYOUT = "LEFT-RIGHT"   # "LEFT-RIGHT"（左右布局） # "TOP-DOWN"（上下
 
 
 # 暗色模式 / 亮色模式
-DARK_MODE = False
+DARK_MODE = True
 
 
 # 发送请求到OpenAI后，等待多久判定为超时
@@ -84,28 +106,11 @@ WEB_PORT = -1
 # 如果OpenAI不响应（网络卡顿、代理失败、KEY失效），重试的次数限制
 MAX_RETRY = 2
 
-# OpenAI模型选择是（gpt4现在只对申请成功的人开放）
-LLM_MODEL = "gpt-3.5-turbo" # 可选 "chatglm"
-AVAIL_LLM_MODELS = ["gpt-3.5-turbo", "gpt-4", "api2d-gpt-4", "api2d-gpt-3.5-turbo", "spark", "azure-gpt-3.5"]
 
 # 插件分类默认选项
 DEFAULT_FN_GROUPS = ['对话', '编程', '学术', '智能体']
 
 
-# 模型选择是 (注意: LLM_MODEL是默认选中的模型, 它*必须*被包含在AVAIL_LLM_MODELS列表中 )
-LLM_MODEL = "gpt-3.5-turbo-16k" # 可选 ↓↓↓
-AVAIL_LLM_MODELS = ["gpt-4-1106-preview", "gpt-4-turbo-preview", "gpt-4-vision-preview",
-                    "gpt-3.5-turbo-1106", "gpt-3.5-turbo-16k", "gpt-3.5-turbo", "azure-gpt-3.5",
-                    "gpt-4", "gpt-4-32k", "azure-gpt-4", "glm-4", "glm-3-turbo",
-                    "gemini-pro", "chatglm3", "claude-2"]
-# P.S. 其他可用的模型还包括 [
-# "moss", "qwen-turbo", "qwen-plus", "qwen-max"
-# "zhipuai", "qianfan", "deepseekcoder", "llama2", "qwen-local", "gpt-3.5-turbo-0613",
-# "gpt-3.5-turbo-16k-0613",  "gpt-3.5-random", "api2d-gpt-3.5-turbo", 'api2d-gpt-3.5-turbo-16k',
-# "spark", "sparkv2", "sparkv3", "chatglm_onnx", "claude-1-100k", "claude-2", "internlm", "jittorllms_pangualpha", "jittorllms_llama"
-# ]
-
-
 # 定义界面上“询问多个GPT模型”插件应该使用哪些模型，请从AVAIL_LLM_MODELS中选择，并在不同模型之间用`&`间隔，例如"gpt-3.5-turbo&chatglm3&azure-gpt-4"
 MULTI_QUERY_LLM_MODELS = "gpt-3.5-turbo&chatglm3"
 
@@ -134,6 +139,7 @@ CHATGLM_PTUNING_CHECKPOINT = "" # 例如"/home/hmp/ChatGLM2-6B/ptuning/output/6b
 LOCAL_MODEL_DEVICE = "cpu" # 可选 "cuda"
 LOCAL_MODEL_QUANT = "FP16" # 默认 "FP16" "INT4" 启用量化INT4版本 "INT8" 启用量化INT8版本
 
+
 # 设置gradio的并行线程数（不需要修改）
 CONCURRENT_COUNT = 100
 
@@ -143,7 +149,7 @@ AUTO_CLEAR_TXT = False
 
 
 # 加一个live2d装饰
-ADD_WAIFU = True
+ADD_WAIFU = False
 
 
 # 设置用户名和密码（不需要修改）（相关功能不稳定，与gradio版本和网络都相关，如果本地使用不建议加这个）
@@ -151,7 +157,8 @@ ADD_WAIFU = True
 AUTHENTICATION = []
 
 
-# 如果需要在二级路径下运行（常规情况下，不要修改!!）（需要配合修改main.py才能生效!）
+# 如果需要在二级路径下运行（常规情况下，不要修改!!）
+# （举例 CUSTOM_PATH = "/gpt_academic"，可以让软件运行在 http://ip:port/gpt_academic/ 下。）
 CUSTOM_PATH = "/"
 
 
@@ -179,14 +186,8 @@ AZURE_ENGINE = "填入你亲手写的部署名"            # 读 docs\use_azure.
 AZURE_CFG_ARRAY = {}
 
 
-# 使用Newbing (不推荐使用，未来将删除)
-NEWBING_STYLE = "creative"  # ["creative", "balanced", "precise"]
-NEWBING_COOKIES = """
-put your new bing cookies here
-"""
-
-
-# 阿里云实时语音识别 配置难度较高 仅建议高手用户使用 参考 https://github.com/binary-husky/gpt_academic/blob/master/docs/use_audio.md
+# 阿里云实时语音识别 配置难度较高
+# 参考 https://github.com/binary-husky/gpt_academic/blob/master/docs/use_audio.md
 ENABLE_AUDIO = False
 ALIYUN_TOKEN=""     # 例如 f37f30e0f9934c34a992f6f64f7eba4f
 ALIYUN_APPKEY=""    # 例如 RoPlZrM88DnAFkZK
@@ -205,16 +206,18 @@ ZHIPUAI_API_KEY = ""
 ZHIPUAI_MODEL = "" # 此选项已废弃，不再需要填写
 
 
-# # 火山引擎YUNQUE大模型
-# YUNQUE_SECRET_KEY = ""
-# YUNQUE_ACCESS_KEY = ""
-# YUNQUE_MODEL = ""
-
-
 # Claude API KEY
 ANTHROPIC_API_KEY = ""
 
 
+# 月之暗面 API KEY
+MOONSHOT_API_KEY = ""
+
+
+# 零一万物(Yi Model) API KEY
+YIMODEL_API_KEY = ""
+
+
 # Mathpix 拥有执行PDF的OCR功能，但是需要注册账号
 MATHPIX_APPID = ""
 MATHPIX_APPKEY = ""
@@ -229,7 +232,7 @@ GEMINI_API_KEY = ''
 
 
 # HUGGINGFACE的TOKEN，下载LLAMA时起作用 https://huggingface.co/docs/hub/security-tokens
-HUGGINGFACE_ACCESS_TOKEN = ""
+HUGGINGFACE_ACCESS_TOKEN = "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"
 
 
 # GROBID服务器地址（填写多个可以均衡负载），用于高质量地读取PDF文档
@@ -273,7 +276,11 @@ PLUGIN_HOT_RELOAD = False
 # 自定义按钮的最大数量限制
 NUM_CUSTOM_BASIC_BTN = 4
 
+
+
 """
+--------------- 配置关联关系说明 ---------------
+
 在线大模型配置关联关系示意图
 │
 ├── "gpt-3.5-turbo" 等openai模型
@@ -297,7 +304,7 @@ NUM_CUSTOM_BASIC_BTN = 4
 │   ├── XFYUN_API_SECRET
 │   └── XFYUN_API_KEY
 │
-├── "claude-1-100k" 等claude模型
+├── "claude-3-opus-20240229" 等claude模型
 │   └── ANTHROPIC_API_KEY
 │
 ├── "stack-claude"
@@ -312,15 +319,19 @@ NUM_CUSTOM_BASIC_BTN = 4
 ├── "glm-4", "glm-3-turbo", "zhipuai" 智谱AI大模型
 │   └── ZHIPUAI_API_KEY
 │
+├── "yi-34b-chat-0205", "yi-34b-chat-200k" 等零一万物(Yi Model)大模型
+│   └── YIMODEL_API_KEY
+│
 ├── "qwen-turbo" 等通义千问大模型
 │   └──  DASHSCOPE_API_KEY
 │
 ├── "Gemini"
 │   └──  GEMINI_API_KEY
 │
-└── "newbing" Newbing接口不再稳定，不推荐使用
-    ├── NEWBING_STYLE
-    └── NEWBING_COOKIES
+└── "one-api-...(max_token=...)" 用一种更方便的方式接入one-api多模型管理界面
+    ├── AVAIL_LLM_MODELS
+    ├── API_KEY
+    └── API_URL_REDIRECT
 
 
 本地大模型示意图

core_functional.py

@@ -34,16 +34,16 @@ def get_core_functions():
             # [6] 文本预处理 （可选参数，默认 None，举例：写个函数移除所有的换行符）
             "PreProcess": None,
         },
-        
-        
+
+
         "总结绘制脑图": {
             # 前缀，会被加在你的输入之前。例如，用来描述你的要求，例如翻译、解释代码、润色等等
-            "Prefix":   r"",
+            "Prefix":   '''"""\n\n''',
             # 后缀，会被加在你的输入之后。例如，配合前缀可以把你的输入内容用引号圈起来
             "Suffix":
                 # dedent() 函数用于去除多行字符串的缩进
-                dedent("\n"+r'''
-                    ==============================
+                dedent("\n\n"+r'''
+                    """
 
                     使用mermaid flowchart对以上文本进行总结，概括上述段落的内容以及内在逻辑关系，例如：
 
@@ -57,15 +57,15 @@ def get_core_functions():
                         C --> |"箭头名2"| F["节点名6"]
                     ```
 
-                    警告：
+                    注意：
                     （1）使用中文
                     （2）节点名字使用引号包裹，如["Laptop"]
                     （3）`|` 和 `"`之间不要存在空格
                     （4）根据情况选择flowchart LR（从左到右）或者flowchart TD（从上到下）
                 '''),
         },
-        
-        
+
+
         "查找语法错误": {
             "Prefix":   r"Help me ensure that the grammar and the spelling is correct. "
                         r"Do not try to polish the text, if no mistake is found, tell me that this paragraph is good. "
@@ -85,14 +85,14 @@ def get_core_functions():
             "Suffix":   r"",
             "PreProcess": clear_line_break,    # 预处理：清除换行符
         },
-        
-        
+
+
         "中译英": {
             "Prefix":   r"Please translate following sentence to English:" + "\n\n",
             "Suffix":   r"",
         },
-        
-        
+
+
         "学术英中互译": {
             "Prefix":   build_gpt_academic_masked_string_langbased(
                             text_show_chinese=
@@ -112,29 +112,29 @@ def get_core_functions():
                         ) + "\n\n",
             "Suffix":   r"",
         },
-        
-        
+
+
         "英译中": {
             "Prefix":   r"翻译成地道的中文：" + "\n\n",
             "Suffix":   r"",
             "Visible":  False,
         },
-        
-        
+
+
         "找图片": {
             "Prefix":   r"我需要你找一张网络图片。使用Unsplash API(https://source.unsplash.com/960x640/?<英语关键词>)获取图片URL，"
                         r"然后请使用Markdown格式封装，并且不要有反斜线，不要用代码块。现在，请按以下描述给我发送图片：" + "\n\n",
             "Suffix":   r"",
             "Visible":  False,
         },
-        
-        
+
+
         "解释代码": {
             "Prefix":   r"请解释以下代码：" + "\n```\n",
             "Suffix":   "\n```\n",
         },
-        
-        
+
+
         "参考文献转Bib": {
             "Prefix":   r"Here are some bibliography items, please transform them into bibtex style."
                         r"Note that, reference styles maybe more than one kind, you should transform each item correctly."

crazy_functions/CodeInterpreter.py

@@ -1,232 +0,0 @@
-from collections.abc import Callable, Iterable, Mapping
-from typing import Any
-from toolbox import CatchException, update_ui, gen_time_str, trimmed_format_exc
-from toolbox import promote_file_to_downloadzone, get_log_folder
-from .crazy_utils import request_gpt_model_in_new_thread_with_ui_alive
-from .crazy_utils import input_clipping, try_install_deps
-from multiprocessing import Process, Pipe
-import os
-import time
-
-templete = """
-```python
-import ...  # Put dependencies here, e.g. import numpy as np
-
-class TerminalFunction(object): # Do not change the name of the class, The name of the class must be `TerminalFunction`
-
-    def run(self, path):    # The name of the function must be `run`, it takes only a positional argument.
-        # rewrite the function you have just written here 
-        ...
-        return generated_file_path
-```
-"""
-
-def inspect_dependency(chatbot, history):
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-    return True
-
-def get_code_block(reply):
-    import re
-    pattern = r"```([\s\S]*?)```" # regex pattern to match code blocks
-    matches = re.findall(pattern, reply) # find all code blocks in text
-    if len(matches) == 1: 
-        return matches[0].strip('python') #  code block
-    for match in matches:
-        if 'class TerminalFunction' in match:
-            return match.strip('python') #  code block
-    raise RuntimeError("GPT is not generating proper code.")
-
-def gpt_interact_multi_step(txt, file_type, llm_kwargs, chatbot, history):
-    # 输入
-    prompt_compose = [
-        f'Your job:\n'
-        f'1. write a single Python function, which takes a path of a `{file_type}` file as the only argument and returns a `string` containing the result of analysis or the path of generated files. \n',
-        f"2. You should write this function to perform following task: " + txt + "\n",
-        f"3. Wrap the output python function with markdown codeblock."
-    ]
-    i_say = "".join(prompt_compose)
-    demo = []
-
-    # 第一步
-    gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-        inputs=i_say, inputs_show_user=i_say, 
-        llm_kwargs=llm_kwargs, chatbot=chatbot, history=demo, 
-        sys_prompt= r"You are a programmer."
-    )
-    history.extend([i_say, gpt_say])
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 界面更新
-
-    # 第二步
-    prompt_compose = [
-        "If previous stage is successful, rewrite the function you have just written to satisfy following templete: \n",
-        templete
-    ]
-    i_say = "".join(prompt_compose); inputs_show_user = "If previous stage is successful, rewrite the function you have just written to satisfy executable templete. "
-    gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-        inputs=i_say, inputs_show_user=inputs_show_user, 
-        llm_kwargs=llm_kwargs, chatbot=chatbot, history=history, 
-        sys_prompt= r"You are a programmer."
-    )
-    code_to_return = gpt_say
-    history.extend([i_say, gpt_say])
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 界面更新
-    
-    # # 第三步
-    # i_say = "Please list to packages to install to run the code above. Then show me how to use `try_install_deps` function to install them."
-    # i_say += 'For instance. `try_install_deps(["opencv-python", "scipy", "numpy"])`'
-    # installation_advance = yield from request_gpt_model_in_new_thread_with_ui_alive(
-    #     inputs=i_say, inputs_show_user=inputs_show_user, 
-    #     llm_kwargs=llm_kwargs, chatbot=chatbot, history=history, 
-    #     sys_prompt= r"You are a programmer."
-    # )
-    # # # 第三步  
-    # i_say = "Show me how to use `pip` to install packages to run the code above. "
-    # i_say += 'For instance. `pip install -r opencv-python scipy numpy`'
-    # installation_advance = yield from request_gpt_model_in_new_thread_with_ui_alive(
-    #     inputs=i_say, inputs_show_user=i_say, 
-    #     llm_kwargs=llm_kwargs, chatbot=chatbot, history=history, 
-    #     sys_prompt= r"You are a programmer."
-    # )
-    installation_advance = ""
-    
-    return code_to_return, installation_advance, txt, file_type, llm_kwargs, chatbot, history
-
-def make_module(code):
-    module_file = 'gpt_fn_' + gen_time_str().replace('-','_')
-    with open(f'{get_log_folder()}/{module_file}.py', 'w', encoding='utf8') as f:
-        f.write(code)
-
-    def get_class_name(class_string):
-        import re
-        # Use regex to extract the class name
-        class_name = re.search(r'class (\w+)\(', class_string).group(1)
-        return class_name
-
-    class_name = get_class_name(code)
-    return f"{get_log_folder().replace('/', '.')}.{module_file}->{class_name}"
-
-def init_module_instance(module):
-    import importlib
-    module_, class_ = module.split('->')
-    init_f = getattr(importlib.import_module(module_), class_)
-    return init_f()
-
-def for_immediate_show_off_when_possible(file_type, fp, chatbot):
-    if file_type in ['png', 'jpg']:
-        image_path = os.path.abspath(fp)
-        chatbot.append(['这是一张图片, 展示如下:',  
-            f'本地文件地址: <br/>`{image_path}`<br/>'+
-            f'本地文件预览: <br/><div align="center"><img src="file={image_path}"></div>'
-        ])
-    return chatbot
-
-def subprocess_worker(instance, file_path, return_dict):
-    return_dict['result'] = instance.run(file_path)
-
-def have_any_recent_upload_files(chatbot):
-    _5min = 5 * 60
-    if not chatbot: return False    # chatbot is None
-    most_recent_uploaded = chatbot._cookies.get("most_recent_uploaded", None)
-    if not most_recent_uploaded: return False   # most_recent_uploaded is None
-    if time.time() - most_recent_uploaded["time"] < _5min: return True # most_recent_uploaded is new
-    else: return False  # most_recent_uploaded is too old
-
-def get_recent_file_prompt_support(chatbot):
-    most_recent_uploaded = chatbot._cookies.get("most_recent_uploaded", None)
-    path = most_recent_uploaded['path']
-    return path
-
-@CatchException
-def 虚空终端CodeInterpreter(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-    """
-    txt             输入栏用户输入的文本，例如需要翻译的一段话，再例如一个包含了待处理文件的路径
-    llm_kwargs      gpt模型参数，如温度和top_p等，一般原样传递下去就行
-    plugin_kwargs   插件模型的参数，暂时没有用武之地
-    chatbot         聊天显示框的句柄，用于显示给用户
-    history         聊天历史，前情提要
-    system_prompt   给gpt的静默提醒
-    web_port        当前软件运行的端口号
-    """
-    raise NotImplementedError
-
-    # 清空历史，以免输入溢出
-    history = []; clear_file_downloadzone(chatbot)
-
-    # 基本信息：功能、贡献者
-    chatbot.append([
-        "函数插件功能？",
-        "CodeInterpreter开源版, 此插件处于开发阶段, 建议暂时不要使用, 插件初始化中 ..."
-    ])
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-
-    if have_any_recent_upload_files(chatbot):
-        file_path = get_recent_file_prompt_support(chatbot)
-    else:
-        chatbot.append(["文件检索", "没有发现任何近期上传的文件。"])
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-
-    # 读取文件
-    if ("recently_uploaded_files" in plugin_kwargs) and (plugin_kwargs["recently_uploaded_files"] == ""): plugin_kwargs.pop("recently_uploaded_files")
-    recently_uploaded_files = plugin_kwargs.get("recently_uploaded_files", None)
-    file_path = recently_uploaded_files[-1]
-    file_type = file_path.split('.')[-1]
-
-    # 粗心检查
-    if is_the_upload_folder(txt):
-        chatbot.append([
-            "...",
-            f"请在输入框内填写需求，然后再次点击该插件（文件路径 {file_path} 已经被记忆）"
-        ])
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        return
-    
-    # 开始干正事
-    for j in range(5):  # 最多重试5次
-        try:
-            code, installation_advance, txt, file_type, llm_kwargs, chatbot, history = \
-                yield from gpt_interact_multi_step(txt, file_type, llm_kwargs, chatbot, history)
-            code = get_code_block(code)
-            res = make_module(code)
-            instance = init_module_instance(res)
-            break
-        except Exception as e:
-            chatbot.append([f"第{j}次代码生成尝试，失败了", f"错误追踪\n```\n{trimmed_format_exc()}\n```\n"])
-            yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-
-    # 代码生成结束, 开始执行
-    try:
-        import multiprocessing
-        manager = multiprocessing.Manager()
-        return_dict = manager.dict()
-
-        p = multiprocessing.Process(target=subprocess_worker, args=(instance, file_path, return_dict))
-        # only has 10 seconds to run
-        p.start(); p.join(timeout=10)
-        if p.is_alive(): p.terminate(); p.join()
-        p.close()
-        res = return_dict['result']
-        # res = instance.run(file_path)
-    except Exception as e:
-        chatbot.append(["执行失败了", f"错误追踪\n```\n{trimmed_format_exc()}\n```\n"])
-        # chatbot.append(["如果是缺乏依赖，请参考以下建议", installation_advance])
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        return
-
-    # 顺利完成，收尾
-    res = str(res)
-    if os.path.exists(res):
-        chatbot.append(["执行成功了，结果是一个有效文件", "结果：" + res])
-        new_file_path = promote_file_to_downloadzone(res, chatbot=chatbot)
-        chatbot = for_immediate_show_off_when_possible(file_type, new_file_path, chatbot)
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 界面更新
-    else:
-        chatbot.append(["执行成功了，结果是一个字符串", "结果：" + res])
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 界面更新   
-
-"""
-测试：
-    裁剪图像，保留下半部分
-    交换图像的蓝色通道和红色通道
-    将图像转为灰度图像
-    将csv文件转excel表格
-"""

crazy_functions/Langchain知识库.py

@@ -1,106 +0,0 @@
-from toolbox import CatchException, update_ui, ProxyNetworkActivate, update_ui_lastest_msg
-from .crazy_utils import request_gpt_model_in_new_thread_with_ui_alive, get_files_from_everything
-
-
-
-@CatchException
-def 知识库问答(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-    """
-    txt             输入栏用户输入的文本，例如需要翻译的一段话，再例如一个包含了待处理文件的路径
-    llm_kwargs      gpt模型参数, 如温度和top_p等, 一般原样传递下去就行
-    plugin_kwargs   插件模型的参数，暂时没有用武之地
-    chatbot         聊天显示框的句柄，用于显示给用户
-    history         聊天历史，前情提要
-    system_prompt   给gpt的静默提醒
-    web_port        当前软件运行的端口号
-    """
-    history = []    # 清空历史，以免输入溢出
-
-    # < --------------------读取参数--------------- >
-    if ("advanced_arg" in plugin_kwargs) and (plugin_kwargs["advanced_arg"] == ""): plugin_kwargs.pop("advanced_arg")
-    kai_id = plugin_kwargs.get("advanced_arg", 'default')
-
-    chatbot.append((f"向`{kai_id}`知识库中添加文件。", "[Local Message] 从一批文件(txt, md, tex)中读取数据构建知识库, 然后进行问答。"))
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-
-    # resolve deps
-    try:
-        from zh_langchain import construct_vector_store
-        from langchain.embeddings.huggingface import HuggingFaceEmbeddings
-        from .crazy_utils import knowledge_archive_interface
-    except Exception as e:
-        chatbot.append(["依赖不足", "导入依赖失败。正在尝试自动安装，请查看终端的输出或耐心等待..."])
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        from .crazy_utils import try_install_deps
-        try_install_deps(['zh_langchain==0.2.1', 'pypinyin'], reload_m=['pypinyin', 'zh_langchain'])
-        yield from update_ui_lastest_msg("安装完成，您可以再次重试。", chatbot, history)
-        return
-
-    # < --------------------读取文件--------------- >
-    file_manifest = []
-    spl = ["txt", "doc", "docx", "email", "epub", "html", "json", "md", "msg", "pdf", "ppt", "pptx", "rtf"]
-    for sp in spl:
-        _, file_manifest_tmp, _ = get_files_from_everything(txt, type=f'.{sp}')
-        file_manifest += file_manifest_tmp
-    
-    if len(file_manifest) == 0:
-        chatbot.append(["没有找到任何可读取文件", "当前支持的格式包括: txt, md, docx, pptx, pdf, json等"])
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        return
-    
-    # < -------------------预热文本向量化模组--------------- >
-    chatbot.append(['<br/>'.join(file_manifest), "正在预热文本向量化模组, 如果是第一次运行, 将消耗较长时间下载中文向量化模型..."])
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-    print('Checking Text2vec ...')
-    from langchain.embeddings.huggingface import HuggingFaceEmbeddings
-    with ProxyNetworkActivate('Download_LLM'):    # 临时地激活代理网络
-        HuggingFaceEmbeddings(model_name="GanymedeNil/text2vec-large-chinese")
-
-    # < -------------------构建知识库--------------- >
-    chatbot.append(['<br/>'.join(file_manifest), "正在构建知识库..."])
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-    print('Establishing knowledge archive ...')
-    with ProxyNetworkActivate('Download_LLM'):    # 临时地激活代理网络
-        kai = knowledge_archive_interface()
-        kai.feed_archive(file_manifest=file_manifest, id=kai_id)
-    kai_files = kai.get_loaded_file()
-    kai_files = '<br/>'.join(kai_files)
-    # chatbot.append(['知识库构建成功', "正在将知识库存储至cookie中"])
-    # yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-    # chatbot._cookies['langchain_plugin_embedding'] = kai.get_current_archive_id()
-    # chatbot._cookies['lock_plugin'] = 'crazy_functions.Langchain知识库->读取知识库作答'
-    # chatbot.append(['完成', "“根据知识库作答”函数插件已经接管问答系统, 提问吧! 但注意, 您接下来不能再使用其他插件了，刷新页面即可以退出知识库问答模式。"])
-    chatbot.append(['构建完成', f"当前知识库内的有效文件：\n\n---\n\n{kai_files}\n\n---\n\n请切换至“知识库问答”插件进行知识库访问, 或者使用此插件继续上传更多文件。"])
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 由于请求gpt需要一段时间，我们先及时地做一次界面更新
-
-@CatchException
-def 读取知识库作答(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port=-1):
-    # resolve deps
-    try:
-        from zh_langchain import construct_vector_store
-        from langchain.embeddings.huggingface import HuggingFaceEmbeddings
-        from .crazy_utils import knowledge_archive_interface
-    except Exception as e:
-        chatbot.append(["依赖不足", "导入依赖失败。正在尝试自动安装，请查看终端的输出或耐心等待..."])
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        from .crazy_utils import try_install_deps
-        try_install_deps(['zh_langchain==0.2.1', 'pypinyin'], reload_m=['pypinyin', 'zh_langchain'])
-        yield from update_ui_lastest_msg("安装完成，您可以再次重试。", chatbot, history)
-        return
-
-    # < -------------------  --------------- >
-    kai = knowledge_archive_interface()
-
-    if ("advanced_arg" in plugin_kwargs) and (plugin_kwargs["advanced_arg"] == ""): plugin_kwargs.pop("advanced_arg")
-    kai_id = plugin_kwargs.get("advanced_arg", 'default')
-    resp, prompt = kai.answer_with_archive_by_id(txt, kai_id)
-
-    chatbot.append((txt, f'[知识库 {kai_id}] ' + prompt))
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 由于请求gpt需要一段时间，我们先及时地做一次界面更新
-    gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-        inputs=prompt, inputs_show_user=txt, 
-        llm_kwargs=llm_kwargs, chatbot=chatbot, history=[], 
-        sys_prompt=system_prompt
-    )
-    history.extend((prompt, gpt_say))
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 由于请求gpt需要一段时间，我们先及时地做一次界面更新

crazy_functions/Latex全文润色.py

@@ -46,7 +46,7 @@ class PaperFileGroup():
                 manifest.append(path + '.polish.tex')
                 f.write(res)
         return manifest
-    
+
     def zip_result(self):
         import os, time
         folder = os.path.dirname(self.file_paths[0])
@@ -59,7 +59,7 @@ def 多文件润色(file_manifest, project_folder, llm_kwargs, plugin_kwargs, ch
     from .crazy_utils import request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency
 
 
-    #  <-------- 读取Latex文件，删除其中的所有注释 ----------> 
+    #  <-------- 读取Latex文件，删除其中的所有注释 ---------->
     pfg = PaperFileGroup()
 
     for index, fp in enumerate(file_manifest):
@@ -73,31 +73,31 @@ def 多文件润色(file_manifest, project_folder, llm_kwargs, plugin_kwargs, ch
             pfg.file_paths.append(fp)
             pfg.file_contents.append(clean_tex_content)
 
-    #  <-------- 拆分过长的latex文件 ----------> 
+    #  <-------- 拆分过长的latex文件 ---------->
     pfg.run_file_split(max_token_limit=1024)
     n_split = len(pfg.sp_file_contents)
 
 
-    #  <-------- 多线程润色开始 ----------> 
+    #  <-------- 多线程润色开始 ---------->
     if language == 'en':
         if mode == 'polish':
-            inputs_array = ["Below is a section from an academic paper, polish this section to meet the academic standard, " + 
-                            "improve the grammar, clarity and overall readability, do not modify any latex command such as \section, \cite and equations:" + 
+            inputs_array = [r"Below is a section from an academic paper, polish this section to meet the academic standard, " +
+                            r"improve the grammar, clarity and overall readability, do not modify any latex command such as \section, \cite and equations:" +
                             f"\n\n{frag}" for frag in pfg.sp_file_contents]
         else:
-            inputs_array = [r"Below is a section from an academic paper, proofread this section." + 
-                            r"Do not modify any latex command such as \section, \cite, \begin, \item and equations. " + 
-                            r"Answer me only with the revised text:" + 
+            inputs_array = [r"Below is a section from an academic paper, proofread this section." +
+                            r"Do not modify any latex command such as \section, \cite, \begin, \item and equations. " +
+                            r"Answer me only with the revised text:" +
                         f"\n\n{frag}" for frag in pfg.sp_file_contents]
         inputs_show_user_array = [f"Polish {f}" for f in pfg.sp_file_tag]
         sys_prompt_array = ["You are a professional academic paper writer." for _ in range(n_split)]
     elif language == 'zh':
         if mode == 'polish':
-            inputs_array = [f"以下是一篇学术论文中的一段内容，请将此部分润色以满足学术标准，提高语法、清晰度和整体可读性，不要修改任何LaTeX命令，例如\section，\cite和方程式：" + 
+            inputs_array = [r"以下是一篇学术论文中的一段内容，请将此部分润色以满足学术标准，提高语法、清晰度和整体可读性，不要修改任何LaTeX命令，例如\section，\cite和方程式：" +
                             f"\n\n{frag}" for frag in pfg.sp_file_contents]
         else:
-            inputs_array = [f"以下是一篇学术论文中的一段内容，请对这部分内容进行语法矫正。不要修改任何LaTeX命令，例如\section，\cite和方程式：" + 
-                            f"\n\n{frag}" for frag in pfg.sp_file_contents] 
+            inputs_array = [r"以下是一篇学术论文中的一段内容，请对这部分内容进行语法矫正。不要修改任何LaTeX命令，例如\section，\cite和方程式：" +
+                            f"\n\n{frag}" for frag in pfg.sp_file_contents]
         inputs_show_user_array = [f"润色 {f}" for f in pfg.sp_file_tag]
         sys_prompt_array=["你是一位专业的中文学术论文作家。" for _ in range(n_split)]
 
@@ -113,7 +113,7 @@ def 多文件润色(file_manifest, project_folder, llm_kwargs, plugin_kwargs, ch
         scroller_max_len = 80
     )
 
-    #  <-------- 文本碎片重组为完整的tex文件，整理结果为压缩包 ----------> 
+    #  <-------- 文本碎片重组为完整的tex文件，整理结果为压缩包 ---------->
     try:
         pfg.sp_file_result = []
         for i_say, gpt_say in zip(gpt_response_collection[0::2], gpt_response_collection[1::2]):
@@ -124,7 +124,7 @@ def 多文件润色(file_manifest, project_folder, llm_kwargs, plugin_kwargs, ch
     except:
         print(trimmed_format_exc())
 
-    #  <-------- 整理结果，退出 ----------> 
+    #  <-------- 整理结果，退出 ---------->
     create_report_file_name = time.strftime("%Y-%m-%d-%H-%M-%S", time.localtime()) + f"-chatgpt.polish.md"
     res = write_history_to_file(gpt_response_collection, file_basename=create_report_file_name)
     promote_file_to_downloadzone(res, chatbot=chatbot)

crazy_functions/Latex全文翻译.py

@@ -39,7 +39,7 @@ def 多文件翻译(file_manifest, project_folder, llm_kwargs, plugin_kwargs, ch
     import time, os, re
     from .crazy_utils import request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency
 
-    #  <-------- 读取Latex文件，删除其中的所有注释 ----------> 
+    #  <-------- 读取Latex文件，删除其中的所有注释 ---------->
     pfg = PaperFileGroup()
 
     for index, fp in enumerate(file_manifest):
@@ -53,11 +53,11 @@ def 多文件翻译(file_manifest, project_folder, llm_kwargs, plugin_kwargs, ch
             pfg.file_paths.append(fp)
             pfg.file_contents.append(clean_tex_content)
 
-    #  <-------- 拆分过长的latex文件 ----------> 
+    #  <-------- 拆分过长的latex文件 ---------->
     pfg.run_file_split(max_token_limit=1024)
     n_split = len(pfg.sp_file_contents)
 
-    #  <-------- 抽取摘要 ----------> 
+    #  <-------- 抽取摘要 ---------->
     # if language == 'en':
     #     abs_extract_inputs = f"Please write an abstract for this paper"
 
@@ -70,14 +70,14 @@ def 多文件翻译(file_manifest, project_folder, llm_kwargs, plugin_kwargs, ch
     #     sys_prompt="Your job is to collect information from materials。",
     # )
 
-    #  <-------- 多线程润色开始 ----------> 
+    #  <-------- 多线程润色开始 ---------->
     if language == 'en->zh':
-        inputs_array = ["Below is a section from an English academic paper, translate it into Chinese, do not modify any latex command such as \section, \cite and equations:" + 
+        inputs_array = ["Below is a section from an English academic paper, translate it into Chinese, do not modify any latex command such as \section, \cite and equations:" +
                         f"\n\n{frag}" for frag in pfg.sp_file_contents]
         inputs_show_user_array = [f"翻译 {f}" for f in pfg.sp_file_tag]
         sys_prompt_array = ["You are a professional academic paper translator." for _ in range(n_split)]
     elif language == 'zh->en':
-        inputs_array = [f"Below is a section from a Chinese academic paper, translate it into English, do not modify any latex command such as \section, \cite and equations:" + 
+        inputs_array = [f"Below is a section from a Chinese academic paper, translate it into English, do not modify any latex command such as \section, \cite and equations:" +
                         f"\n\n{frag}" for frag in pfg.sp_file_contents]
         inputs_show_user_array = [f"翻译 {f}" for f in pfg.sp_file_tag]
         sys_prompt_array = ["You are a professional academic paper translator." for _ in range(n_split)]
@@ -93,7 +93,7 @@ def 多文件翻译(file_manifest, project_folder, llm_kwargs, plugin_kwargs, ch
         scroller_max_len = 80
     )
 
-    #  <-------- 整理结果，退出 ----------> 
+    #  <-------- 整理结果，退出 ---------->
     create_report_file_name = time.strftime("%Y-%m-%d-%H-%M-%S", time.localtime()) + f"-chatgpt.polish.md"
     res = write_history_to_file(gpt_response_collection, create_report_file_name)
     promote_file_to_downloadzone(res, chatbot=chatbot)

crazy_functions/Latex输出PDF.py

@@ -1,4 +1,4 @@
-from toolbox import update_ui, trimmed_format_exc, get_conf, get_log_folder, promote_file_to_downloadzone
+from toolbox import update_ui, trimmed_format_exc, get_conf, get_log_folder, promote_file_to_downloadzone, check_repeat_upload, map_file_to_sha256
 from toolbox import CatchException, report_exception, update_ui_lastest_msg, zip_result, gen_time_str
 from functools import partial
 import glob, os, requests, time, json, tarfile
@@ -40,7 +40,7 @@ def switch_prompt(pfg, mode, more_requirement):
 
 
 def desend_to_extracted_folder_if_exist(project_folder):
-    """ 
+    """
     Descend into the extracted folder if it exists, otherwise return the original folder.
 
     Args:
@@ -56,7 +56,7 @@ def desend_to_extracted_folder_if_exist(project_folder):
 
 
 def move_project(project_folder, arxiv_id=None):
-    """ 
+    """
     Create a new work folder and copy the project folder to it.
 
     Args:
@@ -112,9 +112,9 @@ def arxiv_download(chatbot, history, txt, allow_cache=True):
     if ('.' in txt) and ('/' not in txt) and is_float(txt[:10]):  # is arxiv ID
         txt = 'https://arxiv.org/abs/' + txt[:10]
 
-    if not txt.startswith('https://arxiv.org'): 
+    if not txt.startswith('https://arxiv.org'):
         return txt, None    # 是本地文件，跳过下载
-    
+
     # <-------------- inspect format ------------->
     chatbot.append([f"检测到arxiv文档连接", '尝试下载 ...'])
     yield from update_ui(chatbot=chatbot, history=history)
@@ -214,7 +214,7 @@ def pdf2tex_project(pdf_file_path):
         return None
 
 
-# =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= 插件主程序1 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=    
+# =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= 插件主程序1 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 
 
 @CatchException
@@ -291,7 +291,7 @@ def Latex英文纠错加PDF对比(txt, llm_kwargs, plugin_kwargs, chatbot, histo
     return success
 
 
-# =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= 插件主程序2 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=    
+# =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= 插件主程序2 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 
 @CatchException
 def Latex翻译中文并重新编译PDF(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, user_request):
@@ -326,7 +326,7 @@ def Latex翻译中文并重新编译PDF(txt, llm_kwargs, plugin_kwargs, chatbot,
         txt, arxiv_id = yield from arxiv_download(chatbot, history, txt, allow_cache)
     except tarfile.ReadError as e:
         yield from update_ui_lastest_msg(
-            "无法自动下载该论文的Latex源码，请前往arxiv打开此论文下载页面，点other Formats，然后download source手动下载latex源码包。接下来调用本地Latex翻译插件即可。", 
+            "无法自动下载该论文的Latex源码，请前往arxiv打开此论文下载页面，点other Formats，然后download source手动下载latex源码包。接下来调用本地Latex翻译插件即可。",
             chatbot=chatbot, history=history)
         return
 
@@ -385,7 +385,7 @@ def Latex翻译中文并重新编译PDF(txt, llm_kwargs, plugin_kwargs, chatbot,
     return success
 
 
-#  =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- 插件主程序3  =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=   
+#  =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- 插件主程序3  =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 
 @CatchException
 def PDF翻译中文并重新编译PDF(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
@@ -438,47 +438,101 @@ def PDF翻译中文并重新编译PDF(txt, llm_kwargs, plugin_kwargs, chatbot, h
         yield from update_ui(chatbot=chatbot, history=history)  # 刷新界面
         return
 
-    # <-------------- convert pdf into tex ------------->
-    project_folder = pdf2tex_project(file_manifest[0])
+    hash_tag = map_file_to_sha256(file_manifest[0])
 
-    # Translate English Latex to Chinese Latex, and compile it
-    file_manifest = [f for f in glob.glob(f'{project_folder}/**/*.tex', recursive=True)]
-    if len(file_manifest) == 0:
-        report_exception(chatbot, history, a=f"解析项目: {txt}", b=f"找不到任何.tex文件: {txt}")
-        yield from update_ui(chatbot=chatbot, history=history)  # 刷新界面
-        return
+    # <-------------- check repeated pdf ------------->
+    chatbot.append([f"检查PDF是否被重复上传", "正在检查..."])
+    yield from update_ui(chatbot=chatbot, history=history)
+    repeat, project_folder = check_repeat_upload(file_manifest[0], hash_tag)
 
-    # <-------------- if is a zip/tar file ------------->
-    project_folder = desend_to_extracted_folder_if_exist(project_folder)
+    except_flag = False
 
-    # <-------------- move latex project away from temp folder ------------->
-    project_folder = move_project(project_folder)
+    if repeat:
+        yield from update_ui_lastest_msg(f"发现重复上传，请查收结果（压缩包）...", chatbot=chatbot, history=history)
 
-    # <-------------- if merge_translate_zh is already generated, skip gpt req ------------->
-    if not os.path.exists(project_folder + '/merge_translate_zh.tex'):
-        yield from Latex精细分解与转化(file_manifest, project_folder, llm_kwargs, plugin_kwargs,
-                                       chatbot, history, system_prompt, mode='translate_zh',
-                                       switch_prompt=_switch_prompt_)
+        try:
+            trans_html_file = [f for f in glob.glob(f'{project_folder}/**/*.trans.html', recursive=True)][0]
+            promote_file_to_downloadzone(trans_html_file, rename_file=None, chatbot=chatbot)
 
-    # <-------------- compile PDF ------------->
-    success = yield from 编译Latex(chatbot, history, main_file_original='merge',
-                                   main_file_modified='merge_translate_zh', mode='translate_zh',
-                                   work_folder_original=project_folder, work_folder_modified=project_folder,
-                                   work_folder=project_folder)
+            translate_pdf = [f for f in glob.glob(f'{project_folder}/**/merge_translate_zh.pdf', recursive=True)][0]
+            promote_file_to_downloadzone(translate_pdf, rename_file=None, chatbot=chatbot)
 
-    # <-------------- zip PDF ------------->
-    zip_res = zip_result(project_folder)
-    if success:
-        chatbot.append((f"成功啦", '请查收结果（压缩包）...'))
-        yield from update_ui(chatbot=chatbot, history=history);
-        time.sleep(1)  # 刷新界面
-        promote_file_to_downloadzone(file=zip_res, chatbot=chatbot)
-    else:
-        chatbot.append((f"失败了",
-                        '虽然PDF生成失败了, 但请查收结果（压缩包）, 内含已经翻译的Tex文档, 您可以到Github Issue区, 用该压缩包进行反馈。如系统是Linux，请检查系统字体（见Github wiki） ...'))
-        yield from update_ui(chatbot=chatbot, history=history);
-        time.sleep(1)  # 刷新界面
-        promote_file_to_downloadzone(file=zip_res, chatbot=chatbot)
+            comparison_pdf = [f for f in glob.glob(f'{project_folder}/**/comparison.pdf', recursive=True)][0]
+            promote_file_to_downloadzone(comparison_pdf, rename_file=None, chatbot=chatbot)        
 
-    # <-------------- we are done ------------->
-    return success
+            zip_res = zip_result(project_folder)
+            promote_file_to_downloadzone(file=zip_res, chatbot=chatbot)
+
+            return True
+        
+        except:
+            report_exception(chatbot, history, b=f"发现重复上传，但是无法找到相关文件")
+            yield from update_ui(chatbot=chatbot, history=history)
+            
+            chatbot.append([f"没有相关文件", '尝试重新翻译PDF...'])
+            yield from update_ui(chatbot=chatbot, history=history)
+
+            except_flag = True
+            
+    
+    elif not repeat or except_flag:
+        yield from update_ui_lastest_msg(f"未发现重复上传", chatbot=chatbot, history=history)
+
+        # <-------------- convert pdf into tex ------------->
+        chatbot.append([f"解析项目: {txt}", "正在将PDF转换为tex项目，请耐心等待..."])
+        yield from update_ui(chatbot=chatbot, history=history)
+        project_folder = pdf2tex_project(file_manifest[0])
+        if project_folder is None:
+            report_exception(chatbot, history, a=f"解析项目: {txt}", b=f"PDF转换为tex项目失败")
+            yield from update_ui(chatbot=chatbot, history=history)
+            return False
+
+        # <-------------- translate latex file into Chinese ------------->
+        yield from update_ui_lastest_msg("正在tex项目将翻译为中文...", chatbot=chatbot, history=history)
+        file_manifest = [f for f in glob.glob(f'{project_folder}/**/*.tex', recursive=True)]
+        if len(file_manifest) == 0:
+            report_exception(chatbot, history, a=f"解析项目: {txt}", b=f"找不到任何.tex文件: {txt}")
+            yield from update_ui(chatbot=chatbot, history=history)  # 刷新界面
+            return
+
+        # <-------------- if is a zip/tar file ------------->
+        project_folder = desend_to_extracted_folder_if_exist(project_folder)
+
+        # <-------------- move latex project away from temp folder ------------->
+        project_folder = move_project(project_folder)
+
+        # <-------------- set a hash tag for repeat-checking ------------->
+        with open(pj(project_folder, hash_tag + '.tag'), 'w') as f:
+            f.write(hash_tag)
+            f.close()
+
+
+        # <-------------- if merge_translate_zh is already generated, skip gpt req ------------->
+        if not os.path.exists(project_folder + '/merge_translate_zh.tex'):
+            yield from Latex精细分解与转化(file_manifest, project_folder, llm_kwargs, plugin_kwargs,
+                                        chatbot, history, system_prompt, mode='translate_zh',
+                                        switch_prompt=_switch_prompt_)
+
+        # <-------------- compile PDF ------------->
+        yield from update_ui_lastest_msg("正在将翻译好的项目tex项目编译为PDF...", chatbot=chatbot, history=history)
+        success = yield from 编译Latex(chatbot, history, main_file_original='merge',
+                                    main_file_modified='merge_translate_zh', mode='translate_zh',
+                                    work_folder_original=project_folder, work_folder_modified=project_folder,
+                                    work_folder=project_folder)
+
+        # <-------------- zip PDF ------------->
+        zip_res = zip_result(project_folder)
+        if success:
+            chatbot.append((f"成功啦", '请查收结果（压缩包）...'))
+            yield from update_ui(chatbot=chatbot, history=history);
+            time.sleep(1)  # 刷新界面
+            promote_file_to_downloadzone(file=zip_res, chatbot=chatbot)
+        else:
+            chatbot.append((f"失败了",
+                            '虽然PDF生成失败了, 但请查收结果（压缩包）, 内含已经翻译的Tex文档, 您可以到Github Issue区, 用该压缩包进行反馈。如系统是Linux，请检查系统字体（见Github wiki） ...'))
+            yield from update_ui(chatbot=chatbot, history=history);
+            time.sleep(1)  # 刷新界面
+            promote_file_to_downloadzone(file=zip_res, chatbot=chatbot)
+
+        # <-------------- we are done ------------->
+        return success

crazy_functions/Latex输出PDF结果.py

@@ -1,306 +0,0 @@
-from toolbox import update_ui, trimmed_format_exc, get_conf, get_log_folder, promote_file_to_downloadzone
-from toolbox import CatchException, report_exception, update_ui_lastest_msg, zip_result, gen_time_str
-from functools import partial
-import glob, os, requests, time
-pj = os.path.join
-ARXIV_CACHE_DIR = os.path.expanduser(f"~/arxiv_cache/")
-
-# =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- 工具函数 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
-# 专业词汇声明  = 'If the term "agent" is used in this section, it should be translated to "智能体". '
-def switch_prompt(pfg, mode, more_requirement):
-    """
-    Generate prompts and system prompts based on the mode for proofreading or translating.
-    Args:
-    - pfg: Proofreader or Translator instance.
-    - mode: A string specifying the mode, either 'proofread' or 'translate_zh'.
-
-    Returns:
-    - inputs_array: A list of strings containing prompts for users to respond to.
-    - sys_prompt_array: A list of strings containing prompts for system prompts.
-    """
-    n_split = len(pfg.sp_file_contents)
-    if mode == 'proofread_en':
-        inputs_array = [r"Below is a section from an academic paper, proofread this section." + 
-                        r"Do not modify any latex command such as \section, \cite, \begin, \item and equations. " + more_requirement +
-                        r"Answer me only with the revised text:" + 
-                        f"\n\n{frag}" for frag in pfg.sp_file_contents]
-        sys_prompt_array = ["You are a professional academic paper writer." for _ in range(n_split)]
-    elif mode == 'translate_zh':
-        inputs_array = [r"Below is a section from an English academic paper, translate it into Chinese. " + more_requirement + 
-                        r"Do not modify any latex command such as \section, \cite, \begin, \item and equations. " + 
-                        r"Answer me only with the translated text:" + 
-                        f"\n\n{frag}" for frag in pfg.sp_file_contents]
-        sys_prompt_array = ["You are a professional translator." for _ in range(n_split)]
-    else:
-        assert False, "未知指令"
-    return inputs_array, sys_prompt_array
-
-def desend_to_extracted_folder_if_exist(project_folder):
-    """ 
-    Descend into the extracted folder if it exists, otherwise return the original folder.
-
-    Args:
-    - project_folder: A string specifying the folder path.
-
-    Returns:
-    - A string specifying the path to the extracted folder, or the original folder if there is no extracted folder.
-    """
-    maybe_dir = [f for f in glob.glob(f'{project_folder}/*') if os.path.isdir(f)]
-    if len(maybe_dir) == 0: return project_folder
-    if maybe_dir[0].endswith('.extract'): return maybe_dir[0]
-    return project_folder
-
-def move_project(project_folder, arxiv_id=None):
-    """ 
-    Create a new work folder and copy the project folder to it.
-
-    Args:
-    - project_folder: A string specifying the folder path of the project.
-
-    Returns:
-    - A string specifying the path to the new work folder.
-    """
-    import shutil, time
-    time.sleep(2)   # avoid time string conflict
-    if arxiv_id is not None:
-        new_workfolder = pj(ARXIV_CACHE_DIR, arxiv_id, 'workfolder')
-    else:
-        new_workfolder = f'{get_log_folder()}/{gen_time_str()}'
-    try:
-        shutil.rmtree(new_workfolder)
-    except:
-        pass
-
-    # align subfolder if there is a folder wrapper
-    items = glob.glob(pj(project_folder,'*'))
-    items = [item for item in items if os.path.basename(item)!='__MACOSX']
-    if len(glob.glob(pj(project_folder,'*.tex'))) == 0 and len(items) == 1:
-        if os.path.isdir(items[0]): project_folder = items[0]
-
-    shutil.copytree(src=project_folder, dst=new_workfolder)
-    return new_workfolder
-
-def arxiv_download(chatbot, history, txt, allow_cache=True):
-    def check_cached_translation_pdf(arxiv_id):
-        translation_dir = pj(ARXIV_CACHE_DIR, arxiv_id, 'translation')
-        if not os.path.exists(translation_dir):
-            os.makedirs(translation_dir)
-        target_file = pj(translation_dir, 'translate_zh.pdf')
-        if os.path.exists(target_file):
-            promote_file_to_downloadzone(target_file, rename_file=None, chatbot=chatbot)
-            target_file_compare = pj(translation_dir, 'comparison.pdf')
-            if os.path.exists(target_file_compare):
-                promote_file_to_downloadzone(target_file_compare, rename_file=None, chatbot=chatbot)
-            return target_file
-        return False
-    def is_float(s):
-        try:
-            float(s)
-            return True
-        except ValueError:
-            return False
-    if ('.' in txt) and ('/' not in txt) and is_float(txt): # is arxiv ID
-        txt = 'https://arxiv.org/abs/' + txt.strip()
-    if ('.' in txt) and ('/' not in txt) and is_float(txt[:10]): # is arxiv ID
-        txt = 'https://arxiv.org/abs/' + txt[:10]
-    if not txt.startswith('https://arxiv.org'): 
-        return txt, None
-    
-    # <-------------- inspect format ------------->
-    chatbot.append([f"检测到arxiv文档连接", '尝试下载 ...']) 
-    yield from update_ui(chatbot=chatbot, history=history)
-    time.sleep(1) # 刷新界面
-
-    url_ = txt   # https://arxiv.org/abs/1707.06690
-    if not txt.startswith('https://arxiv.org/abs/'): 
-        msg = f"解析arxiv网址失败, 期望格式例如: https://arxiv.org/abs/1707.06690。实际得到格式: {url_}。"
-        yield from update_ui_lastest_msg(msg, chatbot=chatbot, history=history) # 刷新界面
-        return msg, None
-    # <-------------- set format ------------->
-    arxiv_id = url_.split('/abs/')[-1]
-    if 'v' in arxiv_id: arxiv_id = arxiv_id[:10]
-    cached_translation_pdf = check_cached_translation_pdf(arxiv_id)
-    if cached_translation_pdf and allow_cache: return cached_translation_pdf, arxiv_id
-
-    url_tar = url_.replace('/abs/', '/e-print/')
-    translation_dir = pj(ARXIV_CACHE_DIR, arxiv_id, 'e-print')
-    extract_dst = pj(ARXIV_CACHE_DIR, arxiv_id, 'extract')
-    os.makedirs(translation_dir, exist_ok=True)
-    
-    # <-------------- download arxiv source file ------------->
-    dst = pj(translation_dir, arxiv_id+'.tar')
-    if os.path.exists(dst):
-        yield from update_ui_lastest_msg("调用缓存", chatbot=chatbot, history=history)  # 刷新界面
-    else:
-        yield from update_ui_lastest_msg("开始下载", chatbot=chatbot, history=history)  # 刷新界面
-        proxies = get_conf('proxies')
-        r = requests.get(url_tar, proxies=proxies)
-        with open(dst, 'wb+') as f:
-            f.write(r.content)
-    # <-------------- extract file ------------->
-    yield from update_ui_lastest_msg("下载完成", chatbot=chatbot, history=history)  # 刷新界面
-    from toolbox import extract_archive
-    extract_archive(file_path=dst, dest_dir=extract_dst)
-    return extract_dst, arxiv_id
-# =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= 插件主程序1 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=    
-
-
-@CatchException
-def Latex英文纠错加PDF对比(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-    # <-------------- information about this plugin ------------->
-    chatbot.append([ "函数插件功能？",
-        "对整个Latex项目进行纠错, 用latex编译为PDF对修正处做高亮。函数插件贡献者: Binary-Husky。注意事项: 目前仅支持GPT3.5/GPT4，其他模型转化效果未知。目前对机器学习类文献转化效果最好，其他类型文献转化效果未知。仅在Windows系统进行了测试，其他操作系统表现未知。"])
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-    
-    # <-------------- more requirements ------------->
-    if ("advanced_arg" in plugin_kwargs) and (plugin_kwargs["advanced_arg"] == ""): plugin_kwargs.pop("advanced_arg")
-    more_req = plugin_kwargs.get("advanced_arg", "")
-    _switch_prompt_ = partial(switch_prompt, more_requirement=more_req)
-
-    # <-------------- check deps ------------->
-    try:
-        import glob, os, time, subprocess
-        subprocess.Popen(['pdflatex', '-version'])
-        from .latex_fns.latex_actions import Latex精细分解与转化, 编译Latex
-    except Exception as e:
-        chatbot.append([ f"解析项目: {txt}",
-            f"尝试执行Latex指令失败。Latex没有安装, 或者不在环境变量PATH中。安装方法https://tug.org/texlive/。报错信息\n\n```\n\n{trimmed_format_exc()}\n\n```\n\n"])
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        return
-    
-
-    # <-------------- clear history and read input ------------->
-    history = []
-    if os.path.exists(txt):
-        project_folder = txt
-    else:
-        if txt == "": txt = '空空如也的输入栏'
-        report_exception(chatbot, history, a = f"解析项目: {txt}", b = f"找不到本地项目或无权访问: {txt}")
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        return
-    file_manifest = [f for f in glob.glob(f'{project_folder}/**/*.tex', recursive=True)]
-    if len(file_manifest) == 0:
-        report_exception(chatbot, history, a = f"解析项目: {txt}", b = f"找不到任何.tex文件: {txt}")
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        return
-    
-
-    # <-------------- if is a zip/tar file ------------->
-    project_folder = desend_to_extracted_folder_if_exist(project_folder)
-
-
-    # <-------------- move latex project away from temp folder ------------->
-    project_folder = move_project(project_folder, arxiv_id=None)
-
-
-    # <-------------- if merge_translate_zh is already generated, skip gpt req ------------->
-    if not os.path.exists(project_folder + '/merge_proofread_en.tex'):
-        yield from Latex精细分解与转化(file_manifest, project_folder, llm_kwargs, plugin_kwargs, 
-                                chatbot, history, system_prompt, mode='proofread_en', switch_prompt=_switch_prompt_)
-
-
-    # <-------------- compile PDF ------------->
-    success = yield from 编译Latex(chatbot, history, main_file_original='merge', main_file_modified='merge_proofread_en', 
-                             work_folder_original=project_folder, work_folder_modified=project_folder, work_folder=project_folder)
-    
-
-    # <-------------- zip PDF ------------->
-    zip_res = zip_result(project_folder)
-    if success:
-        chatbot.append((f"成功啦", '请查收结果（压缩包）...'))
-        yield from update_ui(chatbot=chatbot, history=history); time.sleep(1) # 刷新界面
-        promote_file_to_downloadzone(file=zip_res, chatbot=chatbot)
-    else:
-        chatbot.append((f"失败了", '虽然PDF生成失败了, 但请查收结果（压缩包）, 内含已经翻译的Tex文档, 也是可读的, 您可以到Github Issue区, 用该压缩包+对话历史存档进行反馈 ...'))
-        yield from update_ui(chatbot=chatbot, history=history); time.sleep(1) # 刷新界面
-        promote_file_to_downloadzone(file=zip_res, chatbot=chatbot)
-
-    # <-------------- we are done ------------->
-    return success
-
-# =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= 插件主程序2 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=    
-
-@CatchException
-def Latex翻译中文并重新编译PDF(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-    # <-------------- information about this plugin ------------->
-    chatbot.append([
-        "函数插件功能？",
-        "对整个Latex项目进行翻译, 生成中文PDF。函数插件贡献者: Binary-Husky。注意事项: 此插件Windows支持最佳，Linux下必须使用Docker安装，详见项目主README.md。目前仅支持GPT3.5/GPT4，其他模型转化效果未知。目前对机器学习类文献转化效果最好，其他类型文献转化效果未知。"])
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-
-    # <-------------- more requirements ------------->
-    if ("advanced_arg" in plugin_kwargs) and (plugin_kwargs["advanced_arg"] == ""): plugin_kwargs.pop("advanced_arg")
-    more_req = plugin_kwargs.get("advanced_arg", "")
-    no_cache = more_req.startswith("--no-cache")
-    if no_cache: more_req.lstrip("--no-cache")
-    allow_cache = not no_cache
-    _switch_prompt_ = partial(switch_prompt, more_requirement=more_req)
-
-    # <-------------- check deps ------------->
-    try:
-        import glob, os, time, subprocess
-        subprocess.Popen(['pdflatex', '-version'])
-        from .latex_fns.latex_actions import Latex精细分解与转化, 编译Latex
-    except Exception as e:
-        chatbot.append([ f"解析项目: {txt}",
-            f"尝试执行Latex指令失败。Latex没有安装, 或者不在环境变量PATH中。安装方法https://tug.org/texlive/。报错信息\n\n```\n\n{trimmed_format_exc()}\n\n```\n\n"])
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        return
-    
-
-    # <-------------- clear history and read input ------------->
-    history = []
-    txt, arxiv_id = yield from arxiv_download(chatbot, history, txt, allow_cache)
-    if txt.endswith('.pdf'):
-        report_exception(chatbot, history, a = f"解析项目: {txt}", b = f"发现已经存在翻译好的PDF文档")
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        return
-    
-
-    if os.path.exists(txt):
-        project_folder = txt
-    else:
-        if txt == "": txt = '空空如也的输入栏'
-        report_exception(chatbot, history, a = f"解析项目: {txt}", b = f"找不到本地项目或无法处理: {txt}")
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        return
-    
-    file_manifest = [f for f in glob.glob(f'{project_folder}/**/*.tex', recursive=True)]
-    if len(file_manifest) == 0:
-        report_exception(chatbot, history, a = f"解析项目: {txt}", b = f"找不到任何.tex文件: {txt}")
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        return
-    
-
-    # <-------------- if is a zip/tar file ------------->
-    project_folder = desend_to_extracted_folder_if_exist(project_folder)
-
-
-    # <-------------- move latex project away from temp folder ------------->
-    project_folder = move_project(project_folder, arxiv_id)
-
-
-    # <-------------- if merge_translate_zh is already generated, skip gpt req ------------->
-    if not os.path.exists(project_folder + '/merge_translate_zh.tex'):
-        yield from Latex精细分解与转化(file_manifest, project_folder, llm_kwargs, plugin_kwargs, 
-                                chatbot, history, system_prompt, mode='translate_zh', switch_prompt=_switch_prompt_)
-
-
-    # <-------------- compile PDF ------------->
-    success = yield from 编译Latex(chatbot, history, main_file_original='merge', main_file_modified='merge_translate_zh', mode='translate_zh', 
-                             work_folder_original=project_folder, work_folder_modified=project_folder, work_folder=project_folder)
-
-    # <-------------- zip PDF ------------->
-    zip_res = zip_result(project_folder)
-    if success:
-        chatbot.append((f"成功啦", '请查收结果（压缩包）...'))
-        yield from update_ui(chatbot=chatbot, history=history); time.sleep(1) # 刷新界面
-        promote_file_to_downloadzone(file=zip_res, chatbot=chatbot)
-    else:
-        chatbot.append((f"失败了", '虽然PDF生成失败了, 但请查收结果（压缩包）, 内含已经翻译的Tex文档, 您可以到Github Issue区, 用该压缩包进行反馈。如系统是Linux，请检查系统字体（见Github wiki） ...'))
-        yield from update_ui(chatbot=chatbot, history=history); time.sleep(1) # 刷新界面
-        promote_file_to_downloadzone(file=zip_res, chatbot=chatbot)
-
-
-    # <-------------- we are done ------------->
-    return success

crazy_functions/agent_fns/pipe.py

@@ -72,7 +72,7 @@ class PluginMultiprocessManager:
         if file_type.lower() in ['png', 'jpg']:
             image_path = os.path.abspath(fp)
             self.chatbot.append([
-                '检测到新生图像:', 
+                '检测到新生图像:',
                 f'本地文件预览: <br/><div align="center"><img src="file={image_path}"></div>'
             ])
             yield from update_ui(chatbot=self.chatbot, history=self.history)
@@ -114,21 +114,21 @@ class PluginMultiprocessManager:
             self.cnt = 1
             self.parent_conn = self.launch_subprocess_with_pipe() # ⭐⭐⭐
         repeated, cmd_to_autogen = self.send_command(txt)
-        if txt == 'exit': 
+        if txt == 'exit':
             self.chatbot.append([f"结束", "结束信号已明确，终止AutoGen程序。"])
             yield from update_ui(chatbot=self.chatbot, history=self.history)
             self.terminate()
             return "terminate"
-        
+
         # patience = 10
-        
+
         while True:
             time.sleep(0.5)
             if not self.alive:
                 # the heartbeat watchdog might have it killed
                 self.terminate()
                 return "terminate"
-            if self.parent_conn.poll(): 
+            if self.parent_conn.poll():
                 self.feed_heartbeat_watchdog()
                 if "[GPT-Academic] 等待中" in self.chatbot[-1][-1]:
                     self.chatbot.pop(-1)  # remove the last line
@@ -152,8 +152,8 @@ class PluginMultiprocessManager:
                     yield from update_ui(chatbot=self.chatbot, history=self.history)
                 if msg.cmd == "interact":
                     yield from self.overwatch_workdir_file_change()
-                    self.chatbot.append([f"程序抵达用户反馈节点.", msg.content + 
-                                         "\n\n等待您的进一步指令." + 
+                    self.chatbot.append([f"程序抵达用户反馈节点.", msg.content +
+                                         "\n\n等待您的进一步指令." +
                                          "\n\n(1) 一般情况下您不需要说什么, 清空输入区, 然后直接点击“提交”以继续. " +
                                          "\n\n(2) 如果您需要补充些什么, 输入要反馈的内容, 直接点击“提交”以继续. " +
                                          "\n\n(3) 如果您想终止程序, 输入exit, 直接点击“提交”以终止AutoGen并解锁. "

crazy_functions/agent_fns/watchdog.py

@@ -8,7 +8,7 @@ class WatchDog():
         self.interval = interval
         self.msg = msg
         self.kill_dog = False
-    
+
     def watch(self):
         while True:
             if self.kill_dog: break

crazy_functions/chatglm微调工具.py

@@ -46,7 +46,7 @@ def 微调数据集生成(txt, llm_kwargs, plugin_kwargs, chatbot, history, syst
     chatbot.append(("这是什么功能？", "[Local Message] 微调数据集生成"))
     if ("advanced_arg" in plugin_kwargs) and (plugin_kwargs["advanced_arg"] == ""): plugin_kwargs.pop("advanced_arg")
     args = plugin_kwargs.get("advanced_arg", None)
-    if args is None: 
+    if args is None:
         chatbot.append(("没给定指令", "退出"))
         yield from update_ui(chatbot=chatbot, history=history); return
     else:
@@ -69,7 +69,7 @@ def 微调数据集生成(txt, llm_kwargs, plugin_kwargs, chatbot, history, syst
             sys_prompt_array=[arguments.system_prompt for _ in (batch)],
             max_workers=10  # OpenAI所允许的最大并行过载
         )
-    
+
         with open(txt+'.generated.json', 'a+', encoding='utf8') as f:
             for b, r in zip(batch, res[1::2]):
                 f.write(json.dumps({"content":b, "summary":r}, ensure_ascii=False)+'\n')
@@ -95,12 +95,12 @@ def 启动微调(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt
     chatbot.append(("这是什么功能？", "[Local Message] 微调数据集生成"))
     if ("advanced_arg" in plugin_kwargs) and (plugin_kwargs["advanced_arg"] == ""): plugin_kwargs.pop("advanced_arg")
     args = plugin_kwargs.get("advanced_arg", None)
-    if args is None: 
+    if args is None:
         chatbot.append(("没给定指令", "退出"))
         yield from update_ui(chatbot=chatbot, history=history); return
     else:
         arguments = string_to_options(arguments=args)
-      
+
 
 
     pre_seq_len = arguments.pre_seq_len             # 128

crazy_functions/crazy_functions_test.py

@@ -1,231 +0,0 @@
-"""
-这是什么？
-    这个文件用于函数插件的单元测试
-    运行方法 python crazy_functions/crazy_functions_test.py
-"""
-    
-# ==============================================================================================================================
-
-def validate_path():
-    import os, sys
-    dir_name = os.path.dirname(__file__)
-    root_dir_assume = os.path.abspath(os.path.dirname(__file__) +  '/..')
-    os.chdir(root_dir_assume)
-    sys.path.append(root_dir_assume)
-validate_path() # validate path so you can run from base directory
-
-# ==============================================================================================================================
-
-from colorful import *
-from toolbox import get_conf, ChatBotWithCookies
-import contextlib
-import os
-import sys
-from functools import wraps
-proxies, WEB_PORT, LLM_MODEL, CONCURRENT_COUNT, AUTHENTICATION, CHATBOT_HEIGHT, LAYOUT, API_KEY = \
-    get_conf('proxies', 'WEB_PORT', 'LLM_MODEL', 'CONCURRENT_COUNT', 'AUTHENTICATION', 'CHATBOT_HEIGHT', 'LAYOUT', 'API_KEY')
-
-llm_kwargs = {
-    'api_key': API_KEY,
-    'llm_model': LLM_MODEL,
-    'top_p':1.0, 
-    'max_length': None,
-    'temperature':1.0,
-}
-plugin_kwargs = { }
-chatbot = ChatBotWithCookies(llm_kwargs)
-history = []
-system_prompt = "Serve me as a writing and programming assistant."
-web_port = 1024
-
-# ==============================================================================================================================
-
-def silence_stdout(func):
-    @wraps(func)
-    def wrapper(*args, **kwargs):
-        _original_stdout = sys.stdout
-        sys.stdout = open(os.devnull, 'w')
-        for q in func(*args, **kwargs):
-            sys.stdout = _original_stdout
-            yield q
-            sys.stdout = open(os.devnull, 'w')
-        sys.stdout.close()
-        sys.stdout = _original_stdout
-    return wrapper
-
-class CLI_Printer():
-    def __init__(self) -> None:
-        self.pre_buf = ""
-
-    def print(self, buf):
-        bufp = ""
-        for index, chat in enumerate(buf):
-            a, b = chat
-            bufp += sprint亮靛('[Me]:' + a) + '\n'
-            bufp += '[GPT]:' + b
-            if index < len(buf)-1: 
-                bufp += '\n'
-
-        if self.pre_buf!="" and bufp.startswith(self.pre_buf):
-            print(bufp[len(self.pre_buf):], end='')
-        else:
-            print('\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n'+bufp, end='')
-        self.pre_buf = bufp
-        return
-    
-cli_printer = CLI_Printer()
-# ==============================================================================================================================
-def test_解析一个Python项目():
-    from crazy_functions.解析项目源代码 import 解析一个Python项目
-    txt = "crazy_functions/test_project/python/dqn"
-    for cookies, cb, hist, msg in 解析一个Python项目(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        print(cb)
-
-def test_解析一个Cpp项目():
-    from crazy_functions.解析项目源代码 import 解析一个C项目
-    txt = "crazy_functions/test_project/cpp/cppipc"
-    for cookies, cb, hist, msg in 解析一个C项目(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        print(cb)
-
-def test_Latex英文润色():
-    from crazy_functions.Latex全文润色 import Latex英文润色
-    txt = "crazy_functions/test_project/latex/attention"
-    for cookies, cb, hist, msg in Latex英文润色(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        print(cb)
-
-def test_Markdown中译英():
-    from crazy_functions.批量Markdown翻译 import Markdown中译英
-    txt = "README.md"
-    for cookies, cb, hist, msg in Markdown中译英(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        print(cb)
-
-def test_批量翻译PDF文档():
-    from crazy_functions.批量翻译PDF文档_多线程 import 批量翻译PDF文档
-    txt = "crazy_functions/test_project/pdf_and_word"
-    for cookies, cb, hist, msg in 批量翻译PDF文档(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        print(cb)
-
-def test_谷歌检索小助手():
-    from crazy_functions.谷歌检索小助手 import 谷歌检索小助手
-    txt = "https://scholar.google.com/scholar?hl=en&as_sdt=0%2C5&q=auto+reinforcement+learning&btnG="
-    for cookies, cb, hist, msg in 谷歌检索小助手(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        print(cb)
-
-def test_总结word文档():
-    from crazy_functions.总结word文档 import 总结word文档
-    txt = "crazy_functions/test_project/pdf_and_word"
-    for cookies, cb, hist, msg in 总结word文档(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        print(cb)
-
-def test_下载arxiv论文并翻译摘要():
-    from crazy_functions.下载arxiv论文翻译摘要 import 下载arxiv论文并翻译摘要
-    txt = "1812.10695"
-    for cookies, cb, hist, msg in 下载arxiv论文并翻译摘要(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        print(cb)
-
-def test_联网回答问题():
-    from crazy_functions.联网的ChatGPT import 连接网络回答问题
-    # txt = "谁是应急食品？"
-    # >>        '根据以上搜索结果可以得知，应急食品是“原神”游戏中的角色派蒙的外号。'
-    # txt = "道路千万条，安全第一条。后面两句是？"
-    # >>        '行车不规范，亲人两行泪。'
-    # txt = "You should have gone for the head. What does that mean?"
-    # >>        The phrase "You should have gone for the head" is a quote from the Marvel movies, Avengers: Infinity War and Avengers: Endgame. It was spoken by the character Thanos in Infinity War and by Thor in Endgame.
-    txt = "AutoGPT是什么？"
-    for cookies, cb, hist, msg in 连接网络回答问题(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port): 
-        print("当前问答：", cb[-1][-1].replace("\n"," "))
-    for i, it in enumerate(cb): print亮蓝(it[0]); print亮黄(it[1])
-
-def test_解析ipynb文件():
-    from crazy_functions.解析JupyterNotebook import 解析ipynb文件
-    txt = "crazy_functions/test_samples"
-    for cookies, cb, hist, msg in 解析ipynb文件(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        print(cb)
-
-
-def test_数学动画生成manim():
-    from crazy_functions.数学动画生成manim import 动画生成
-    txt = "A ball split into 2, and then split into 4, and finally split into 8."
-    for cookies, cb, hist, msg in 动画生成(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        print(cb)
-
-
-
-def test_Markdown多语言():
-    from crazy_functions.批量Markdown翻译 import Markdown翻译指定语言
-    txt = "README.md"
-    history = []
-    for lang in ["English", "French", "Japanese", "Korean", "Russian", "Italian", "German", "Portuguese", "Arabic"]:
-        plugin_kwargs = {"advanced_arg": lang}
-        for cookies, cb, hist, msg in Markdown翻译指定语言(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-            print(cb)
-
-def test_Langchain知识库():
-    from crazy_functions.Langchain知识库 import 知识库问答
-    txt = "./"
-    chatbot = ChatBotWithCookies(llm_kwargs)
-    for cookies, cb, hist, msg in silence_stdout(知识库问答)(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        cli_printer.print(cb)   #  print(cb)
-    
-    chatbot = ChatBotWithCookies(cookies)
-    from crazy_functions.Langchain知识库 import 读取知识库作答
-    txt = "What is the installation method？"
-    for cookies, cb, hist, msg in silence_stdout(读取知识库作答)(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        cli_printer.print(cb)   #  print(cb)
-
-def test_Langchain知识库读取():
-    from crazy_functions.Langchain知识库 import 读取知识库作答
-    txt = "远程云服务器部署？"
-    for cookies, cb, hist, msg in silence_stdout(读取知识库作答)(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        cli_printer.print(cb)   #  print(cb)
-
-def test_Latex():
-    from crazy_functions.Latex输出PDF结果 import Latex英文纠错加PDF对比, Latex翻译中文并重新编译PDF
-
-    # txt = r"https://arxiv.org/abs/1706.03762"
-    # txt = r"https://arxiv.org/abs/1902.03185"
-    # txt = r"https://arxiv.org/abs/2305.18290"
-    # txt = r"https://arxiv.org/abs/2305.17608"
-    # txt = r"https://arxiv.org/abs/2211.16068"                     #  ACE
-    # txt = r"C:\Users\x\arxiv_cache\2211.16068\workfolder"  #  ACE
-    # txt = r"https://arxiv.org/abs/2002.09253"
-    # txt = r"https://arxiv.org/abs/2306.07831"
-    # txt = r"https://arxiv.org/abs/2212.10156"
-    # txt = r"https://arxiv.org/abs/2211.11559"
-    # txt = r"https://arxiv.org/abs/2303.08774"
-    txt = r"https://arxiv.org/abs/2303.12712"
-    # txt = r"C:\Users\fuqingxu\arxiv_cache\2303.12712\workfolder"
-    
-
-    for cookies, cb, hist, msg in (Latex翻译中文并重新编译PDF)(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-        cli_printer.print(cb)   #  print(cb)
-
-
-
-    # txt = "2302.02948.tar"
-    # print(txt)
-    # main_tex, work_folder = Latex预处理(txt)
-    # print('main tex:', main_tex)
-    # res = 编译Latex(main_tex, work_folder)
-    # # for cookies, cb, hist, msg in silence_stdout(编译Latex)(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
-    #     cli_printer.print(cb)   #  print(cb)
-
-
-
-# test_解析一个Python项目()
-# test_Latex英文润色()
-# test_Markdown中译英()
-# test_批量翻译PDF文档()
-# test_谷歌检索小助手()
-# test_总结word文档()
-# test_下载arxiv论文并翻译摘要()
-# test_解析一个Cpp项目()
-# test_联网回答问题()
-# test_解析ipynb文件()
-# test_数学动画生成manim()
-# test_Langchain知识库()
-# test_Langchain知识库读取()
-if __name__ == "__main__":
-    test_Latex()
-    input("程序完成，回车退出。")
-    print("退出。")

crazy_functions/crazy_utils.py

@@ -135,13 +135,25 @@ def request_gpt_model_in_new_thread_with_ui_alive(
     yield from update_ui(chatbot=chatbot, history=[]) # 如果最后成功了，则删除报错信息
     return final_result
 
-def can_multi_process(llm):
-    if llm.startswith('gpt-'): return True
-    if llm.startswith('api2d-'): return True
-    if llm.startswith('azure-'): return True
-    if llm.startswith('spark'): return True
-    if llm.startswith('zhipuai') or llm.startswith('glm-'): return True
-    return False
+def can_multi_process(llm) -> bool:
+    from request_llms.bridge_all import model_info
+
+    def default_condition(llm) -> bool:
+        # legacy condition
+        if llm.startswith('gpt-'): return True
+        if llm.startswith('api2d-'): return True
+        if llm.startswith('azure-'): return True
+        if llm.startswith('spark'): return True
+        if llm.startswith('zhipuai') or llm.startswith('glm-'): return True
+        return False
+
+    if llm in model_info:
+        if 'can_multi_thread' in model_info[llm]:
+            return model_info[llm]['can_multi_thread']
+        else:
+            return default_condition(llm)
+    else:
+        return default_condition(llm)
 
 def request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency(
         inputs_array, inputs_show_user_array, llm_kwargs,

crazy_functions/diagram_fns/file_tree.py

@@ -10,7 +10,7 @@ class FileNode:
         self.parenting_ship = []
         self.comment = ""
         self.comment_maxlen_show = 50
-        
+
     @staticmethod
     def add_linebreaks_at_spaces(string, interval=10):
         return '\n'.join(string[i:i+interval] for i in range(0, len(string), interval))

crazy_functions/game_fns/game_ascii_art.py

@@ -8,7 +8,7 @@ import random
 
 class MiniGame_ASCII_Art(GptAcademicGameBaseState):
     def step(self, prompt, chatbot, history):
-        if self.step_cnt == 0:  
+        if self.step_cnt == 0:
             chatbot.append(["我画你猜（动物）", "请稍等..."])
         else:
             if prompt.strip() == 'exit':

crazy_functions/game_fns/game_interactive_story.py

@@ -88,8 +88,8 @@ class MiniGame_ResumeStory(GptAcademicGameBaseState):
         self.story = []
         chatbot.append(["互动写故事", f"这次的故事开头是：{self.headstart}"])
         self.sys_prompt_ = '你是一个想象力丰富的杰出作家。正在与你的朋友互动，一起写故事，因此你每次写的故事段落应少于300字（结局除外）。'
-        
-        
+
+
     def generate_story_image(self, story_paragraph):
         try:
             from crazy_functions.图片生成 import gen_image
@@ -98,13 +98,13 @@ class MiniGame_ResumeStory(GptAcademicGameBaseState):
             return f'<br/><div align="center"><img src="file={image_path}"></div>'
         except:
             return ''
-        
+
     def step(self, prompt, chatbot, history):
-        
+
         """
         首先，处理游戏初始化等特殊情况
         """
-        if self.step_cnt == 0:  
+        if self.step_cnt == 0:
             self.begin_game_step_0(prompt, chatbot, history)
             self.lock_plugin(chatbot)
             self.cur_task = 'head_start'
@@ -132,7 +132,7 @@ class MiniGame_ResumeStory(GptAcademicGameBaseState):
             inputs_ = prompts_hs.format(headstart=self.headstart)
             history_ = []
             story_paragraph = yield from request_gpt_model_in_new_thread_with_ui_alive(
-                inputs_, '故事开头', self.llm_kwargs, 
+                inputs_, '故事开头', self.llm_kwargs,
                 chatbot, history_, self.sys_prompt_
             )
             self.story.append(story_paragraph)
@@ -147,7 +147,7 @@ class MiniGame_ResumeStory(GptAcademicGameBaseState):
             inputs_ = prompts_interact.format(previously_on_story=previously_on_story)
             history_ = []
             self.next_choices = yield from request_gpt_model_in_new_thread_with_ui_alive(
-                inputs_, '请在以下几种故事走向中，选择一种（当然，您也可以选择给出其他故事走向）：', self.llm_kwargs, 
+                inputs_, '请在以下几种故事走向中，选择一种（当然，您也可以选择给出其他故事走向）：', self.llm_kwargs,
                 chatbot,
                 history_,
                 self.sys_prompt_
@@ -166,7 +166,7 @@ class MiniGame_ResumeStory(GptAcademicGameBaseState):
             inputs_ = prompts_resume.format(previously_on_story=previously_on_story, choice=self.next_choices, user_choice=prompt)
             history_ = []
             story_paragraph = yield from request_gpt_model_in_new_thread_with_ui_alive(
-                inputs_, f'下一段故事（您的选择是：{prompt}）。', self.llm_kwargs, 
+                inputs_, f'下一段故事（您的选择是：{prompt}）。', self.llm_kwargs,
                 chatbot, history_, self.sys_prompt_
             )
             self.story.append(story_paragraph)
@@ -181,10 +181,10 @@ class MiniGame_ResumeStory(GptAcademicGameBaseState):
             inputs_ = prompts_interact.format(previously_on_story=previously_on_story)
             history_ = []
             self.next_choices = yield from request_gpt_model_in_new_thread_with_ui_alive(
-                inputs_, 
-                '请在以下几种故事走向中，选择一种。当然，您也可以给出您心中的其他故事走向。另外，如果您希望剧情立即收尾，请输入剧情走向，并以“剧情收尾”四个字提示程序。', self.llm_kwargs, 
-                chatbot, 
-                history_, 
+                inputs_,
+                '请在以下几种故事走向中，选择一种。当然，您也可以给出您心中的其他故事走向。另外，如果您希望剧情立即收尾，请输入剧情走向，并以“剧情收尾”四个字提示程序。', self.llm_kwargs,
+                chatbot,
+                history_,
                 self.sys_prompt_
             )
             self.cur_task = 'user_choice'
@@ -200,7 +200,7 @@ class MiniGame_ResumeStory(GptAcademicGameBaseState):
             inputs_ = prompts_terminate.format(previously_on_story=previously_on_story, user_choice=prompt)
             history_ = []
             story_paragraph = yield from request_gpt_model_in_new_thread_with_ui_alive(
-                inputs_, f'故事收尾（您的选择是：{prompt}）。', self.llm_kwargs, 
+                inputs_, f'故事收尾（您的选择是：{prompt}）。', self.llm_kwargs,
                 chatbot, history_, self.sys_prompt_
             )
             # # 配图

crazy_functions/game_fns/game_utils.py

@@ -5,7 +5,7 @@ def get_code_block(reply):
     import re
     pattern = r"```([\s\S]*?)```" # regex pattern to match code blocks
     matches = re.findall(pattern, reply) # find all code blocks in text
-    if len(matches) == 1: 
+    if len(matches) == 1:
         return "```" + matches[0] + "```" #  code block
     raise RuntimeError("GPT is not generating proper code.")
 
@@ -13,10 +13,10 @@ def is_same_thing(a, b, llm_kwargs):
     from pydantic import BaseModel, Field
     class IsSameThing(BaseModel):
         is_same_thing: bool = Field(description="determine whether two objects are same thing.", default=False)
-        
-    def run_gpt_fn(inputs, sys_prompt, history=[]): 
+
+    def run_gpt_fn(inputs, sys_prompt, history=[]):
         return predict_no_ui_long_connection(
-            inputs=inputs, llm_kwargs=llm_kwargs, 
+            inputs=inputs, llm_kwargs=llm_kwargs,
             history=history, sys_prompt=sys_prompt, observe_window=[]
         )
 
@@ -24,7 +24,7 @@ def is_same_thing(a, b, llm_kwargs):
     inputs_01 = "Identity whether the user input and the target is the same thing: \n target object: {a} \n user input object: {b} \n\n\n".format(a=a, b=b)
     inputs_01 += "\n\n\n Note that the user may describe the target object with a different language, e.g. cat and 猫 are the same thing."
     analyze_res_cot_01 = run_gpt_fn(inputs_01, "", [])
-    
+
     inputs_02 = inputs_01 + gpt_json_io.format_instructions
     analyze_res = run_gpt_fn(inputs_02, "", [inputs_01, analyze_res_cot_01])
 

crazy_functions/gen_fns/gen_fns_shared.py

@@ -41,11 +41,11 @@ def is_function_successfully_generated(fn_path, class_name, return_dict):
         # Now you can create an instance of the class
         instance = some_class()
         return_dict['success'] = True
-        return 
+        return
     except:
         return_dict['traceback'] = trimmed_format_exc()
         return
-    
+
 def subprocess_worker(code, file_path, return_dict):
     return_dict['result'] = None
     return_dict['success'] = False

crazy_functions/ipc_fns/mp.py

@@ -1,4 +1,4 @@
-import platform 
+import platform
 import pickle
 import multiprocessing
 

crazy_functions/json_fns/pydantic_io.py

@@ -89,7 +89,7 @@ class GptJsonIO():
                  error + "\n\n" + \
                 "Now, fix this json string. \n\n"
         return prompt
-    
+
     def generate_output_auto_repair(self, response, gpt_gen_fn):
         """
         response: string containing canidate json

crazy_functions/latex_fns/latex_actions.py

@@ -90,16 +90,16 @@ class LatexPaperSplit():
             "版权归原文作者所有。翻译内容可靠性无保障，请仔细鉴别并以原文为准。" + \
             "项目Github地址 \\url{https://github.com/binary-husky/gpt_academic/}。"
         # 请您不要删除或修改这行警告，除非您是论文的原作者（如果您是论文原作者，欢迎加REAME中的QQ联系开发者）
-        self.msg_declare = "为了防止大语言模型的意外谬误产生扩散影响，禁止移除或修改此警告。}}\\\\" 
+        self.msg_declare = "为了防止大语言模型的意外谬误产生扩散影响，禁止移除或修改此警告。}}\\\\"
         self.title = "unknown"
         self.abstract = "unknown"
 
     def read_title_and_abstract(self, txt):
         try:
             title, abstract = find_title_and_abs(txt)
-            if title is not None: 
+            if title is not None:
                 self.title = title.replace('\n', ' ').replace('\\\\', ' ').replace('  ', '').replace('  ', '')
-            if abstract is not None: 
+            if abstract is not None:
                 self.abstract = abstract.replace('\n', ' ').replace('\\\\', ' ').replace('  ', '').replace('  ', '')
         except:
             pass
@@ -111,7 +111,7 @@ class LatexPaperSplit():
         result_string = ""
         node_cnt = 0
         line_cnt = 0
-        
+
         for node in self.nodes:
             if node.preserve:
                 line_cnt += node.string.count('\n')
@@ -144,7 +144,7 @@ class LatexPaperSplit():
         return result_string
 
 
-    def split(self, txt, project_folder, opts): 
+    def split(self, txt, project_folder, opts):
         """
         break down latex file to a linked list,
         each node use a preserve flag to indicate whether it should
@@ -155,7 +155,7 @@ class LatexPaperSplit():
         manager = multiprocessing.Manager()
         return_dict = manager.dict()
         p = multiprocessing.Process(
-            target=split_subprocess, 
+            target=split_subprocess,
             args=(txt, project_folder, return_dict, opts))
         p.start()
         p.join()
@@ -217,13 +217,13 @@ def Latex精细分解与转化(file_manifest, project_folder, llm_kwargs, plugin
     from ..crazy_utils import request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency
     from .latex_actions import LatexPaperFileGroup, LatexPaperSplit
 
-    #  <-------- 寻找主tex文件 ----------> 
+    #  <-------- 寻找主tex文件 ---------->
     maintex = find_main_tex_file(file_manifest, mode)
     chatbot.append((f"定位主Latex文件", f'[Local Message] 分析结果：该项目的Latex主文件是{maintex}, 如果分析错误, 请立即终止程序, 删除或修改歧义文件, 然后重试。主程序即将开始, 请稍候。'))
     yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
     time.sleep(3)
 
-    #  <-------- 读取Latex文件, 将多文件tex工程融合为一个巨型tex ----------> 
+    #  <-------- 读取Latex文件, 将多文件tex工程融合为一个巨型tex ---------->
     main_tex_basename = os.path.basename(maintex)
     assert main_tex_basename.endswith('.tex')
     main_tex_basename_bare = main_tex_basename[:-4]
@@ -240,13 +240,13 @@ def Latex精细分解与转化(file_manifest, project_folder, llm_kwargs, plugin
     with open(project_folder + '/merge.tex', 'w', encoding='utf-8', errors='replace') as f:
         f.write(merged_content)
 
-    #  <-------- 精细切分latex文件 ----------> 
+    #  <-------- 精细切分latex文件 ---------->
     chatbot.append((f"Latex文件融合完成", f'[Local Message] 正在精细切分latex文件，这需要一段时间计算，文档越长耗时越长，请耐心等待。'))
     yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
     lps = LatexPaperSplit()
     lps.read_title_and_abstract(merged_content)
     res = lps.split(merged_content, project_folder, opts) # 消耗时间的函数
-    #  <-------- 拆分过长的latex片段 ----------> 
+    #  <-------- 拆分过长的latex片段 ---------->
     pfg = LatexPaperFileGroup()
     for index, r in enumerate(res):
         pfg.file_paths.append('segment-' + str(index))
@@ -255,17 +255,17 @@ def Latex精细分解与转化(file_manifest, project_folder, llm_kwargs, plugin
     pfg.run_file_split(max_token_limit=1024)
     n_split = len(pfg.sp_file_contents)
 
-    #  <-------- 根据需要切换prompt ----------> 
+    #  <-------- 根据需要切换prompt ---------->
     inputs_array, sys_prompt_array = switch_prompt(pfg, mode)
     inputs_show_user_array = [f"{mode} {f}" for f in pfg.sp_file_tag]
 
     if os.path.exists(pj(project_folder,'temp.pkl')):
 
-        #  <-------- 【仅调试】如果存在调试缓存文件，则跳过GPT请求环节 ----------> 
+        #  <-------- 【仅调试】如果存在调试缓存文件，则跳过GPT请求环节 ---------->
         pfg = objload(file=pj(project_folder,'temp.pkl'))
 
     else:
-        #  <-------- gpt 多线程请求 ----------> 
+        #  <-------- gpt 多线程请求 ---------->
         history_array = [[""] for _ in range(n_split)]
         # LATEX_EXPERIMENTAL, = get_conf('LATEX_EXPERIMENTAL')
         # if LATEX_EXPERIMENTAL:
@@ -284,32 +284,32 @@ def Latex精细分解与转化(file_manifest, project_folder, llm_kwargs, plugin
             scroller_max_len = 40
         )
 
-        #  <-------- 文本碎片重组为完整的tex片段 ----------> 
+        #  <-------- 文本碎片重组为完整的tex片段 ---------->
         pfg.sp_file_result = []
         for i_say, gpt_say, orig_content in zip(gpt_response_collection[0::2], gpt_response_collection[1::2], pfg.sp_file_contents):
             pfg.sp_file_result.append(gpt_say)
         pfg.merge_result()
 
-        # <-------- 临时存储用于调试 ----------> 
+        # <-------- 临时存储用于调试 ---------->
         pfg.get_token_num = None
         objdump(pfg, file=pj(project_folder,'temp.pkl'))
 
     write_html(pfg.sp_file_contents, pfg.sp_file_result, chatbot=chatbot, project_folder=project_folder)
 
-    #  <-------- 写出文件 ----------> 
+    #  <-------- 写出文件 ---------->
     msg = f"当前大语言模型: {llm_kwargs['llm_model']}，当前语言模型温度设定: {llm_kwargs['temperature']}。"
     final_tex = lps.merge_result(pfg.file_result, mode, msg)
     objdump((lps, pfg.file_result, mode, msg), file=pj(project_folder,'merge_result.pkl'))
 
     with open(project_folder + f'/merge_{mode}.tex', 'w', encoding='utf-8', errors='replace') as f:
         if mode != 'translate_zh' or "binary" in final_tex: f.write(final_tex)
-        
 
-    #  <-------- 整理结果, 退出 ----------> 
+
+    #  <-------- 整理结果, 退出 ---------->
     chatbot.append((f"完成了吗？", 'GPT结果已输出, 即将编译PDF'))
     yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
 
-    #  <-------- 返回 ----------> 
+    #  <-------- 返回 ---------->
     return project_folder + f'/merge_{mode}.tex'
 
 
@@ -362,7 +362,7 @@ def 编译Latex(chatbot, history, main_file_original, main_file_modified, work_f
 
         yield from update_ui_lastest_msg(f'尝试第 {n_fix}/{max_try} 次编译, 编译转化后的PDF ...', chatbot, history)   # 刷新Gradio前端界面
         ok = compile_latex_with_timeout(f'pdflatex -interaction=batchmode -file-line-error {main_file_modified}.tex', work_folder_modified)
-        
+
         if ok and os.path.exists(pj(work_folder_modified, f'{main_file_modified}.pdf')):
             # 只有第二步成功，才能继续下面的步骤
             yield from update_ui_lastest_msg(f'尝试第 {n_fix}/{max_try} 次编译, 编译BibTex ...', chatbot, history)    # 刷新Gradio前端界面
@@ -393,9 +393,9 @@ def 编译Latex(chatbot, history, main_file_original, main_file_modified, work_f
         original_pdf_success = os.path.exists(pj(work_folder_original, f'{main_file_original}.pdf'))
         modified_pdf_success = os.path.exists(pj(work_folder_modified, f'{main_file_modified}.pdf'))
         diff_pdf_success     = os.path.exists(pj(work_folder, f'merge_diff.pdf'))
-        results_ += f"原始PDF编译是否成功: {original_pdf_success};" 
-        results_ += f"转化PDF编译是否成功: {modified_pdf_success};" 
-        results_ += f"对比PDF编译是否成功: {diff_pdf_success};" 
+        results_ += f"原始PDF编译是否成功: {original_pdf_success};"
+        results_ += f"转化PDF编译是否成功: {modified_pdf_success};"
+        results_ += f"对比PDF编译是否成功: {diff_pdf_success};"
         yield from update_ui_lastest_msg(f'第{n_fix}编译结束:<br/>{results_}...', chatbot, history) # 刷新Gradio前端界面
 
         if diff_pdf_success:
@@ -409,7 +409,7 @@ def 编译Latex(chatbot, history, main_file_original, main_file_modified, work_f
                 shutil.copyfile(result_pdf, pj(work_folder, '..', 'translation', 'translate_zh.pdf'))
             promote_file_to_downloadzone(result_pdf, rename_file=None, chatbot=chatbot)  # promote file to web UI
             # 将两个PDF拼接
-            if original_pdf_success: 
+            if original_pdf_success:
                 try:
                     from .latex_toolbox import merge_pdfs
                     concat_pdf = pj(work_folder_modified, f'comparison.pdf')
@@ -425,7 +425,7 @@ def 编译Latex(chatbot, history, main_file_original, main_file_modified, work_f
             if n_fix>=max_try: break
             n_fix += 1
             can_retry, main_file_modified, buggy_lines = remove_buggy_lines(
-                file_path=pj(work_folder_modified, f'{main_file_modified}.tex'), 
+                file_path=pj(work_folder_modified, f'{main_file_modified}.tex'),
                 log_path=pj(work_folder_modified, f'{main_file_modified}.log'),
                 tex_name=f'{main_file_modified}.tex',
                 tex_name_pure=f'{main_file_modified}',
@@ -445,14 +445,14 @@ def write_html(sp_file_contents, sp_file_result, chatbot, project_folder):
         import shutil
         from crazy_functions.pdf_fns.report_gen_html import construct_html
         from toolbox import gen_time_str
-        ch = construct_html() 
+        ch = construct_html()
         orig = ""
         trans = ""
         final = []
-        for c,r in zip(sp_file_contents, sp_file_result): 
+        for c,r in zip(sp_file_contents, sp_file_result):
             final.append(c)
             final.append(r)
-        for i, k in enumerate(final): 
+        for i, k in enumerate(final):
             if i%2==0:
                 orig = k
             if i%2==1:

crazy_functions/latex_utils.py

@@ -1,788 +0,0 @@
-from toolbox import update_ui, update_ui_lastest_msg    # 刷新Gradio前端界面
-from toolbox import zip_folder, objdump, objload, promote_file_to_downloadzone
-import os, shutil
-import re
-import numpy as np
-pj = os.path.join
-
-"""
-========================================================================
-Part One
-Latex segmentation with a binary mask (PRESERVE=0, TRANSFORM=1)
-========================================================================
-"""
-PRESERVE = 0
-TRANSFORM = 1
-
-def set_forbidden_text(text, mask, pattern, flags=0):
-    """
-    Add a preserve text area in this paper
-    e.g. with pattern = r"\\begin\{algorithm\}(.*?)\\end\{algorithm\}"
-    you can mask out (mask = PRESERVE so that text become untouchable for GPT) 
-    everything between "\begin{equation}" and "\end{equation}"
-    """
-    if isinstance(pattern, list): pattern = '|'.join(pattern)
-    pattern_compile = re.compile(pattern, flags)
-    for res in pattern_compile.finditer(text):
-        mask[res.span()[0]:res.span()[1]] = PRESERVE
-    return text, mask
-
-def reverse_forbidden_text(text, mask, pattern, flags=0, forbid_wrapper=True):
-    """
-    Move area out of preserve area (make text editable for GPT)
-    count the number of the braces so as to catch compelete text area. 
-    e.g.
-    \begin{abstract} blablablablablabla. \end{abstract} 
-    """
-    if isinstance(pattern, list): pattern = '|'.join(pattern)
-    pattern_compile = re.compile(pattern, flags)
-    for res in pattern_compile.finditer(text):
-        if not forbid_wrapper:
-            mask[res.span()[0]:res.span()[1]] = TRANSFORM
-        else:
-            mask[res.regs[0][0]: res.regs[1][0]] = PRESERVE   # '\\begin{abstract}'
-            mask[res.regs[1][0]: res.regs[1][1]] = TRANSFORM   # abstract
-            mask[res.regs[1][1]: res.regs[0][1]] = PRESERVE   # abstract
-    return text, mask
-
-def set_forbidden_text_careful_brace(text, mask, pattern, flags=0):
-    """
-    Add a preserve text area in this paper (text become untouchable for GPT).
-    count the number of the braces so as to catch compelete text area. 
-    e.g.
-    \caption{blablablablabla\texbf{blablabla}blablabla.} 
-    """
-    pattern_compile = re.compile(pattern, flags)
-    for res in pattern_compile.finditer(text):
-        brace_level = -1
-        p = begin = end = res.regs[0][0]
-        for _ in range(1024*16):
-            if text[p] == '}' and brace_level == 0: break
-            elif text[p] == '}':  brace_level -= 1
-            elif text[p] == '{':  brace_level += 1
-            p += 1
-        end = p+1
-        mask[begin:end] = PRESERVE
-    return text, mask
-
-def reverse_forbidden_text_careful_brace(text, mask, pattern, flags=0, forbid_wrapper=True):
-    """
-    Move area out of preserve area (make text editable for GPT)
-    count the number of the braces so as to catch compelete text area. 
-    e.g.
-    \caption{blablablablabla\texbf{blablabla}blablabla.} 
-    """
-    pattern_compile = re.compile(pattern, flags)
-    for res in pattern_compile.finditer(text):
-        brace_level = 0
-        p = begin = end = res.regs[1][0]
-        for _ in range(1024*16):
-            if text[p] == '}' and brace_level == 0: break
-            elif text[p] == '}':  brace_level -= 1
-            elif text[p] == '{':  brace_level += 1
-            p += 1
-        end = p
-        mask[begin:end] = TRANSFORM
-        if forbid_wrapper:
-            mask[res.regs[0][0]:begin] = PRESERVE
-            mask[end:res.regs[0][1]] = PRESERVE
-    return text, mask
-
-def set_forbidden_text_begin_end(text, mask, pattern, flags=0, limit_n_lines=42):
-    """
-    Find all \begin{} ... \end{} text block that with less than limit_n_lines lines.
-    Add it to preserve area
-    """
-    pattern_compile = re.compile(pattern, flags)
-    def search_with_line_limit(text, mask):
-        for res in pattern_compile.finditer(text):
-            cmd = res.group(1)  # begin{what}
-            this = res.group(2) # content between begin and end
-            this_mask = mask[res.regs[2][0]:res.regs[2][1]]
-            white_list = ['document', 'abstract', 'lemma', 'definition', 'sproof', 
-                          'em', 'emph', 'textit', 'textbf', 'itemize', 'enumerate']
-            if (cmd in white_list) or this.count('\n') >= limit_n_lines: # use a magical number 42
-                this, this_mask = search_with_line_limit(this, this_mask)
-                mask[res.regs[2][0]:res.regs[2][1]] = this_mask
-            else:
-                mask[res.regs[0][0]:res.regs[0][1]] = PRESERVE
-        return text, mask
-    return search_with_line_limit(text, mask) 
-
-class LinkedListNode():
-    """
-    Linked List Node
-    """
-    def __init__(self, string, preserve=True) -> None:
-        self.string = string
-        self.preserve = preserve
-        self.next = None
-        # self.begin_line = 0
-        # self.begin_char = 0
-
-def convert_to_linklist(text, mask):
-    root = LinkedListNode("", preserve=True)
-    current_node = root
-    for c, m, i in zip(text, mask, range(len(text))):
-        if (m==PRESERVE and current_node.preserve) \
-            or (m==TRANSFORM and not current_node.preserve):
-            # add
-            current_node.string += c
-        else:
-            current_node.next = LinkedListNode(c, preserve=(m==PRESERVE))
-            current_node = current_node.next
-    return root
-"""
-========================================================================
-Latex Merge File
-========================================================================
-"""
-
-def 寻找Latex主文件(file_manifest, mode):
-    """
-    在多Tex文档中，寻找主文件，必须包含documentclass，返回找到的第一个。
-    P.S. 但愿没人把latex模板放在里面传进来 (6.25 加入判定latex模板的代码)
-    """
-    canidates = []
-    for texf in file_manifest:
-        if os.path.basename(texf).startswith('merge'):
-            continue
-        with open(texf, 'r', encoding='utf8') as f:
-            file_content = f.read()
-        if r'\documentclass' in file_content:
-            canidates.append(texf)
-        else:
-            continue
-
-    if len(canidates) == 0:
-        raise RuntimeError('无法找到一个主Tex文件（包含documentclass关键字）')
-    elif len(canidates) == 1:
-        return canidates[0]
-    else: # if len(canidates) >= 2 通过一些Latex模板中常见（但通常不会出现在正文）的单词，对不同latex源文件扣分，取评分最高者返回
-        canidates_score = []
-        # 给出一些判定模板文档的词作为扣分项
-        unexpected_words = ['\LaTeX', 'manuscript', 'Guidelines', 'font', 'citations', 'rejected', 'blind review', 'reviewers']
-        expected_words = ['\input', '\ref', '\cite']
-        for texf in canidates:
-            canidates_score.append(0)
-            with open(texf, 'r', encoding='utf8') as f:
-                file_content = f.read()
-            for uw in unexpected_words:
-                if uw in file_content:
-                    canidates_score[-1] -= 1
-            for uw in expected_words:
-                if uw in file_content:
-                    canidates_score[-1] += 1
-        select = np.argmax(canidates_score) # 取评分最高者返回
-        return canidates[select]
-    
-def rm_comments(main_file):
-    new_file_remove_comment_lines = []
-    for l in main_file.splitlines():
-        # 删除整行的空注释
-        if l.lstrip().startswith("%"):
-            pass
-        else:
-            new_file_remove_comment_lines.append(l)
-    main_file = '\n'.join(new_file_remove_comment_lines)
-    # main_file = re.sub(r"\\include{(.*?)}", r"\\input{\1}", main_file)  # 将 \include 命令转换为 \input 命令
-    main_file = re.sub(r'(?<!\\)%.*', '', main_file)  # 使用正则表达式查找半行注释, 并替换为空字符串
-    return main_file
-
-def merge_tex_files_(project_foler, main_file, mode):
-    """
-    Merge Tex project recrusively
-    """
-    main_file = rm_comments(main_file)
-    for s in reversed([q for q in re.finditer(r"\\input\{(.*?)\}", main_file, re.M)]):
-        f = s.group(1)
-        fp = os.path.join(project_foler, f)
-        if os.path.exists(fp):  
-            # e.g., \input{srcs/07_appendix.tex}
-            with open(fp, 'r', encoding='utf-8', errors='replace') as fx:
-                c = fx.read()
-        else:  
-            # e.g., \input{srcs/07_appendix}
-            with open(fp+'.tex', 'r', encoding='utf-8', errors='replace') as fx:
-                c = fx.read()
-        c = merge_tex_files_(project_foler, c, mode)
-        main_file = main_file[:s.span()[0]] + c + main_file[s.span()[1]:]
-    return main_file
-
-def merge_tex_files(project_foler, main_file, mode):
-    """
-    Merge Tex project recrusively
-    P.S. 顺便把CTEX塞进去以支持中文
-    P.S. 顺便把Latex的注释去除
-    """
-    main_file = merge_tex_files_(project_foler, main_file, mode)
-    main_file = rm_comments(main_file)
-
-    if mode == 'translate_zh':
-        # find paper documentclass
-        pattern = re.compile(r'\\documentclass.*\n')
-        match = pattern.search(main_file)
-        assert match is not None, "Cannot find documentclass statement!"
-        position = match.end()
-        add_ctex = '\\usepackage{ctex}\n'
-        add_url = '\\usepackage{url}\n' if '{url}' not in main_file else ''
-        main_file = main_file[:position] + add_ctex + add_url + main_file[position:]
-        # fontset=windows
-        import platform
-        main_file = re.sub(r"\\documentclass\[(.*?)\]{(.*?)}", r"\\documentclass[\1,fontset=windows,UTF8]{\2}",main_file)
-        main_file = re.sub(r"\\documentclass{(.*?)}", r"\\documentclass[fontset=windows,UTF8]{\1}",main_file)
-        # find paper abstract
-        pattern_opt1 = re.compile(r'\\begin\{abstract\}.*\n')
-        pattern_opt2 = re.compile(r"\\abstract\{(.*?)\}", flags=re.DOTALL)
-        match_opt1 = pattern_opt1.search(main_file)
-        match_opt2 = pattern_opt2.search(main_file)
-        assert (match_opt1 is not None) or (match_opt2 is not None), "Cannot find paper abstract section!"
-    return main_file
-
-
-
-"""
-========================================================================
-Post process
-========================================================================
-"""
-def mod_inbraket(match):
-    """
-    为啥chatgpt会把cite里面的逗号换成中文逗号呀 
-    """
-    # get the matched string
-    cmd = match.group(1)
-    str_to_modify = match.group(2)
-    # modify the matched string
-    str_to_modify = str_to_modify.replace('：', ':')    # 前面是中文冒号，后面是英文冒号
-    str_to_modify = str_to_modify.replace('，', ',')    # 前面是中文逗号，后面是英文逗号
-    # str_to_modify = 'BOOM'
-    return "\\" + cmd + "{" + str_to_modify + "}"
-
-def fix_content(final_tex, node_string):
-    """
-    Fix common GPT errors to increase success rate
-    """
-    final_tex = re.sub(r"(?<!\\)%", "\\%", final_tex)
-    final_tex = re.sub(r"\\([a-z]{2,10})\ \{", r"\\\1{", string=final_tex)
-    final_tex = re.sub(r"\\\ ([a-z]{2,10})\{", r"\\\1{", string=final_tex)
-    final_tex = re.sub(r"\\([a-z]{2,10})\{([^\}]*?)\}", mod_inbraket, string=final_tex)
-
-    if "Traceback" in final_tex and "[Local Message]" in final_tex:
-        final_tex = node_string # 出问题了，还原原文
-    if node_string.count('\\begin') != final_tex.count('\\begin'):
-        final_tex = node_string # 出问题了，还原原文
-    if node_string.count('\_') > 0 and node_string.count('\_') > final_tex.count('\_'):
-        # walk and replace any _ without \
-        final_tex = re.sub(r"(?<!\\)_", "\\_", final_tex)
-
-    def compute_brace_level(string):
-        # this function count the number of { and }
-        brace_level = 0
-        for c in string:
-            if c == "{": brace_level += 1
-            elif c == "}": brace_level -= 1
-        return brace_level
-    def join_most(tex_t, tex_o):
-        # this function join translated string and original string when something goes wrong
-        p_t = 0
-        p_o = 0
-        def find_next(string, chars, begin):
-            p = begin
-            while p < len(string):
-                if string[p] in chars: return p, string[p]
-                p += 1
-            return None, None
-        while True:
-            res1, char = find_next(tex_o, ['{','}'], p_o)
-            if res1 is None: break
-            res2, char = find_next(tex_t, [char], p_t)
-            if res2 is None: break
-            p_o = res1 + 1
-            p_t = res2 + 1
-        return tex_t[:p_t] + tex_o[p_o:]
-
-    if compute_brace_level(final_tex) != compute_brace_level(node_string):
-        # 出问题了，还原部分原文，保证括号正确
-        final_tex = join_most(final_tex, node_string)
-    return final_tex
-
-def split_subprocess(txt, project_folder, return_dict, opts):
-    """
-    break down latex file to a linked list,
-    each node use a preserve flag to indicate whether it should
-    be proccessed by GPT.
-    """
-    text = txt
-    mask = np.zeros(len(txt), dtype=np.uint8) + TRANSFORM
-
-    # 吸收title与作者以上的部分
-    text, mask = set_forbidden_text(text, mask, r"(.*?)\\maketitle", re.DOTALL)
-    # 吸收iffalse注释
-    text, mask = set_forbidden_text(text, mask, r"\\iffalse(.*?)\\fi", re.DOTALL)
-    # 吸收在42行以内的begin-end组合
-    text, mask = set_forbidden_text_begin_end(text, mask, r"\\begin\{([a-z\*]*)\}(.*?)\\end\{\1\}", re.DOTALL, limit_n_lines=42)
-    # 吸收匿名公式
-    text, mask = set_forbidden_text(text, mask, [ r"\$\$(.*?)\$\$",  r"\\\[.*?\\\]" ], re.DOTALL)
-    # 吸收其他杂项
-    text, mask = set_forbidden_text(text, mask, [ r"\\section\{(.*?)\}", r"\\section\*\{(.*?)\}", r"\\subsection\{(.*?)\}", r"\\subsubsection\{(.*?)\}" ])
-    text, mask = set_forbidden_text(text, mask, [ r"\\bibliography\{(.*?)\}", r"\\bibliographystyle\{(.*?)\}" ])
-    text, mask = set_forbidden_text(text, mask, r"\\begin\{thebibliography\}.*?\\end\{thebibliography\}", re.DOTALL)
-    text, mask = set_forbidden_text(text, mask, r"\\begin\{lstlisting\}(.*?)\\end\{lstlisting\}", re.DOTALL)
-    text, mask = set_forbidden_text(text, mask, r"\\begin\{wraptable\}(.*?)\\end\{wraptable\}", re.DOTALL)
-    text, mask = set_forbidden_text(text, mask, r"\\begin\{algorithm\}(.*?)\\end\{algorithm\}", re.DOTALL)
-    text, mask = set_forbidden_text(text, mask, [r"\\begin\{wrapfigure\}(.*?)\\end\{wrapfigure\}", r"\\begin\{wrapfigure\*\}(.*?)\\end\{wrapfigure\*\}"], re.DOTALL)
-    text, mask = set_forbidden_text(text, mask, [r"\\begin\{figure\}(.*?)\\end\{figure\}", r"\\begin\{figure\*\}(.*?)\\end\{figure\*\}"], re.DOTALL)
-    text, mask = set_forbidden_text(text, mask, [r"\\begin\{multline\}(.*?)\\end\{multline\}", r"\\begin\{multline\*\}(.*?)\\end\{multline\*\}"], re.DOTALL)
-    text, mask = set_forbidden_text(text, mask, [r"\\begin\{table\}(.*?)\\end\{table\}", r"\\begin\{table\*\}(.*?)\\end\{table\*\}"], re.DOTALL)
-    text, mask = set_forbidden_text(text, mask, [r"\\begin\{minipage\}(.*?)\\end\{minipage\}", r"\\begin\{minipage\*\}(.*?)\\end\{minipage\*\}"], re.DOTALL)
-    text, mask = set_forbidden_text(text, mask, [r"\\begin\{align\*\}(.*?)\\end\{align\*\}", r"\\begin\{align\}(.*?)\\end\{align\}"], re.DOTALL)
-    text, mask = set_forbidden_text(text, mask, [r"\\begin\{equation\}(.*?)\\end\{equation\}", r"\\begin\{equation\*\}(.*?)\\end\{equation\*\}"], re.DOTALL)
-    text, mask = set_forbidden_text(text, mask, [r"\\includepdf\[(.*?)\]\{(.*?)\}", r"\\clearpage", r"\\newpage", r"\\appendix", r"\\tableofcontents", r"\\include\{(.*?)\}"])
-    text, mask = set_forbidden_text(text, mask, [r"\\vspace\{(.*?)\}", r"\\hspace\{(.*?)\}", r"\\label\{(.*?)\}", r"\\begin\{(.*?)\}", r"\\end\{(.*?)\}", r"\\item "])
-    text, mask = set_forbidden_text_careful_brace(text, mask, r"\\hl\{(.*?)\}", re.DOTALL)
-    # reverse 操作必须放在最后
-    text, mask = reverse_forbidden_text_careful_brace(text, mask, r"\\caption\{(.*?)\}", re.DOTALL, forbid_wrapper=True)
-    text, mask = reverse_forbidden_text_careful_brace(text, mask, r"\\abstract\{(.*?)\}", re.DOTALL, forbid_wrapper=True)
-    text, mask = reverse_forbidden_text(text, mask, r"\\begin\{abstract\}(.*?)\\end\{abstract\}", re.DOTALL, forbid_wrapper=True)
-    root = convert_to_linklist(text, mask)
-
-    # 修复括号
-    node = root
-    while True:
-        string = node.string
-        if node.preserve: 
-            node = node.next
-            if node is None: break
-            continue
-        def break_check(string):
-            str_stack = [""] # (lv, index)
-            for i, c in enumerate(string):
-                if c == '{':
-                    str_stack.append('{')
-                elif c == '}':
-                    if len(str_stack) == 1:
-                        print('stack fix')
-                        return i
-                    str_stack.pop(-1)
-                else:
-                    str_stack[-1] += c
-            return -1
-        bp = break_check(string)
-
-        if bp == -1:
-            pass
-        elif bp == 0:
-            node.string = string[:1]
-            q = LinkedListNode(string[1:], False)
-            q.next = node.next
-            node.next = q
-        else:
-            node.string = string[:bp]
-            q = LinkedListNode(string[bp:], False)
-            q.next = node.next
-            node.next = q
-
-        node = node.next
-        if node is None: break
-
-    # 屏蔽空行和太短的句子
-    node = root
-    while True:
-        if len(node.string.strip('\n').strip(''))==0: node.preserve = True
-        if len(node.string.strip('\n').strip(''))<42: node.preserve = True
-        node = node.next
-        if node is None: break
-    node = root
-    while True:
-        if node.next and node.preserve and node.next.preserve:
-            node.string += node.next.string
-            node.next = node.next.next
-        node = node.next
-        if node is None: break
-
-    # 将前后断行符脱离
-    node = root
-    prev_node = None
-    while True:
-        if not node.preserve:
-            lstriped_ = node.string.lstrip().lstrip('\n')
-            if (prev_node is not None) and (prev_node.preserve) and (len(lstriped_)!=len(node.string)):
-                prev_node.string += node.string[:-len(lstriped_)]
-                node.string = lstriped_
-            rstriped_ = node.string.rstrip().rstrip('\n')
-            if (node.next is not None) and (node.next.preserve) and (len(rstriped_)!=len(node.string)):
-                node.next.string = node.string[len(rstriped_):] + node.next.string
-                node.string = rstriped_
-        # =====
-        prev_node = node
-        node = node.next
-        if node is None: break
-    # 输出html调试文件，用红色标注处保留区（PRESERVE），用黑色标注转换区（TRANSFORM）
-    with open(pj(project_folder, 'debug_log.html'), 'w', encoding='utf8') as f:
-        segment_parts_for_gpt = []
-        nodes = []
-        node = root
-        while True:
-            nodes.append(node)
-            show_html = node.string.replace('\n','<br/>')
-            if not node.preserve:
-                segment_parts_for_gpt.append(node.string)
-                f.write(f'<p style="color:black;">#{show_html}#</p>')
-            else:
-                f.write(f'<p style="color:red;">{show_html}</p>')
-            node = node.next
-            if node is None: break
-
-    for n in nodes: n.next = None   # break
-    return_dict['nodes'] = nodes
-    return_dict['segment_parts_for_gpt'] = segment_parts_for_gpt
-    return return_dict
-
-
-
-class LatexPaperSplit():
-    """
-    break down latex file to a linked list,
-    each node use a preserve flag to indicate whether it should
-    be proccessed by GPT.
-    """
-    def __init__(self) -> None:
-        self.nodes = None
-        self.msg = "*{\\scriptsize\\textbf{警告：该PDF由GPT-Academic开源项目调用大语言模型+Latex翻译插件一键生成，" + \
-            "版权归原文作者所有。翻译内容可靠性无保障，请仔细鉴别并以原文为准。" + \
-            "项目Github地址 \\url{https://github.com/binary-husky/gpt_academic/}。"
-        # 请您不要删除或修改这行警告，除非您是论文的原作者（如果您是论文原作者，欢迎加REAME中的QQ联系开发者）
-        self.msg_declare = "为了防止大语言模型的意外谬误产生扩散影响，禁止移除或修改此警告。}}\\\\" 
-
-    def merge_result(self, arr, mode, msg):
-        """
-        Merge the result after the GPT process completed
-        """
-        result_string = ""
-        p = 0
-        for node in self.nodes:
-            if node.preserve:
-                result_string += node.string
-            else:
-                result_string += fix_content(arr[p], node.string)
-                p += 1
-        if mode == 'translate_zh':
-            pattern = re.compile(r'\\begin\{abstract\}.*\n')
-            match = pattern.search(result_string)
-            if not match:
-                # match \abstract{xxxx}
-                pattern_compile = re.compile(r"\\abstract\{(.*?)\}", flags=re.DOTALL)
-                match = pattern_compile.search(result_string)
-                position = match.regs[1][0]
-            else:
-                # match \begin{abstract}xxxx\end{abstract}
-                position = match.end()
-            result_string = result_string[:position] + self.msg + msg + self.msg_declare + result_string[position:]
-        return result_string
-
-    def split(self, txt, project_folder, opts): 
-        """
-        break down latex file to a linked list,
-        each node use a preserve flag to indicate whether it should
-        be proccessed by GPT.
-        P.S. use multiprocessing to avoid timeout error
-        """
-        import multiprocessing
-        manager = multiprocessing.Manager()
-        return_dict = manager.dict()
-        p = multiprocessing.Process(
-            target=split_subprocess, 
-            args=(txt, project_folder, return_dict, opts))
-        p.start()
-        p.join()
-        p.close()
-        self.nodes = return_dict['nodes']
-        self.sp = return_dict['segment_parts_for_gpt']
-        return self.sp
-
-
-
-class LatexPaperFileGroup():
-    """
-    use tokenizer to break down text according to max_token_limit
-    """
-    def __init__(self):
-        self.file_paths = []
-        self.file_contents = []
-        self.sp_file_contents = []
-        self.sp_file_index = []
-        self.sp_file_tag = []
-
-        # count_token
-        from request_llm.bridge_all import model_info
-        enc = model_info["gpt-3.5-turbo"]['tokenizer']
-        def get_token_num(txt): return len(enc.encode(txt, disallowed_special=()))
-        self.get_token_num = get_token_num
-
-    def run_file_split(self, max_token_limit=1900):
-        """
-        use tokenizer to break down text according to max_token_limit
-        """
-        for index, file_content in enumerate(self.file_contents):
-            if self.get_token_num(file_content) < max_token_limit:
-                self.sp_file_contents.append(file_content)
-                self.sp_file_index.append(index)
-                self.sp_file_tag.append(self.file_paths[index])
-            else:
-                from .crazy_utils import breakdown_txt_to_satisfy_token_limit_for_pdf
-                segments = breakdown_txt_to_satisfy_token_limit_for_pdf(file_content, self.get_token_num, max_token_limit)
-                for j, segment in enumerate(segments):
-                    self.sp_file_contents.append(segment)
-                    self.sp_file_index.append(index)
-                    self.sp_file_tag.append(self.file_paths[index] + f".part-{j}.tex")
-        print('Segmentation: done')
-
-    def merge_result(self):
-        self.file_result = ["" for _ in range(len(self.file_paths))]
-        for r, k in zip(self.sp_file_result, self.sp_file_index):
-            self.file_result[k] += r
-
-    def write_result(self):
-        manifest = []
-        for path, res in zip(self.file_paths, self.file_result):
-            with open(path + '.polish.tex', 'w', encoding='utf8') as f:
-                manifest.append(path + '.polish.tex')
-                f.write(res)
-        return manifest
-
-def write_html(sp_file_contents, sp_file_result, chatbot, project_folder):
-
-    # write html
-    try:
-        import shutil
-        from .crazy_utils import construct_html
-        from toolbox import gen_time_str
-        ch = construct_html() 
-        orig = ""
-        trans = ""
-        final = []
-        for c,r in zip(sp_file_contents, sp_file_result): 
-            final.append(c)
-            final.append(r)
-        for i, k in enumerate(final): 
-            if i%2==0:
-                orig = k
-            if i%2==1:
-                trans = k
-                ch.add_row(a=orig, b=trans)
-        create_report_file_name = f"{gen_time_str()}.trans.html"
-        ch.save_file(create_report_file_name)
-        shutil.copyfile(pj('./gpt_log/', create_report_file_name), pj(project_folder, create_report_file_name))
-        promote_file_to_downloadzone(file=f'./gpt_log/{create_report_file_name}', chatbot=chatbot)
-    except:
-        from toolbox import trimmed_format_exc
-        print('writing html result failed:', trimmed_format_exc())
-
-def Latex精细分解与转化(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, mode='proofread', switch_prompt=None, opts=[]):
-    import time, os, re
-    from .crazy_utils import request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency
-    from .latex_utils import LatexPaperFileGroup, merge_tex_files, LatexPaperSplit, 寻找Latex主文件
-
-    #  <-------- 寻找主tex文件 ----------> 
-    maintex = 寻找Latex主文件(file_manifest, mode)
-    chatbot.append((f"定位主Latex文件", f'[Local Message] 分析结果：该项目的Latex主文件是{maintex}, 如果分析错误, 请立即终止程序, 删除或修改歧义文件, 然后重试。主程序即将开始, 请稍候。'))
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-    time.sleep(3)
-
-    #  <-------- 读取Latex文件, 将多文件tex工程融合为一个巨型tex ----------> 
-    main_tex_basename = os.path.basename(maintex)
-    assert main_tex_basename.endswith('.tex')
-    main_tex_basename_bare = main_tex_basename[:-4]
-    may_exist_bbl = pj(project_folder, f'{main_tex_basename_bare}.bbl')
-    if os.path.exists(may_exist_bbl):
-        shutil.copyfile(may_exist_bbl, pj(project_folder, f'merge.bbl'))
-        shutil.copyfile(may_exist_bbl, pj(project_folder, f'merge_{mode}.bbl'))
-        shutil.copyfile(may_exist_bbl, pj(project_folder, f'merge_diff.bbl'))
-
-    with open(maintex, 'r', encoding='utf-8', errors='replace') as f:
-        content = f.read()
-        merged_content = merge_tex_files(project_folder, content, mode)
-
-    with open(project_folder + '/merge.tex', 'w', encoding='utf-8', errors='replace') as f:
-        f.write(merged_content)
-
-    #  <-------- 精细切分latex文件 ----------> 
-    chatbot.append((f"Latex文件融合完成", f'[Local Message] 正在精细切分latex文件，这需要一段时间计算，文档越长耗时越长，请耐心等待。'))
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-    lps = LatexPaperSplit()
-    res = lps.split(merged_content, project_folder, opts) # 消耗时间的函数
-
-    #  <-------- 拆分过长的latex片段 ----------> 
-    pfg = LatexPaperFileGroup()
-    for index, r in enumerate(res):
-        pfg.file_paths.append('segment-' + str(index))
-        pfg.file_contents.append(r)
-
-    pfg.run_file_split(max_token_limit=1024)
-    n_split = len(pfg.sp_file_contents)
-
-    #  <-------- 根据需要切换prompt ----------> 
-    inputs_array, sys_prompt_array = switch_prompt(pfg, mode)
-    inputs_show_user_array = [f"{mode} {f}" for f in pfg.sp_file_tag]
-
-    if os.path.exists(pj(project_folder,'temp.pkl')):
-
-        #  <-------- 【仅调试】如果存在调试缓存文件，则跳过GPT请求环节 ----------> 
-        pfg = objload(file=pj(project_folder,'temp.pkl'))
-
-    else:
-        #  <-------- gpt 多线程请求 ----------> 
-        gpt_response_collection = yield from request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency(
-            inputs_array=inputs_array,
-            inputs_show_user_array=inputs_show_user_array,
-            llm_kwargs=llm_kwargs,
-            chatbot=chatbot,
-            history_array=[[""] for _ in range(n_split)],
-            sys_prompt_array=sys_prompt_array,
-            # max_workers=5,  # 并行任务数量限制, 最多同时执行5个, 其他的排队等待
-            scroller_max_len = 40
-        )
-
-        #  <-------- 文本碎片重组为完整的tex片段 ----------> 
-        pfg.sp_file_result = []
-        for i_say, gpt_say, orig_content in zip(gpt_response_collection[0::2], gpt_response_collection[1::2], pfg.sp_file_contents):
-            pfg.sp_file_result.append(gpt_say)
-        pfg.merge_result()
-
-        # <-------- 临时存储用于调试 ----------> 
-        pfg.get_token_num = None
-        objdump(pfg, file=pj(project_folder,'temp.pkl'))
-
-    write_html(pfg.sp_file_contents, pfg.sp_file_result, chatbot=chatbot, project_folder=project_folder)
-
-    #  <-------- 写出文件 ----------> 
-    msg = f"当前大语言模型: {llm_kwargs['llm_model']}，当前语言模型温度设定: {llm_kwargs['temperature']}。"
-    final_tex = lps.merge_result(pfg.file_result, mode, msg)
-    with open(project_folder + f'/merge_{mode}.tex', 'w', encoding='utf-8', errors='replace') as f:
-        if mode != 'translate_zh' or "binary" in final_tex: f.write(final_tex)
-        
-
-    #  <-------- 整理结果, 退出 ----------> 
-    chatbot.append((f"完成了吗？", 'GPT结果已输出, 正在编译PDF'))
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-
-    #  <-------- 返回 ----------> 
-    return project_folder + f'/merge_{mode}.tex'
-
-
-
-def remove_buggy_lines(file_path, log_path, tex_name, tex_name_pure, n_fix, work_folder_modified):
-    try:
-        with open(log_path, 'r', encoding='utf-8', errors='replace') as f:
-            log = f.read()
-        with open(file_path, 'r', encoding='utf-8', errors='replace') as f:
-            file_lines = f.readlines()
-        import re
-        buggy_lines = re.findall(tex_name+':([0-9]{1,5}):', log)
-        buggy_lines = [int(l) for l in buggy_lines]
-        buggy_lines = sorted(buggy_lines)
-        print("removing lines that has errors", buggy_lines)
-        file_lines.pop(buggy_lines[0]-1)
-        with open(pj(work_folder_modified, f"{tex_name_pure}_fix_{n_fix}.tex"), 'w', encoding='utf-8', errors='replace') as f:
-            f.writelines(file_lines)
-        return True, f"{tex_name_pure}_fix_{n_fix}", buggy_lines
-    except:
-        print("Fatal error occurred, but we cannot identify error, please download zip, read latex log, and compile manually.")
-        return False, -1, [-1]
-    
-def compile_latex_with_timeout(command, cwd, timeout=60):
-    import subprocess
-    process = subprocess.Popen(command, shell=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE, cwd=cwd)
-    try:
-        stdout, stderr = process.communicate(timeout=timeout)
-    except subprocess.TimeoutExpired:
-        process.kill()
-        stdout, stderr = process.communicate()
-        print("Process timed out!")
-        return False
-    return True
-
-def 编译Latex(chatbot, history, main_file_original, main_file_modified, work_folder_original, work_folder_modified, work_folder, mode='default'):
-    import os, time
-    current_dir = os.getcwd()
-    n_fix = 1
-    max_try = 32
-    chatbot.append([f"正在编译PDF文档", f'编译已经开始。当前工作路径为{work_folder}，如果程序停顿5分钟以上，请直接去该路径下取回翻译结果，或者重启之后再度尝试 ...']); yield from update_ui(chatbot=chatbot, history=history)
-    chatbot.append([f"正在编译PDF文档", '...']); yield from update_ui(chatbot=chatbot, history=history); time.sleep(1); chatbot[-1] = list(chatbot[-1]) # 刷新界面
-    yield from update_ui_lastest_msg('编译已经开始...', chatbot, history)   # 刷新Gradio前端界面
-
-    while True:
-        import os
-
-        # https://stackoverflow.com/questions/738755/dont-make-me-manually-abort-a-latex-compile-when-theres-an-error
-        yield from update_ui_lastest_msg(f'尝试第 {n_fix}/{max_try} 次编译, 编译原始PDF ...', chatbot, history)   # 刷新Gradio前端界面
-        ok = compile_latex_with_timeout(f'pdflatex -interaction=batchmode -file-line-error {main_file_original}.tex', work_folder_original)
-
-        yield from update_ui_lastest_msg(f'尝试第 {n_fix}/{max_try} 次编译, 编译转化后的PDF ...', chatbot, history)   # 刷新Gradio前端界面
-        ok = compile_latex_with_timeout(f'pdflatex -interaction=batchmode -file-line-error {main_file_modified}.tex', work_folder_modified)
-        
-        if ok and os.path.exists(pj(work_folder_modified, f'{main_file_modified}.pdf')):
-            # 只有第二步成功，才能继续下面的步骤
-            yield from update_ui_lastest_msg(f'尝试第 {n_fix}/{max_try} 次编译, 编译BibTex ...', chatbot, history)    # 刷新Gradio前端界面
-            if not os.path.exists(pj(work_folder_original, f'{main_file_original}.bbl')):
-                ok = compile_latex_with_timeout(f'bibtex  {main_file_original}.aux', work_folder_original)
-            if not os.path.exists(pj(work_folder_modified, f'{main_file_modified}.bbl')):
-                ok = compile_latex_with_timeout(f'bibtex  {main_file_modified}.aux', work_folder_modified)
-
-            yield from update_ui_lastest_msg(f'尝试第 {n_fix}/{max_try} 次编译, 编译文献交叉引用 ...', chatbot, history)  # 刷新Gradio前端界面
-            ok = compile_latex_with_timeout(f'pdflatex -interaction=batchmode -file-line-error {main_file_original}.tex', work_folder_original)
-            ok = compile_latex_with_timeout(f'pdflatex -interaction=batchmode -file-line-error {main_file_modified}.tex', work_folder_modified)
-            ok = compile_latex_with_timeout(f'pdflatex -interaction=batchmode -file-line-error {main_file_original}.tex', work_folder_original)
-            ok = compile_latex_with_timeout(f'pdflatex -interaction=batchmode -file-line-error {main_file_modified}.tex', work_folder_modified)
-
-            if mode!='translate_zh':
-                yield from update_ui_lastest_msg(f'尝试第 {n_fix}/{max_try} 次编译, 使用latexdiff生成论文转化前后对比 ...', chatbot, history) # 刷新Gradio前端界面
-                print(    f'latexdiff --encoding=utf8 --append-safecmd=subfile {work_folder_original}/{main_file_original}.tex  {work_folder_modified}/{main_file_modified}.tex --flatten > {work_folder}/merge_diff.tex')
-                ok = compile_latex_with_timeout(f'latexdiff --encoding=utf8 --append-safecmd=subfile {work_folder_original}/{main_file_original}.tex  {work_folder_modified}/{main_file_modified}.tex --flatten > {work_folder}/merge_diff.tex')
-
-                yield from update_ui_lastest_msg(f'尝试第 {n_fix}/{max_try} 次编译, 正在编译对比PDF ...', chatbot, history)   # 刷新Gradio前端界面
-                ok = compile_latex_with_timeout(f'pdflatex  -interaction=batchmode -file-line-error merge_diff.tex', work_folder)
-                ok = compile_latex_with_timeout(f'bibtex    merge_diff.aux', work_folder)
-                ok = compile_latex_with_timeout(f'pdflatex  -interaction=batchmode -file-line-error merge_diff.tex', work_folder)
-                ok = compile_latex_with_timeout(f'pdflatex  -interaction=batchmode -file-line-error merge_diff.tex', work_folder)
-
-
-        # <---------- 检查结果 ----------->
-        results_ = ""
-        original_pdf_success = os.path.exists(pj(work_folder_original, f'{main_file_original}.pdf'))
-        modified_pdf_success = os.path.exists(pj(work_folder_modified, f'{main_file_modified}.pdf'))
-        diff_pdf_success     = os.path.exists(pj(work_folder, f'merge_diff.pdf'))
-        results_ += f"原始PDF编译是否成功: {original_pdf_success};" 
-        results_ += f"转化PDF编译是否成功: {modified_pdf_success};" 
-        results_ += f"对比PDF编译是否成功: {diff_pdf_success};" 
-        yield from update_ui_lastest_msg(f'第{n_fix}编译结束:<br/>{results_}...', chatbot, history) # 刷新Gradio前端界面
-
-        if diff_pdf_success:
-            result_pdf = pj(work_folder_modified, f'merge_diff.pdf')    # get pdf path
-            promote_file_to_downloadzone(result_pdf, rename_file=None, chatbot=chatbot)  # promote file to web UI
-        if modified_pdf_success:
-            yield from update_ui_lastest_msg(f'转化PDF编译已经成功, 即将退出 ...', chatbot, history)    # 刷新Gradio前端界面
-            result_pdf = pj(work_folder_modified, f'{main_file_modified}.pdf') # get pdf path
-            if os.path.exists(pj(work_folder, '..', 'translation')):
-                shutil.copyfile(result_pdf, pj(work_folder, '..', 'translation', 'translate_zh.pdf'))
-            promote_file_to_downloadzone(result_pdf, rename_file=None, chatbot=chatbot)  # promote file to web UI
-            return True # 成功啦
-        else:
-            if n_fix>=max_try: break
-            n_fix += 1
-            can_retry, main_file_modified, buggy_lines = remove_buggy_lines(
-                file_path=pj(work_folder_modified, f'{main_file_modified}.tex'), 
-                log_path=pj(work_folder_modified, f'{main_file_modified}.log'),
-                tex_name=f'{main_file_modified}.tex',
-                tex_name_pure=f'{main_file_modified}',
-                n_fix=n_fix,
-                work_folder_modified=work_folder_modified,
-            )
-            yield from update_ui_lastest_msg(f'由于最为关键的转化PDF编译失败, 将根据报错信息修正tex源文件并重试, 当前报错的latex代码处于第{buggy_lines}行 ...', chatbot, history)   # 刷新Gradio前端界面
-            if not can_retry: break
-
-    return False # 失败啦
-
-
-

crazy_functions/live_audio/aliyunASR.py

@@ -85,8 +85,8 @@ def write_numpy_to_wave(filename, rate, data, add_header=False):
 
 def is_speaker_speaking(vad, data, sample_rate):
     # Function to detect if the speaker is speaking
-    # The WebRTC VAD only accepts 16-bit mono PCM audio, 
-    # sampled at 8000, 16000, 32000 or 48000 Hz. 
+    # The WebRTC VAD only accepts 16-bit mono PCM audio,
+    # sampled at 8000, 16000, 32000 or 48000 Hz.
     # A frame must be either 10, 20, or 30 ms in duration:
     frame_duration = 30
     n_bit_each = int(sample_rate * frame_duration / 1000)*2 # x2 because audio is 16 bit (2 bytes)
@@ -94,7 +94,7 @@ def is_speaker_speaking(vad, data, sample_rate):
     for t in range(len(data)):
         if t!=0 and t % n_bit_each == 0:
             res_list.append(vad.is_speech(data[t-n_bit_each:t], sample_rate))
-    
+
     info = ''.join(['^' if r else '.' for r in res_list])
     info = info[:10]
     if any(res_list):
@@ -186,10 +186,10 @@ class AliyunASR():
         keep_alive_last_send_time = time.time()
         while not self.stop:
             # time.sleep(self.capture_interval)
-            audio = rad.read(uuid.hex) 
+            audio = rad.read(uuid.hex)
             if audio is not None:
                 # convert to pcm file
-                temp_file = f'{temp_folder}/{uuid.hex}.pcm' # 
+                temp_file = f'{temp_folder}/{uuid.hex}.pcm' #
                 dsdata = change_sample_rate(audio, rad.rate, NEW_SAMPLERATE) # 48000 --> 16000
                 write_numpy_to_wave(temp_file, NEW_SAMPLERATE, dsdata)
                 # read pcm binary

crazy_functions/live_audio/audio_io.py

@@ -3,12 +3,12 @@ from scipy import interpolate
 
 def Singleton(cls):
     _instance = {}
- 
+
     def _singleton(*args, **kargs):
         if cls not in _instance:
             _instance[cls] = cls(*args, **kargs)
         return _instance[cls]
- 
+
     return _singleton
 
 
@@ -39,7 +39,7 @@ class RealtimeAudioDistribution():
         else:
             res = None
         return res
-    
+
 def change_sample_rate(audio, old_sr, new_sr):
     duration = audio.shape[0] / old_sr
 

crazy_functions/multi_stage/multi_stage_utils.py

@@ -40,7 +40,7 @@ class GptAcademicState():
 
 class GptAcademicGameBaseState():
     """
-    1. first init: __init__ -> 
+    1. first init: __init__ ->
     """
     def init_game(self, chatbot, lock_plugin):
         self.plugin_name = None
@@ -53,7 +53,7 @@ class GptAcademicGameBaseState():
             raise ValueError("callback_fn is None")
         chatbot._cookies['lock_plugin'] = self.callback_fn
         self.dump_state(chatbot)
-        
+
     def get_plugin_name(self):
         if self.plugin_name is None:
             raise ValueError("plugin_name is None")
@@ -71,7 +71,7 @@ class GptAcademicGameBaseState():
         state = chatbot._cookies.get(f'plugin_state/{plugin_name}', None)
         if state is not None:
             state = pickle.loads(state)
-        else: 
+        else:
             state = cls()
             state.init_game(chatbot, lock_plugin)
         state.plugin_name = plugin_name
@@ -79,7 +79,7 @@ class GptAcademicGameBaseState():
         state.chatbot = chatbot
         state.callback_fn = callback_fn
         return state
-    
+
     def continue_game(self, prompt, chatbot, history):
         # 游戏主体
         yield from self.step(prompt, chatbot, history)

crazy_functions/pdf_fns/breakdown_txt.py

@@ -35,7 +35,7 @@ def cut(limit, get_token_fn, txt_tocut, must_break_at_empty_line, break_anyway=F
     remain_txt_to_cut_storage = ""
     # 为了加速计算，我们采样一个特殊的手段。当 remain_txt_to_cut > `_max` 时， 我们把 _max 后的文字转存至 remain_txt_to_cut_storage
     remain_txt_to_cut, remain_txt_to_cut_storage = maintain_storage(remain_txt_to_cut, remain_txt_to_cut_storage)
-    
+
     while True:
         if get_token_fn(remain_txt_to_cut) <= limit:
             # 如果剩余文本的token数小于限制，那么就不用切了

crazy_functions/pdf_fns/parse_pdf.py

@@ -64,8 +64,8 @@ def produce_report_markdown(gpt_response_collection, meta, paper_meta_info, chat
             # 再做一个小修改：重新修改当前part的标题，默认用英文的
             cur_value += value
             translated_res_array.append(cur_value)
-    res_path = write_history_to_file(meta +  ["# Meta Translation" , paper_meta_info] + translated_res_array, 
-                                     file_basename = f"{gen_time_str()}-translated_only.md", 
+    res_path = write_history_to_file(meta +  ["# Meta Translation" , paper_meta_info] + translated_res_array,
+                                     file_basename = f"{gen_time_str()}-translated_only.md",
                                      file_fullname = None,
                                      auto_caption = False)
     promote_file_to_downloadzone(res_path, rename_file=os.path.basename(res_path)+'.md', chatbot=chatbot)
@@ -144,11 +144,11 @@ def translate_pdf(article_dict, llm_kwargs, chatbot, fp, generated_conclusion_fi
     produce_report_markdown(gpt_response_collection, meta, paper_meta_info, chatbot, fp, generated_conclusion_files)
 
     # -=-=-=-=-=-=-=-= 写出HTML文件 -=-=-=-=-=-=-=-=
-    ch = construct_html() 
+    ch = construct_html()
     orig = ""
     trans = ""
     gpt_response_collection_html = copy.deepcopy(gpt_response_collection)
-    for i,k in enumerate(gpt_response_collection_html): 
+    for i,k in enumerate(gpt_response_collection_html):
         if i%2==0:
             gpt_response_collection_html[i] = inputs_show_user_array[i//2]
         else:
@@ -159,7 +159,7 @@ def translate_pdf(article_dict, llm_kwargs, chatbot, fp, generated_conclusion_fi
 
     final = ["", "", "一、论文概况",  "", "Abstract", paper_meta_info,  "二、论文翻译",  ""]
     final.extend(gpt_response_collection_html)
-    for i, k in enumerate(final): 
+    for i, k in enumerate(final):
         if i%2==0:
             orig = k
         if i%2==1:

crazy_functions/pdf_fns/parse_word.py

@@ -22,10 +22,10 @@ def extract_text_from_files(txt, chatbot, history):
     file_manifest = []
     excption = ""
 
-    if txt == "": 
+    if txt == "":
         final_result.append(txt)
         return False, final_result, page_one, file_manifest, excption   #如输入区内容不是文件则直接返回输入区内容
-    
+
     #查找输入区内容中的文件
     file_pdf,pdf_manifest,folder_pdf = get_files_from_everything(txt, '.pdf')
     file_md,md_manifest,folder_md = get_files_from_everything(txt, '.md')
@@ -35,12 +35,12 @@ def extract_text_from_files(txt, chatbot, history):
     if file_doc:
         excption = "word"
         return False, final_result, page_one, file_manifest, excption
-    
+
     file_num = len(pdf_manifest) + len(md_manifest) + len(word_manifest)
     if file_num == 0:
         final_result.append(txt)
         return False, final_result, page_one, file_manifest, excption   #如输入区内容不是文件则直接返回输入区内容
-    
+
     if file_pdf:
         try:    # 尝试导入依赖，如果缺少依赖，则给出安装建议
             import fitz
@@ -61,7 +61,7 @@ def extract_text_from_files(txt, chatbot, history):
                 file_content = f.read()
             file_content = file_content.encode('utf-8', 'ignore').decode()
             headers = re.findall(r'^#\s(.*)$', file_content, re.MULTILINE)  #接下来提取md中的一级/二级标题作为摘要
-            if len(headers) > 0: 
+            if len(headers) > 0:
                 page_one.append("\n".join(headers)) #合并所有的标题,以换行符分割
             else:
                 page_one.append("")
@@ -81,5 +81,5 @@ def extract_text_from_files(txt, chatbot, history):
             page_one.append(file_content[:200])
             final_result.append(file_content)
             file_manifest.append(os.path.relpath(fp, folder_word))
-            
+
     return True, final_result, page_one, file_manifest, excption

crazy_functions/test_project/cpp/cppipc/buffer.cpp

@@ -1,87 +0,0 @@
-#include "libipc/buffer.h"
-#include "libipc/utility/pimpl.h"
-
-#include <cstring>
-
-namespace ipc {
-
-bool operator==(buffer const & b1, buffer const & b2) {
-    return (b1.size() == b2.size()) && (std::memcmp(b1.data(), b2.data(), b1.size()) == 0);
-}
-
-bool operator!=(buffer const & b1, buffer const & b2) {
-    return !(b1 == b2);
-}
-
-class buffer::buffer_ : public pimpl<buffer_> {
-public:
-    void*       p_;
-    std::size_t s_;
-    void*       a_;
-    buffer::destructor_t d_;
-
-    buffer_(void* p, std::size_t s, buffer::destructor_t d, void* a)
-        : p_(p), s_(s), a_(a), d_(d) {
-    }
-
-    ~buffer_() {
-        if (d_ == nullptr) return;
-        d_((a_ == nullptr) ? p_ : a_, s_);
-    }
-};
-
-buffer::buffer()
-    : buffer(nullptr, 0, nullptr, nullptr) {
-}
-
-buffer::buffer(void* p, std::size_t s, destructor_t d)
-    : p_(p_->make(p, s, d, nullptr)) {
-}
-
-buffer::buffer(void* p, std::size_t s, destructor_t d, void* additional)
-    : p_(p_->make(p, s, d, additional)) {
-}
-
-buffer::buffer(void* p, std::size_t s)
-    : buffer(p, s, nullptr) {
-}
-
-buffer::buffer(char const & c)
-    : buffer(const_cast<char*>(&c), 1) {
-}
-
-buffer::buffer(buffer&& rhs)
-    : buffer() {
-    swap(rhs);
-}
-
-buffer::~buffer() {
-    p_->clear();
-}
-
-void buffer::swap(buffer& rhs) {
-    std::swap(p_, rhs.p_);
-}
-
-buffer& buffer::operator=(buffer rhs) {
-    swap(rhs);
-    return *this;
-}
-
-bool buffer::empty() const noexcept {
-    return (impl(p_)->p_ == nullptr) || (impl(p_)->s_ == 0);
-}
-
-void* buffer::data() noexcept {
-    return impl(p_)->p_;
-}
-
-void const * buffer::data() const noexcept {
-    return impl(p_)->p_;
-}
-
-std::size_t buffer::size() const noexcept {
-    return impl(p_)->s_;
-}
-
-} // namespace ipc

crazy_functions/test_project/cpp/cppipc/ipc.cpp

@@ -1,701 +0,0 @@
-
-#include <type_traits>
-#include <cstring>
-#include <algorithm>
-#include <utility>          // std::pair, std::move, std::forward
-#include <atomic>
-#include <type_traits>      // aligned_storage_t
-#include <string>
-#include <vector>
-#include <array>
-#include <cassert>
-
-#include "libipc/ipc.h"
-#include "libipc/def.h"
-#include "libipc/shm.h"
-#include "libipc/pool_alloc.h"
-#include "libipc/queue.h"
-#include "libipc/policy.h"
-#include "libipc/rw_lock.h"
-#include "libipc/waiter.h"
-
-#include "libipc/utility/log.h"
-#include "libipc/utility/id_pool.h"
-#include "libipc/utility/scope_guard.h"
-#include "libipc/utility/utility.h"
-
-#include "libipc/memory/resource.h"
-#include "libipc/platform/detail.h"
-#include "libipc/circ/elem_array.h"
-
-namespace {
-
-using msg_id_t = std::uint32_t;
-using acc_t    = std::atomic<msg_id_t>;
-
-template <std::size_t DataSize, std::size_t AlignSize>
-struct msg_t;
-
-template <std::size_t AlignSize>
-struct msg_t<0, AlignSize> {
-    msg_id_t     cc_id_;
-    msg_id_t     id_;
-    std::int32_t remain_;
-    bool         storage_;
-};
-
-template <std::size_t DataSize, std::size_t AlignSize>
-struct msg_t : msg_t<0, AlignSize> {
-    std::aligned_storage_t<DataSize, AlignSize> data_ {};
-
-    msg_t() = default;
-    msg_t(msg_id_t cc_id, msg_id_t id, std::int32_t remain, void const * data, std::size_t size)
-        : msg_t<0, AlignSize> {cc_id, id, remain, (data == nullptr) || (size == 0)} {
-        if (this->storage_) {
-            if (data != nullptr) {
-                // copy storage-id
-                *reinterpret_cast<ipc::storage_id_t*>(&data_) =
-                     *static_cast<ipc::storage_id_t const *>(data);
-            }
-        }
-        else std::memcpy(&data_, data, size);
-    }
-};
-
-template <typename T>
-ipc::buff_t make_cache(T& data, std::size_t size) {
-    auto ptr = ipc::mem::alloc(size);
-    std::memcpy(ptr, &data, (ipc::detail::min)(sizeof(data), size));
-    return { ptr, size, ipc::mem::free };
-}
-
-struct cache_t {
-    std::size_t fill_;
-    ipc::buff_t buff_;
-
-    cache_t(std::size_t f, ipc::buff_t && b)
-        : fill_(f), buff_(std::move(b))
-    {}
-
-    void append(void const * data, std::size_t size) {
-        if (fill_ >= buff_.size() || data == nullptr || size == 0) return;
-        auto new_fill = (ipc::detail::min)(fill_ + size, buff_.size());
-        std::memcpy(static_cast<ipc::byte_t*>(buff_.data()) + fill_, data, new_fill - fill_);
-        fill_ = new_fill;
-    }
-};
-
-auto cc_acc() {
-    static ipc::shm::handle acc_h("__CA_CONN__", sizeof(acc_t));
-    return static_cast<acc_t*>(acc_h.get());
-}
-
-IPC_CONSTEXPR_ std::size_t align_chunk_size(std::size_t size) noexcept {
-    return (((size - 1) / ipc::large_msg_align) + 1) * ipc::large_msg_align;
-}
-
-IPC_CONSTEXPR_ std::size_t calc_chunk_size(std::size_t size) noexcept {
-    return ipc::make_align(alignof(std::max_align_t), align_chunk_size(
-           ipc::make_align(alignof(std::max_align_t), sizeof(std::atomic<ipc::circ::cc_t>)) + size));
-}
-
-struct chunk_t {
-    std::atomic<ipc::circ::cc_t> &conns() noexcept {
-        return *reinterpret_cast<std::atomic<ipc::circ::cc_t> *>(this);
-    }
-
-    void *data() noexcept {
-        return reinterpret_cast<ipc::byte_t *>(this)
-             + ipc::make_align(alignof(std::max_align_t), sizeof(std::atomic<ipc::circ::cc_t>));
-    }
-};
-
-struct chunk_info_t {
-    ipc::id_pool<> pool_;
-    ipc::spin_lock lock_;
-
-    IPC_CONSTEXPR_ static std::size_t chunks_mem_size(std::size_t chunk_size) noexcept {
-        return ipc::id_pool<>::max_count * chunk_size;
-    }
-
-    ipc::byte_t *chunks_mem() noexcept {
-        return reinterpret_cast<ipc::byte_t *>(this + 1);
-    }
-
-    chunk_t *at(std::size_t chunk_size, ipc::storage_id_t id) noexcept {
-        if (id < 0) return nullptr;
-        return reinterpret_cast<chunk_t *>(chunks_mem() + (chunk_size * id));
-    }
-};
-
-auto& chunk_storages() {
-    class chunk_handle_t {
-        ipc::shm::handle handle_;
-
-    public:
-        chunk_info_t *get_info(std::size_t chunk_size) {
-            if (!handle_.valid() &&
-                !handle_.acquire( ("__CHUNK_INFO__" + ipc::to_string(chunk_size)).c_str(), 
-                                  sizeof(chunk_info_t) + chunk_info_t::chunks_mem_size(chunk_size) )) {
-                ipc::error("[chunk_storages] chunk_shm.id_info_.acquire failed: chunk_size = %zd\n", chunk_size);
-                return nullptr;
-            }
-            auto info = static_cast<chunk_info_t*>(handle_.get());
-            if (info == nullptr) {
-                ipc::error("[chunk_storages] chunk_shm.id_info_.get failed: chunk_size = %zd\n", chunk_size);
-                return nullptr;
-            }
-            return info;
-        }
-    };
-    static ipc::map<std::size_t, chunk_handle_t> chunk_hs;
-    return chunk_hs;
-}
-
-chunk_info_t *chunk_storage_info(std::size_t chunk_size) {
-    auto &storages = chunk_storages();
-    std::decay_t<decltype(storages)>::iterator it;
-    {
-        static ipc::rw_lock lock;
-        IPC_UNUSED_ std::shared_lock<ipc::rw_lock> guard {lock};
-        if ((it = storages.find(chunk_size)) == storages.end()) {
-            using chunk_handle_t = std::decay_t<decltype(storages)>::value_type::second_type;
-            guard.unlock();
-            IPC_UNUSED_ std::lock_guard<ipc::rw_lock> guard {lock};
-            it = storages.emplace(chunk_size, chunk_handle_t{}).first;
-        }
-    }
-    return it->second.get_info(chunk_size);
-}
-
-std::pair<ipc::storage_id_t, void*> acquire_storage(std::size_t size, ipc::circ::cc_t conns) {
-    std::size_t chunk_size = calc_chunk_size(size);
-    auto info = chunk_storage_info(chunk_size);
-    if (info == nullptr) return {};
-
-    info->lock_.lock();
-    info->pool_.prepare();
-    // got an unique id
-    auto id = info->pool_.acquire();
-    info->lock_.unlock();
-
-    auto chunk = info->at(chunk_size, id);
-    if (chunk == nullptr) return {};
-    chunk->conns().store(conns, std::memory_order_relaxed);
-    return { id, chunk->data() };
-}
-
-void *find_storage(ipc::storage_id_t id, std::size_t size) {
-    if (id < 0) {
-        ipc::error("[find_storage] id is invalid: id = %ld, size = %zd\n", (long)id, size);
-        return nullptr;
-    }
-    std::size_t chunk_size = calc_chunk_size(size);
-    auto info = chunk_storage_info(chunk_size);
-    if (info == nullptr) return nullptr;
-    return info->at(chunk_size, id)->data();
-}
-
-void release_storage(ipc::storage_id_t id, std::size_t size) {
-    if (id < 0) {
-        ipc::error("[release_storage] id is invalid: id = %ld, size = %zd\n", (long)id, size);
-        return;
-    }
-    std::size_t chunk_size = calc_chunk_size(size);
-    auto info = chunk_storage_info(chunk_size);
-    if (info == nullptr) return;
-    info->lock_.lock();
-    info->pool_.release(id);
-    info->lock_.unlock();
-}
-
-template <ipc::relat Rp, ipc::relat Rc>
-bool sub_rc(ipc::wr<Rp, Rc, ipc::trans::unicast>, 
-            std::atomic<ipc::circ::cc_t> &/*conns*/, ipc::circ::cc_t /*curr_conns*/, ipc::circ::cc_t /*conn_id*/) noexcept {
-    return true;
-}
-
-template <ipc::relat Rp, ipc::relat Rc>
-bool sub_rc(ipc::wr<Rp, Rc, ipc::trans::broadcast>, 
-            std::atomic<ipc::circ::cc_t> &conns, ipc::circ::cc_t curr_conns, ipc::circ::cc_t conn_id) noexcept {
-    auto last_conns = curr_conns & ~conn_id;
-    for (unsigned k = 0;;) {
-        auto chunk_conns  = conns.load(std::memory_order_acquire);
-        if (conns.compare_exchange_weak(chunk_conns, chunk_conns & last_conns, std::memory_order_release)) {
-            return (chunk_conns & last_conns) == 0;
-        }
-        ipc::yield(k);
-    }
-}
-
-template <typename Flag>
-void recycle_storage(ipc::storage_id_t id, std::size_t size, ipc::circ::cc_t curr_conns, ipc::circ::cc_t conn_id) {
-    if (id < 0) {
-        ipc::error("[recycle_storage] id is invalid: id = %ld, size = %zd\n", (long)id, size);
-        return;
-    }
-    std::size_t chunk_size = calc_chunk_size(size);
-    auto info = chunk_storage_info(chunk_size);
-    if (info == nullptr) return;
-
-    auto chunk = info->at(chunk_size, id);
-    if (chunk == nullptr) return;
-
-    if (!sub_rc(Flag{}, chunk->conns(), curr_conns, conn_id)) {
-        return;
-    }
-    info->lock_.lock();
-    info->pool_.release(id);
-    info->lock_.unlock();
-}
-
-template <typename MsgT>
-bool clear_message(void* p) {
-    auto msg = static_cast<MsgT*>(p);
-    if (msg->storage_) {
-        std::int32_t r_size = static_cast<std::int32_t>(ipc::data_length) + msg->remain_;
-        if (r_size <= 0) {
-            ipc::error("[clear_message] invalid msg size: %d\n", (int)r_size);
-            return true;
-        }
-        release_storage(
-            *reinterpret_cast<ipc::storage_id_t*>(&msg->data_),
-            static_cast<std::size_t>(r_size));
-    }
-    return true;
-}
-
-struct conn_info_head {
-
-    ipc::string name_;
-    msg_id_t    cc_id_; // connection-info id
-    ipc::detail::waiter cc_waiter_, wt_waiter_, rd_waiter_;
-    ipc::shm::handle acc_h_;
-
-    conn_info_head(char const * name)
-        : name_     {name}
-        , cc_id_    {(cc_acc() == nullptr) ? 0 : cc_acc()->fetch_add(1, std::memory_order_relaxed)}
-        , cc_waiter_{("__CC_CONN__" + name_).c_str()}
-        , wt_waiter_{("__WT_CONN__" + name_).c_str()}
-        , rd_waiter_{("__RD_CONN__" + name_).c_str()}
-        , acc_h_    {("__AC_CONN__" + name_).c_str(), sizeof(acc_t)} {
-    }
-
-    void quit_waiting() {
-        cc_waiter_.quit_waiting();
-        wt_waiter_.quit_waiting();
-        rd_waiter_.quit_waiting();
-    }
-
-    auto acc() {
-        return static_cast<acc_t*>(acc_h_.get());
-    }
-
-    auto& recv_cache() {
-        thread_local ipc::unordered_map<msg_id_t, cache_t> tls;
-        return tls;
-    }
-};
-
-template <typename W, typename F>
-bool wait_for(W& waiter, F&& pred, std::uint64_t tm) {
-    if (tm == 0) return !pred();
-    for (unsigned k = 0; pred();) {
-        bool ret = true;
-        ipc::sleep(k, [&k, &ret, &waiter, &pred, tm] {
-            ret = waiter.wait_if(std::forward<F>(pred), tm);
-            k   = 0;
-        });
-        if (!ret) return false; // timeout or fail
-        if (k == 0) break; // k has been reset
-    }
-    return true;
-}
-
-template <typename Policy,
-          std::size_t DataSize  = ipc::data_length,
-          std::size_t AlignSize = (ipc::detail::min)(DataSize, alignof(std::max_align_t))>
-struct queue_generator {
-
-    using queue_t = ipc::queue<msg_t<DataSize, AlignSize>, Policy>;
-
-    struct conn_info_t : conn_info_head {
-        queue_t que_;
-
-        conn_info_t(char const * name)
-            : conn_info_head{name}
-            , que_{("__QU_CONN__" +
-                    ipc::to_string(DataSize) + "__" +
-                    ipc::to_string(AlignSize) + "__" + name).c_str()} {
-        }
-
-        void disconnect_receiver() {
-            bool dis = que_.disconnect();
-            this->quit_waiting();
-            if (dis) {
-                this->recv_cache().clear();
-            }
-        }
-    };
-};
-
-template <typename Policy>
-struct detail_impl {
-
-using policy_t    = Policy;
-using flag_t      = typename policy_t::flag_t;
-using queue_t     = typename queue_generator<policy_t>::queue_t;
-using conn_info_t = typename queue_generator<policy_t>::conn_info_t;
-
-constexpr static conn_info_t* info_of(ipc::handle_t h) noexcept {
-    return static_cast<conn_info_t*>(h);
-}
-
-constexpr static queue_t* queue_of(ipc::handle_t h) noexcept {
-    return (info_of(h) == nullptr) ? nullptr : &(info_of(h)->que_);
-}
-
-/* API implementations */
-
-static void disconnect(ipc::handle_t h) {
-    auto que = queue_of(h);
-    if (que == nullptr) {
-        return;
-    }
-    que->shut_sending();
-    assert(info_of(h) != nullptr);
-    info_of(h)->disconnect_receiver();
-}
-
-static bool reconnect(ipc::handle_t * ph, bool start_to_recv) {
-    assert(ph != nullptr);
-    assert(*ph != nullptr);
-    auto que = queue_of(*ph);
-    if (que == nullptr) {
-        return false;
-    }
-    if (start_to_recv) {
-        que->shut_sending();
-        if (que->connect()) { // wouldn't connect twice
-            info_of(*ph)->cc_waiter_.broadcast();
-            return true;
-        }
-        return false;
-    }
-    // start_to_recv == false
-    if (que->connected()) {
-        info_of(*ph)->disconnect_receiver();
-    }
-    return que->ready_sending();
-}
-
-static bool connect(ipc::handle_t * ph, char const * name, bool start_to_recv) {
-    assert(ph != nullptr);
-    if (*ph == nullptr) {
-        *ph = ipc::mem::alloc<conn_info_t>(name);
-    }
-    return reconnect(ph, start_to_recv);
-}
-
-static void destroy(ipc::handle_t h) {
-    disconnect(h);
-    ipc::mem::free(info_of(h));
-}
-
-static std::size_t recv_count(ipc::handle_t h) noexcept {
-    auto que = queue_of(h);
-    if (que == nullptr) {
-        return ipc::invalid_value;
-    }
-    return que->conn_count();
-}
-
-static bool wait_for_recv(ipc::handle_t h, std::size_t r_count, std::uint64_t tm) {
-    auto que = queue_of(h);
-    if (que == nullptr) {
-        return false;
-    }
-    return wait_for(info_of(h)->cc_waiter_, [que, r_count] {
-        return que->conn_count() < r_count;
-    }, tm);
-}
-
-template <typename F>
-static bool send(F&& gen_push, ipc::handle_t h, void const * data, std::size_t size) {
-    if (data == nullptr || size == 0) {
-        ipc::error("fail: send(%p, %zd)\n", data, size);
-        return false;
-    }
-    auto que = queue_of(h);
-    if (que == nullptr) {
-        ipc::error("fail: send, queue_of(h) == nullptr\n");
-        return false;
-    }
-    if (que->elems() == nullptr) {
-        ipc::error("fail: send, queue_of(h)->elems() == nullptr\n");
-        return false;
-    }
-    if (!que->ready_sending()) {
-        ipc::error("fail: send, que->ready_sending() == false\n");
-        return false;
-    }
-    ipc::circ::cc_t conns = que->elems()->connections(std::memory_order_relaxed);
-    if (conns == 0) {
-        ipc::error("fail: send, there is no receiver on this connection.\n");
-        return false;
-    }
-    // calc a new message id
-    auto acc = info_of(h)->acc();
-    if (acc == nullptr) {
-        ipc::error("fail: send, info_of(h)->acc() == nullptr\n");
-        return false;
-    }
-    auto msg_id   = acc->fetch_add(1, std::memory_order_relaxed);
-    auto try_push = std::forward<F>(gen_push)(info_of(h), que, msg_id);
-    if (size > ipc::large_msg_limit) {
-        auto   dat = acquire_storage(size, conns);
-        void * buf = dat.second;
-        if (buf != nullptr) {
-            std::memcpy(buf, data, size);
-            return try_push(static_cast<std::int32_t>(size) - 
-                            static_cast<std::int32_t>(ipc::data_length), &(dat.first), 0);
-        }
-        // try using message fragment
-        //ipc::log("fail: shm::handle for big message. msg_id: %zd, size: %zd\n", msg_id, size);
-    }
-    // push message fragment
-    std::int32_t offset = 0;
-    for (std::int32_t i = 0; i < static_cast<std::int32_t>(size / ipc::data_length); ++i, offset += ipc::data_length) {
-        if (!try_push(static_cast<std::int32_t>(size) - offset - static_cast<std::int32_t>(ipc::data_length),
-                      static_cast<ipc::byte_t const *>(data) + offset, ipc::data_length)) {
-            return false;
-        }
-    }
-    // if remain > 0, this is the last message fragment
-    std::int32_t remain = static_cast<std::int32_t>(size) - offset;
-    if (remain > 0) {
-        if (!try_push(remain - static_cast<std::int32_t>(ipc::data_length),
-                      static_cast<ipc::byte_t const *>(data) + offset, 
-                      static_cast<std::size_t>(remain))) {
-            return false;
-        }
-    }
-    return true;
-}
-
-static bool send(ipc::handle_t h, void const * data, std::size_t size, std::uint64_t tm) {
-    return send([tm](auto info, auto que, auto msg_id) {
-        return [tm, info, que, msg_id](std::int32_t remain, void const * data, std::size_t size) {
-            if (!wait_for(info->wt_waiter_, [&] {
-                    return !que->push(
-                        [](void*) { return true; },
-                        info->cc_id_, msg_id, remain, data, size);
-                }, tm)) {
-                ipc::log("force_push: msg_id = %zd, remain = %d, size = %zd\n", msg_id, remain, size);
-                if (!que->force_push(
-                        clear_message<typename queue_t::value_t>,
-                        info->cc_id_, msg_id, remain, data, size)) {
-                    return false;
-                }
-            }
-            info->rd_waiter_.broadcast();
-            return true;
-        };
-    }, h, data, size);
-}
-
-static bool try_send(ipc::handle_t h, void const * data, std::size_t size, std::uint64_t tm) {
-    return send([tm](auto info, auto que, auto msg_id) {
-        return [tm, info, que, msg_id](std::int32_t remain, void const * data, std::size_t size) {
-            if (!wait_for(info->wt_waiter_, [&] {
-                    return !que->push(
-                        [](void*) { return true; },
-                        info->cc_id_, msg_id, remain, data, size);
-                }, tm)) {
-                return false;
-            }
-            info->rd_waiter_.broadcast();
-            return true;
-        };
-    }, h, data, size);
-}
-
-static ipc::buff_t recv(ipc::handle_t h, std::uint64_t tm) {
-    auto que = queue_of(h);
-    if (que == nullptr) {
-        ipc::error("fail: recv, queue_of(h) == nullptr\n");
-        return {};
-    }
-    if (!que->connected()) {
-        // hasn't connected yet, just return.
-        return {};
-    }
-    auto& rc = info_of(h)->recv_cache();
-    for (;;) {
-        // pop a new message
-        typename queue_t::value_t msg;
-        if (!wait_for(info_of(h)->rd_waiter_, [que, &msg] {
-                return !que->pop(msg);
-            }, tm)) {
-            // pop failed, just return.
-            return {};
-        }
-        info_of(h)->wt_waiter_.broadcast();
-        if ((info_of(h)->acc() != nullptr) && (msg.cc_id_ == info_of(h)->cc_id_)) {
-            continue; // ignore message to self
-        }
-        // msg.remain_ may minus & abs(msg.remain_) < data_length
-        std::int32_t r_size = static_cast<std::int32_t>(ipc::data_length) + msg.remain_;
-        if (r_size <= 0) {
-            ipc::error("fail: recv, r_size = %d\n", (int)r_size);
-            return {};
-        }
-        std::size_t msg_size = static_cast<std::size_t>(r_size);
-        // large message
-        if (msg.storage_) {
-            ipc::storage_id_t buf_id = *reinterpret_cast<ipc::storage_id_t*>(&msg.data_);
-            void* buf = find_storage(buf_id, msg_size);
-            if (buf != nullptr) {
-                struct recycle_t {
-                    ipc::storage_id_t storage_id;
-                    ipc::circ::cc_t   curr_conns;
-                    ipc::circ::cc_t   conn_id;
-                } *r_info = ipc::mem::alloc<recycle_t>(recycle_t{
-                    buf_id, que->elems()->connections(std::memory_order_relaxed), que->connected_id()
-                });
-                if (r_info == nullptr) {
-                    ipc::log("fail: ipc::mem::alloc<recycle_t>.\n");
-                    return ipc::buff_t{buf, msg_size}; // no recycle
-                } else {
-                    return ipc::buff_t{buf, msg_size, [](void* p_info, std::size_t size) {
-                        auto r_info = static_cast<recycle_t *>(p_info);
-                        IPC_UNUSED_ auto finally = ipc::guard([r_info] {
-                            ipc::mem::free(r_info);
-                        });
-                        recycle_storage<flag_t>(r_info->storage_id, size, r_info->curr_conns, r_info->conn_id);
-                    }, r_info};
-                }
-            } else {
-                ipc::log("fail: shm::handle for large message. msg_id: %zd, buf_id: %zd, size: %zd\n", msg.id_, buf_id, msg_size);
-                continue;
-            }
-        }
-        // find cache with msg.id_
-        auto cac_it = rc.find(msg.id_);
-        if (cac_it == rc.end()) {
-            if (msg_size <= ipc::data_length) {
-                return make_cache(msg.data_, msg_size);
-            }
-            // gc
-            if (rc.size() > 1024) {
-                std::vector<msg_id_t> need_del;
-                for (auto const & pair : rc) {
-                    auto cmp = std::minmax(msg.id_, pair.first);
-                    if (cmp.second - cmp.first > 8192) {
-                        need_del.push_back(pair.first);
-                    }
-                }
-                for (auto id : need_del) rc.erase(id);
-            }
-            // cache the first message fragment
-            rc.emplace(msg.id_, cache_t { ipc::data_length, make_cache(msg.data_, msg_size) });
-        }
-        // has cached before this message
-        else {
-            auto& cac = cac_it->second;
-            // this is the last message fragment
-            if (msg.remain_ <= 0) {
-                cac.append(&(msg.data_), msg_size);
-                // finish this message, erase it from cache
-                auto buff = std::move(cac.buff_);
-                rc.erase(cac_it);
-                return buff;
-            }
-            // there are remain datas after this message
-            cac.append(&(msg.data_), ipc::data_length);
-        }
-    }
-}
-
-static ipc::buff_t try_recv(ipc::handle_t h) {
-    return recv(h, 0);
-}
-
-}; // detail_impl<Policy>
-
-template <typename Flag>
-using policy_t = ipc::policy::choose<ipc::circ::elem_array, Flag>;
-
-} // internal-linkage
-
-namespace ipc {
-
-template <typename Flag>
-ipc::handle_t chan_impl<Flag>::inited() {
-    ipc::detail::waiter::init();
-    return nullptr;
-}
-
-template <typename Flag>
-bool chan_impl<Flag>::connect(ipc::handle_t * ph, char const * name, unsigned mode) {
-    return detail_impl<policy_t<Flag>>::connect(ph, name, mode & receiver);
-}
-
-template <typename Flag>
-bool chan_impl<Flag>::reconnect(ipc::handle_t * ph, unsigned mode) {
-    return detail_impl<policy_t<Flag>>::reconnect(ph, mode & receiver);
-}
-
-template <typename Flag>
-void chan_impl<Flag>::disconnect(ipc::handle_t h) {
-    detail_impl<policy_t<Flag>>::disconnect(h);
-}
-
-template <typename Flag>
-void chan_impl<Flag>::destroy(ipc::handle_t h) {
-    detail_impl<policy_t<Flag>>::destroy(h);
-}
-
-template <typename Flag>
-char const * chan_impl<Flag>::name(ipc::handle_t h) {
-    auto info = detail_impl<policy_t<Flag>>::info_of(h);
-    return (info == nullptr) ? nullptr : info->name_.c_str();
-}
-
-template <typename Flag>
-std::size_t chan_impl<Flag>::recv_count(ipc::handle_t h) {
-    return detail_impl<policy_t<Flag>>::recv_count(h);
-}
-
-template <typename Flag>
-bool chan_impl<Flag>::wait_for_recv(ipc::handle_t h, std::size_t r_count, std::uint64_t tm) {
-    return detail_impl<policy_t<Flag>>::wait_for_recv(h, r_count, tm);
-}
-
-template <typename Flag>
-bool chan_impl<Flag>::send(ipc::handle_t h, void const * data, std::size_t size, std::uint64_t tm) {
-    return detail_impl<policy_t<Flag>>::send(h, data, size, tm);
-}
-
-template <typename Flag>
-buff_t chan_impl<Flag>::recv(ipc::handle_t h, std::uint64_t tm) {
-    return detail_impl<policy_t<Flag>>::recv(h, tm);
-}
-
-template <typename Flag>
-bool chan_impl<Flag>::try_send(ipc::handle_t h, void const * data, std::size_t size, std::uint64_t tm) {
-    return detail_impl<policy_t<Flag>>::try_send(h, data, size, tm);
-}
-
-template <typename Flag>
-buff_t chan_impl<Flag>::try_recv(ipc::handle_t h) {
-    return detail_impl<policy_t<Flag>>::try_recv(h);
-}
-
-template struct chan_impl<ipc::wr<relat::single, relat::single, trans::unicast  >>;
-// template struct chan_impl<ipc::wr<relat::single, relat::multi , trans::unicast  >>; // TBD
-// template struct chan_impl<ipc::wr<relat::multi , relat::multi , trans::unicast  >>; // TBD
-template struct chan_impl<ipc::wr<relat::single, relat::multi , trans::broadcast>>;
-template struct chan_impl<ipc::wr<relat::multi , relat::multi , trans::broadcast>>;
-
-} // namespace ipc

crazy_functions/test_project/cpp/cppipc/policy.h

@@ -1,25 +0,0 @@
-#pragma once
-
-#include <type_traits>
-
-#include "libipc/def.h"
-#include "libipc/prod_cons.h"
-
-#include "libipc/circ/elem_array.h"
-
-namespace ipc {
-namespace policy {
-
-template <template <typename, std::size_t...> class Elems, typename Flag>
-struct choose;
-
-template <typename Flag>
-struct choose<circ::elem_array, Flag> {
-    using flag_t = Flag;
-
-    template <std::size_t DataSize, std::size_t AlignSize>
-    using elems_t = circ::elem_array<ipc::prod_cons_impl<flag_t>, DataSize, AlignSize>;
-};
-
-} // namespace policy
-} // namespace ipc

crazy_functions/test_project/cpp/cppipc/pool_alloc.cpp

@@ -1,17 +0,0 @@
-#include "libipc/pool_alloc.h"
-
-#include "libipc/memory/resource.h"
-
-namespace ipc {
-namespace mem {
-
-void* pool_alloc::alloc(std::size_t size) {
-    return async_pool_alloc::alloc(size);
-}
-
-void pool_alloc::free(void* p, std::size_t size) {
-    async_pool_alloc::free(p, size);
-}
-
-} // namespace mem
-} // namespace ipc

crazy_functions/test_project/cpp/cppipc/prod_cons.h

@@ -1,433 +0,0 @@
-#pragma once
-
-#include <atomic>
-#include <utility>
-#include <cstring>
-#include <type_traits>
-#include <cstdint>
-
-#include "libipc/def.h"
-
-#include "libipc/platform/detail.h"
-#include "libipc/circ/elem_def.h"
-#include "libipc/utility/log.h"
-#include "libipc/utility/utility.h"
-
-namespace ipc {
-
-////////////////////////////////////////////////////////////////
-/// producer-consumer implementation
-////////////////////////////////////////////////////////////////
-
-template <typename Flag>
-struct prod_cons_impl;
-
-template <>
-struct prod_cons_impl<wr<relat::single, relat::single, trans::unicast>> {
-
-    template <std::size_t DataSize, std::size_t AlignSize>
-    struct elem_t {
-        std::aligned_storage_t<DataSize, AlignSize> data_ {};
-    };
-
-    alignas(cache_line_size) std::atomic<circ::u2_t> rd_; // read index
-    alignas(cache_line_size) std::atomic<circ::u2_t> wt_; // write index
-
-    constexpr circ::u2_t cursor() const noexcept {
-        return 0;
-    }
-
-    template <typename W, typename F, typename E>
-    bool push(W* /*wrapper*/, F&& f, E* elems) {
-        auto cur_wt = circ::index_of(wt_.load(std::memory_order_relaxed));
-        if (cur_wt == circ::index_of(rd_.load(std::memory_order_acquire) - 1)) {
-            return false; // full
-        }
-        std::forward<F>(f)(&(elems[cur_wt].data_));
-        wt_.fetch_add(1, std::memory_order_release);
-        return true;
-    }
-
-    /**
-     * In single-single-unicast, 'force_push' means 'no reader' or 'the only one reader is dead'.
-     * So we could just disconnect all connections of receiver, and return false.
-    */
-    template <typename W, typename F, typename E>
-    bool force_push(W* wrapper, F&&, E*) {
-        wrapper->elems()->disconnect_receiver(~static_cast<circ::cc_t>(0u));
-        return false;
-    }
-
-    template <typename W, typename F, typename R, typename E>
-    bool pop(W* /*wrapper*/, circ::u2_t& /*cur*/, F&& f, R&& out, E* elems) {
-        auto cur_rd = circ::index_of(rd_.load(std::memory_order_relaxed));
-        if (cur_rd == circ::index_of(wt_.load(std::memory_order_acquire))) {
-            return false; // empty
-        }
-        std::forward<F>(f)(&(elems[cur_rd].data_));
-        std::forward<R>(out)(true);
-        rd_.fetch_add(1, std::memory_order_release);
-        return true;
-    }
-};
-
-template <>
-struct prod_cons_impl<wr<relat::single, relat::multi , trans::unicast>>
-     : prod_cons_impl<wr<relat::single, relat::single, trans::unicast>> {
-
-    template <typename W, typename F, typename E>
-    bool force_push(W* wrapper, F&&, E*) {
-        wrapper->elems()->disconnect_receiver(1);
-        return false;
-    }
-
-    template <typename W, typename F, typename R, 
-              template <std::size_t, std::size_t> class E, std::size_t DS, std::size_t AS>
-    bool pop(W* /*wrapper*/, circ::u2_t& /*cur*/, F&& f, R&& out, E<DS, AS>* elems) {
-        byte_t buff[DS];
-        for (unsigned k = 0;;) {
-            auto cur_rd = rd_.load(std::memory_order_relaxed);
-            if (circ::index_of(cur_rd) ==
-                circ::index_of(wt_.load(std::memory_order_acquire))) {
-                return false; // empty
-            }
-            std::memcpy(buff, &(elems[circ::index_of(cur_rd)].data_), sizeof(buff));
-            if (rd_.compare_exchange_weak(cur_rd, cur_rd + 1, std::memory_order_release)) {
-                std::forward<F>(f)(buff);
-                std::forward<R>(out)(true);
-                return true;
-            }
-            ipc::yield(k);
-        }
-    }
-};
-
-template <>
-struct prod_cons_impl<wr<relat::multi , relat::multi, trans::unicast>>
-     : prod_cons_impl<wr<relat::single, relat::multi, trans::unicast>> {
-
-    using flag_t = std::uint64_t;
-
-    template <std::size_t DataSize, std::size_t AlignSize>
-    struct elem_t {
-        std::aligned_storage_t<DataSize, AlignSize> data_ {};
-        std::atomic<flag_t> f_ct_ { 0 }; // commit flag
-    };
-
-    alignas(cache_line_size) std::atomic<circ::u2_t> ct_; // commit index
-
-    template <typename W, typename F, typename E>
-    bool push(W* /*wrapper*/, F&& f, E* elems) {
-        circ::u2_t cur_ct, nxt_ct;
-        for (unsigned k = 0;;) {
-            cur_ct = ct_.load(std::memory_order_relaxed);
-            if (circ::index_of(nxt_ct = cur_ct + 1) ==
-                circ::index_of(rd_.load(std::memory_order_acquire))) {
-                return false; // full
-            }
-            if (ct_.compare_exchange_weak(cur_ct, nxt_ct, std::memory_order_acq_rel)) {
-                break;
-            }
-            ipc::yield(k);
-        }
-        auto* el = elems + circ::index_of(cur_ct);
-        std::forward<F>(f)(&(el->data_));
-        // set flag & try update wt
-        el->f_ct_.store(~static_cast<flag_t>(cur_ct), std::memory_order_release);
-        while (1) {
-            auto cac_ct = el->f_ct_.load(std::memory_order_acquire);
-            if (cur_ct != wt_.load(std::memory_order_relaxed)) {
-                return true;
-            }
-            if ((~cac_ct) != cur_ct) {
-                return true;
-            }
-            if (!el->f_ct_.compare_exchange_strong(cac_ct, 0, std::memory_order_relaxed)) {
-                return true;
-            }
-            wt_.store(nxt_ct, std::memory_order_release);
-            cur_ct = nxt_ct;
-            nxt_ct = cur_ct + 1;
-            el = elems + circ::index_of(cur_ct);
-        }
-        return true;
-    }
-
-    template <typename W, typename F, typename E>
-    bool force_push(W* wrapper, F&&, E*) {
-        wrapper->elems()->disconnect_receiver(1);
-        return false;
-    }
-
-    template <typename W, typename F, typename R, 
-              template <std::size_t, std::size_t> class E, std::size_t DS, std::size_t AS>
-    bool pop(W* /*wrapper*/, circ::u2_t& /*cur*/, F&& f, R&& out, E<DS, AS>* elems) {
-        byte_t buff[DS];
-        for (unsigned k = 0;;) {
-            auto cur_rd = rd_.load(std::memory_order_relaxed);
-            auto cur_wt = wt_.load(std::memory_order_acquire);
-            auto id_rd  = circ::index_of(cur_rd);
-            auto id_wt  = circ::index_of(cur_wt);
-            if (id_rd == id_wt) {
-                auto* el = elems + id_wt;
-                auto cac_ct = el->f_ct_.load(std::memory_order_acquire);
-                if ((~cac_ct) != cur_wt) {
-                    return false; // empty
-                }
-                if (el->f_ct_.compare_exchange_weak(cac_ct, 0, std::memory_order_relaxed)) {
-                    wt_.store(cur_wt + 1, std::memory_order_release);
-                }
-                k = 0;
-            }
-            else {
-                std::memcpy(buff, &(elems[circ::index_of(cur_rd)].data_), sizeof(buff));
-                if (rd_.compare_exchange_weak(cur_rd, cur_rd + 1, std::memory_order_release)) {
-                    std::forward<F>(f)(buff);
-                    std::forward<R>(out)(true);
-                    return true;
-                }
-                ipc::yield(k);
-            }
-        }
-    }
-};
-
-template <>
-struct prod_cons_impl<wr<relat::single, relat::multi, trans::broadcast>> {
-
-    using rc_t = std::uint64_t;
-
-    enum : rc_t {
-        ep_mask = 0x00000000ffffffffull,
-        ep_incr = 0x0000000100000000ull
-    };
-
-    template <std::size_t DataSize, std::size_t AlignSize>
-    struct elem_t {
-        std::aligned_storage_t<DataSize, AlignSize> data_ {};
-        std::atomic<rc_t> rc_ { 0 }; // read-counter
-    };
-
-    alignas(cache_line_size) std::atomic<circ::u2_t> wt_;   // write index
-    alignas(cache_line_size) rc_t epoch_ { 0 };             // only one writer
-
-    circ::u2_t cursor() const noexcept {
-        return wt_.load(std::memory_order_acquire);
-    }
-
-    template <typename W, typename F, typename E>
-    bool push(W* wrapper, F&& f, E* elems) {
-        E* el;
-        for (unsigned k = 0;;) {
-            circ::cc_t cc = wrapper->elems()->connections(std::memory_order_relaxed);
-            if (cc == 0) return false; // no reader
-            el = elems + circ::index_of(wt_.load(std::memory_order_relaxed));
-            // check all consumers have finished reading this element
-            auto cur_rc = el->rc_.load(std::memory_order_acquire);
-            circ::cc_t rem_cc = cur_rc & ep_mask;
-            if ((cc & rem_cc) && ((cur_rc & ~ep_mask) == epoch_)) {
-                return false; // has not finished yet
-            }
-            // consider rem_cc to be 0 here
-            if (el->rc_.compare_exchange_weak(
-                        cur_rc, epoch_ | static_cast<rc_t>(cc), std::memory_order_release)) {
-                break;
-            }
-            ipc::yield(k);
-        }
-        std::forward<F>(f)(&(el->data_));
-        wt_.fetch_add(1, std::memory_order_release);
-        return true;
-    }
-
-    template <typename W, typename F, typename E>
-    bool force_push(W* wrapper, F&& f, E* elems) {
-        E* el;
-        epoch_ += ep_incr;
-        for (unsigned k = 0;;) {
-            circ::cc_t cc = wrapper->elems()->connections(std::memory_order_relaxed);
-            if (cc == 0) return false; // no reader
-            el = elems + circ::index_of(wt_.load(std::memory_order_relaxed));
-            // check all consumers have finished reading this element
-            auto cur_rc = el->rc_.load(std::memory_order_acquire);
-            circ::cc_t rem_cc = cur_rc & ep_mask;
-            if (cc & rem_cc) {
-                ipc::log("force_push: k = %u, cc = %u, rem_cc = %u\n", k, cc, rem_cc);
-                cc = wrapper->elems()->disconnect_receiver(rem_cc); // disconnect all invalid readers
-                if (cc == 0) return false; // no reader
-            }
-            // just compare & exchange
-            if (el->rc_.compare_exchange_weak(
-                        cur_rc, epoch_ | static_cast<rc_t>(cc), std::memory_order_release)) {
-                break;
-            }
-            ipc::yield(k);
-        }
-        std::forward<F>(f)(&(el->data_));
-        wt_.fetch_add(1, std::memory_order_release);
-        return true;
-    }
-
-    template <typename W, typename F, typename R, typename E>
-    bool pop(W* wrapper, circ::u2_t& cur, F&& f, R&& out, E* elems) {
-        if (cur == cursor()) return false; // acquire
-        auto* el = elems + circ::index_of(cur++);
-        std::forward<F>(f)(&(el->data_));
-        for (unsigned k = 0;;) {
-            auto cur_rc = el->rc_.load(std::memory_order_acquire);
-            if ((cur_rc & ep_mask) == 0) {
-                std::forward<R>(out)(true);
-                return true;
-            }
-            auto nxt_rc = cur_rc & ~static_cast<rc_t>(wrapper->connected_id());
-            if (el->rc_.compare_exchange_weak(cur_rc, nxt_rc, std::memory_order_release)) {
-                std::forward<R>(out)((nxt_rc & ep_mask) == 0);
-                return true;
-            }
-            ipc::yield(k);
-        }
-    }
-};
-
-template <>
-struct prod_cons_impl<wr<relat::multi, relat::multi, trans::broadcast>> {
-
-    using rc_t   = std::uint64_t;
-    using flag_t = std::uint64_t;
-
-    enum : rc_t {
-        rc_mask = 0x00000000ffffffffull,
-        ep_mask = 0x00ffffffffffffffull,
-        ep_incr = 0x0100000000000000ull,
-        ic_mask = 0xff000000ffffffffull,
-        ic_incr = 0x0000000100000000ull
-    };
-
-    template <std::size_t DataSize, std::size_t AlignSize>
-    struct elem_t {
-        std::aligned_storage_t<DataSize, AlignSize> data_ {};
-        std::atomic<rc_t  > rc_   { 0 }; // read-counter
-        std::atomic<flag_t> f_ct_ { 0 }; // commit flag
-    };
-
-    alignas(cache_line_size) std::atomic<circ::u2_t> ct_;   // commit index
-    alignas(cache_line_size) std::atomic<rc_t> epoch_ { 0 };
-
-    circ::u2_t cursor() const noexcept {
-        return ct_.load(std::memory_order_acquire);
-    }
-
-    constexpr static rc_t inc_rc(rc_t rc) noexcept {
-        return (rc & ic_mask) | ((rc + ic_incr) & ~ic_mask);
-    }
-
-    constexpr static rc_t inc_mask(rc_t rc) noexcept {
-        return inc_rc(rc) & ~rc_mask;
-    }
-
-    template <typename W, typename F, typename E>
-    bool push(W* wrapper, F&& f, E* elems) {
-        E* el;
-        circ::u2_t cur_ct;
-        rc_t epoch = epoch_.load(std::memory_order_acquire);
-        for (unsigned k = 0;;) {
-            circ::cc_t cc = wrapper->elems()->connections(std::memory_order_relaxed);
-            if (cc == 0) return false; // no reader
-            el = elems + circ::index_of(cur_ct = ct_.load(std::memory_order_relaxed));
-            // check all consumers have finished reading this element
-            auto cur_rc = el->rc_.load(std::memory_order_relaxed);
-            circ::cc_t rem_cc = cur_rc & rc_mask;
-            if ((cc & rem_cc) && ((cur_rc & ~ep_mask) == epoch)) {
-                return false; // has not finished yet
-            }
-            else if (!rem_cc) {
-                auto cur_fl = el->f_ct_.load(std::memory_order_acquire);
-                if ((cur_fl != cur_ct) && cur_fl) {
-                    return false; // full
-                }
-            }
-            // consider rem_cc to be 0 here
-            if (el->rc_.compare_exchange_weak(
-                        cur_rc, inc_mask(epoch | (cur_rc & ep_mask)) | static_cast<rc_t>(cc), std::memory_order_relaxed) &&
-                epoch_.compare_exchange_weak(epoch, epoch, std::memory_order_acq_rel)) {
-                break;
-            }
-            ipc::yield(k);
-        }
-        // only one thread/process would touch here at one time
-        ct_.store(cur_ct + 1, std::memory_order_release);
-        std::forward<F>(f)(&(el->data_));
-        // set flag & try update wt
-        el->f_ct_.store(~static_cast<flag_t>(cur_ct), std::memory_order_release);
-        return true;
-    }
-
-    template <typename W, typename F, typename E>
-    bool force_push(W* wrapper, F&& f, E* elems) {
-        E* el;
-        circ::u2_t cur_ct;
-        rc_t epoch = epoch_.fetch_add(ep_incr, std::memory_order_release) + ep_incr;
-        for (unsigned k = 0;;) {
-            circ::cc_t cc = wrapper->elems()->connections(std::memory_order_relaxed);
-            if (cc == 0) return false; // no reader
-            el = elems + circ::index_of(cur_ct = ct_.load(std::memory_order_relaxed));
-            // check all consumers have finished reading this element
-            auto cur_rc = el->rc_.load(std::memory_order_acquire);
-            circ::cc_t rem_cc = cur_rc & rc_mask;
-            if (cc & rem_cc) {
-                ipc::log("force_push: k = %u, cc = %u, rem_cc = %u\n", k, cc, rem_cc);
-                cc = wrapper->elems()->disconnect_receiver(rem_cc); // disconnect all invalid readers
-                if (cc == 0) return false; // no reader
-            }
-            // just compare & exchange
-            if (el->rc_.compare_exchange_weak(
-                        cur_rc, inc_mask(epoch | (cur_rc & ep_mask)) | static_cast<rc_t>(cc), std::memory_order_relaxed)) {
-                if (epoch == epoch_.load(std::memory_order_acquire)) {
-                    break;
-                }
-                else if (push(wrapper, std::forward<F>(f), elems)) {
-                    return true;
-                }
-                epoch = epoch_.fetch_add(ep_incr, std::memory_order_release) + ep_incr;
-            }
-            ipc::yield(k);
-        }
-        // only one thread/process would touch here at one time
-        ct_.store(cur_ct + 1, std::memory_order_release);
-        std::forward<F>(f)(&(el->data_));
-        // set flag & try update wt
-        el->f_ct_.store(~static_cast<flag_t>(cur_ct), std::memory_order_release);
-        return true;
-    }
-
-    template <typename W, typename F, typename R, typename E, std::size_t N>
-    bool pop(W* wrapper, circ::u2_t& cur, F&& f, R&& out, E(& elems)[N]) {
-        auto* el = elems + circ::index_of(cur);
-        auto cur_fl = el->f_ct_.load(std::memory_order_acquire);
-        if (cur_fl != ~static_cast<flag_t>(cur)) {
-            return false; // empty
-        }
-        ++cur;
-        std::forward<F>(f)(&(el->data_));
-        for (unsigned k = 0;;) {
-            auto cur_rc = el->rc_.load(std::memory_order_acquire);
-            if ((cur_rc & rc_mask) == 0) {
-                std::forward<R>(out)(true);
-                el->f_ct_.store(cur + N - 1, std::memory_order_release);
-                return true;
-            }
-            auto nxt_rc = inc_rc(cur_rc) & ~static_cast<rc_t>(wrapper->connected_id());
-            bool last_one = false;
-            if ((last_one = (nxt_rc & rc_mask) == 0)) {
-                el->f_ct_.store(cur + N - 1, std::memory_order_release);
-            }
-            if (el->rc_.compare_exchange_weak(cur_rc, nxt_rc, std::memory_order_release)) {
-                std::forward<R>(out)(last_one);
-                return true;
-            }
-            ipc::yield(k);
-        }
-    }
-};
-
-} // namespace ipc

crazy_functions/test_project/cpp/cppipc/queue.h

@@ -1,216 +0,0 @@
-#pragma once
-
-#include <type_traits>
-#include <new>
-#include <utility>  // [[since C++14]]: std::exchange
-#include <algorithm>
-#include <atomic>
-#include <tuple>
-#include <thread>
-#include <chrono>
-#include <string>
-#include <cassert>  // assert
-
-#include "libipc/def.h"
-#include "libipc/shm.h"
-#include "libipc/rw_lock.h"
-
-#include "libipc/utility/log.h"
-#include "libipc/platform/detail.h"
-#include "libipc/circ/elem_def.h"
-
-namespace ipc {
-namespace detail {
-
-class queue_conn {
-protected:
-    circ::cc_t connected_ = 0;
-    shm::handle elems_h_;
-
-    template <typename Elems>
-    Elems* open(char const * name) {
-        if (name == nullptr || name[0] == '\0') {
-            ipc::error("fail open waiter: name is empty!\n");
-            return nullptr;
-        }
-        if (!elems_h_.acquire(name, sizeof(Elems))) {
-            return nullptr;
-        }
-        auto elems = static_cast<Elems*>(elems_h_.get());
-        if (elems == nullptr) {
-            ipc::error("fail acquire elems: %s\n", name);
-            return nullptr;
-        }
-        elems->init();
-        return elems;
-    }
-
-    void close() {
-        elems_h_.release();
-    }
-
-public:
-    queue_conn() = default;
-    queue_conn(const queue_conn&) = delete;
-    queue_conn& operator=(const queue_conn&) = delete;
-
-    bool connected() const noexcept {
-        return connected_ != 0;
-    }
-
-    circ::cc_t connected_id() const noexcept {
-        return connected_;
-    }
-
-    template <typename Elems>
-    auto connect(Elems* elems) noexcept
-                         /*needs 'optional' here*/
-     -> std::tuple<bool, bool, decltype(std::declval<Elems>().cursor())> {
-        if (elems == nullptr) return {};
-        // if it's already connected, just return
-        if (connected()) return {connected(), false, 0};
-        connected_ = elems->connect_receiver();
-        return {connected(), true, elems->cursor()};
-    }
-
-    template <typename Elems>
-    bool disconnect(Elems* elems) noexcept {
-        if (elems == nullptr) return false;
-        // if it's already disconnected, just return false
-        if (!connected()) return false;
-        elems->disconnect_receiver(std::exchange(connected_, 0));
-        return true;
-    }
-};
-
-template <typename Elems>
-class queue_base : public queue_conn {
-    using base_t = queue_conn;
-
-public:
-    using elems_t  = Elems;
-    using policy_t = typename elems_t::policy_t;
-
-protected:
-    elems_t * elems_ = nullptr;
-    decltype(std::declval<elems_t>().cursor()) cursor_ = 0;
-    bool sender_flag_ = false;
-
-public:
-    using base_t::base_t;
-
-    queue_base() = default;
-
-    explicit queue_base(char const * name)
-        : queue_base{} {
-        elems_ = open<elems_t>(name);
-    }
-
-    explicit queue_base(elems_t * elems) noexcept
-        : queue_base{} {
-        assert(elems != nullptr);
-        elems_ = elems;
-    }
-
-    /* not virtual */ ~queue_base() {
-        base_t::close();
-    }
-
-    elems_t       * elems()       noexcept { return elems_; }
-    elems_t const * elems() const noexcept { return elems_; }
-
-    bool ready_sending() noexcept {
-        if (elems_ == nullptr) return false;
-        return sender_flag_ || (sender_flag_ = elems_->connect_sender());
-    }
-
-    void shut_sending() noexcept {
-        if (elems_ == nullptr) return;
-        if (!sender_flag_) return;
-        elems_->disconnect_sender();
-    }
-
-    bool connect() noexcept {
-        auto tp = base_t::connect(elems_);
-        if (std::get<0>(tp) && std::get<1>(tp)) {
-            cursor_ = std::get<2>(tp);
-            return true;
-        }
-        return std::get<0>(tp);
-    }
-
-    bool disconnect() noexcept {
-        return base_t::disconnect(elems_);
-    }
-
-    std::size_t conn_count() const noexcept {
-        return (elems_ == nullptr) ? static_cast<std::size_t>(invalid_value) : elems_->conn_count();
-    }
-
-    bool valid() const noexcept {
-        return elems_ != nullptr;
-    }
-
-    bool empty() const noexcept {
-        return !valid() || (cursor_ == elems_->cursor());
-    }
-
-    template <typename T, typename F, typename... P>
-    bool push(F&& prep, P&&... params) {
-        if (elems_ == nullptr) return false;
-        return elems_->push(this, [&](void* p) {
-            if (prep(p)) ::new (p) T(std::forward<P>(params)...);
-        });
-    }
-
-    template <typename T, typename F, typename... P>
-    bool force_push(F&& prep, P&&... params) {
-        if (elems_ == nullptr) return false;
-        return elems_->force_push(this, [&](void* p) {
-            if (prep(p)) ::new (p) T(std::forward<P>(params)...);
-        });
-    }
-
-    template <typename T, typename F>
-    bool pop(T& item, F&& out) {
-        if (elems_ == nullptr) {
-            return false;
-        }
-        return elems_->pop(this, &(this->cursor_), [&item](void* p) {
-            ::new (&item) T(std::move(*static_cast<T*>(p)));
-        }, std::forward<F>(out));
-    }
-};
-
-} // namespace detail
-
-template <typename T, typename Policy>
-class queue final : public detail::queue_base<typename Policy::template elems_t<sizeof(T), alignof(T)>> {
-    using base_t = detail::queue_base<typename Policy::template elems_t<sizeof(T), alignof(T)>>;
-
-public:
-    using value_t = T;
-
-    using base_t::base_t;
-
-    template <typename... P>
-    bool push(P&&... params) {
-        return base_t::template push<T>(std::forward<P>(params)...);
-    }
-
-    template <typename... P>
-    bool force_push(P&&... params) {
-        return base_t::template force_push<T>(std::forward<P>(params)...);
-    }
-
-    bool pop(T& item) {
-        return base_t::pop(item, [](bool) {});
-    }
-
-    template <typename F>
-    bool pop(T& item, F&& out) {
-        return base_t::pop(item, std::forward<F>(out));
-    }
-};
-
-} // namespace ipc

crazy_functions/test_project/cpp/cppipc/shm.cpp

@@ -1,103 +0,0 @@
-
-#include <string>
-#include <utility>
-
-#include "libipc/shm.h"
-
-#include "libipc/utility/pimpl.h"
-#include "libipc/memory/resource.h"
-
-namespace ipc {
-namespace shm {
-
-class handle::handle_ : public pimpl<handle_> {
-public:
-    shm::id_t id_ = nullptr;
-    void*     m_  = nullptr;
-
-    ipc::string n_;
-    std::size_t s_ = 0;
-};
-
-handle::handle()
-    : p_(p_->make()) {
-}
-
-handle::handle(char const * name, std::size_t size, unsigned mode)
-    : handle() {
-    acquire(name, size, mode);
-}
-
-handle::handle(handle&& rhs)
-    : handle() {
-    swap(rhs);
-}
-
-handle::~handle() {
-    release();
-    p_->clear();
-}
-
-void handle::swap(handle& rhs) {
-    std::swap(p_, rhs.p_);
-}
-
-handle& handle::operator=(handle rhs) {
-    swap(rhs);
-    return *this;
-}
-
-bool handle::valid() const noexcept {
-    return impl(p_)->m_ != nullptr;
-}
-
-std::size_t handle::size() const noexcept {
-    return impl(p_)->s_;
-}
-
-char const * handle::name() const noexcept {
-    return impl(p_)->n_.c_str();
-}
-
-std::int32_t handle::ref() const noexcept {
-    return shm::get_ref(impl(p_)->id_);
-}
-
-void handle::sub_ref() noexcept {
-    shm::sub_ref(impl(p_)->id_);
-}
-
-bool handle::acquire(char const * name, std::size_t size, unsigned mode) {
-    release();
-    impl(p_)->id_ = shm::acquire((impl(p_)->n_ = name).c_str(), size, mode);
-    impl(p_)->m_  = shm::get_mem(impl(p_)->id_, &(impl(p_)->s_));
-    return valid();
-}
-
-std::int32_t handle::release() {
-    if (impl(p_)->id_ == nullptr) return -1;
-    return shm::release(detach());
-}
-
-void* handle::get() const {
-    return impl(p_)->m_;
-}
-
-void handle::attach(id_t id) {
-    if (id == nullptr) return;
-    release();
-    impl(p_)->id_ = id;
-    impl(p_)->m_  = shm::get_mem(impl(p_)->id_, &(impl(p_)->s_));
-}
-
-id_t handle::detach() {
-    auto old = impl(p_)->id_;
-    impl(p_)->id_ = nullptr;
-    impl(p_)->m_  = nullptr;
-    impl(p_)->s_  = 0;
-    impl(p_)->n_.clear();
-    return old;
-}
-
-} // namespace shm
-} // namespace ipc

crazy_functions/test_project/cpp/cppipc/waiter.h

@@ -1,83 +0,0 @@
-#pragma once
-
-#include <utility>
-#include <string>
-#include <mutex>
-#include <atomic>
-
-#include "libipc/def.h"
-#include "libipc/mutex.h"
-#include "libipc/condition.h"
-#include "libipc/platform/detail.h"
-
-namespace ipc {
-namespace detail {
-
-class waiter {
-    ipc::sync::condition cond_;
-    ipc::sync::mutex     lock_;
-    std::atomic<bool>    quit_ {false};
-
-public:
-    static void init();
-
-    waiter() = default;
-    waiter(char const *name) {
-        open(name);
-    }
-
-    ~waiter() {
-        close();
-    }
-
-    bool valid() const noexcept {
-        return cond_.valid() && lock_.valid();
-    }
-
-    bool open(char const *name) noexcept {
-        quit_.store(false, std::memory_order_relaxed);
-        if (!cond_.open((std::string{"_waiter_cond_"} + name).c_str())) {
-            return false;
-        }
-        if (!lock_.open((std::string{"_waiter_lock_"} + name).c_str())) {
-            cond_.close();
-            return false;
-        }
-        return valid();
-    }
-
-    void close() noexcept {
-        cond_.close();
-        lock_.close();
-    }
-
-    template <typename F>
-    bool wait_if(F &&pred, std::uint64_t tm = ipc::invalid_value) noexcept {
-        IPC_UNUSED_ std::lock_guard<ipc::sync::mutex> guard {lock_};
-        while ([this, &pred] {
-                    return !quit_.load(std::memory_order_relaxed)
-                        && std::forward<F>(pred)();
-                }()) {
-            if (!cond_.wait(lock_, tm)) return false;
-        }
-        return true;
-    }
-
-    bool notify() noexcept {
-        std::lock_guard<ipc::sync::mutex>{lock_}; // barrier
-        return cond_.notify(lock_);
-    }
-
-    bool broadcast() noexcept {
-        std::lock_guard<ipc::sync::mutex>{lock_}; // barrier
-        return cond_.broadcast(lock_);
-    }
-
-    bool quit_waiting() {
-        quit_.store(true, std::memory_order_release);
-        return broadcast();
-    }
-};
-
-} // namespace detail
-} // namespace ipc

crazy_functions/test_project/cpp/cppipc/来源

@@ -1,3 +0,0 @@
-https://github.com/mutouyun/cpp-ipc
-
-A high-performance inter-process communication library using shared memory on Linux/Windows.

crazy_functions/test_project/cpp/libJPG/jpgd.h

@@ -1,316 +0,0 @@
-// jpgd.h - C++ class for JPEG decompression.
-// Public domain, Rich Geldreich <richgel99@gmail.com>
-#ifndef JPEG_DECODER_H
-#define JPEG_DECODER_H
-
-#include <stdlib.h>
-#include <stdio.h>
-#include <setjmp.h>
-
-namespace jpgd
-{
-  typedef unsigned char  uint8;
-  typedef   signed short int16;
-  typedef unsigned short uint16;
-  typedef unsigned int   uint;
-  typedef   signed int   int32;
-
-  // Loads a JPEG image from a memory buffer or a file.
-  // req_comps can be 1 (grayscale), 3 (RGB), or 4 (RGBA).
-  // On return, width/height will be set to the image's dimensions, and actual_comps will be set to the either 1 (grayscale) or 3 (RGB).
-  // Notes: For more control over where and how the source data is read, see the decompress_jpeg_image_from_stream() function below, or call the jpeg_decoder class directly.
-  // Requesting a 8 or 32bpp image is currently a little faster than 24bpp because the jpeg_decoder class itself currently always unpacks to either 8 or 32bpp.
-// BEGIN EPIC MOD
-//unsigned char *decompress_jpeg_image_from_memory(const unsigned char *pSrc_data, int src_data_size, int *width, int *height, int *actual_comps, int req_comps);
-  unsigned char *decompress_jpeg_image_from_memory(const unsigned char *pSrc_data, int src_data_size, int *width, int *height, int *actual_comps, int req_comps, int format);
-// END EPIC MOD
-  unsigned char *decompress_jpeg_image_from_file(const char *pSrc_filename, int *width, int *height, int *actual_comps, int req_comps);
-
-  // Success/failure error codes.
-  enum jpgd_status
-  {
-    JPGD_SUCCESS = 0, JPGD_FAILED = -1, JPGD_DONE = 1,
-    JPGD_BAD_DHT_COUNTS = -256, JPGD_BAD_DHT_INDEX, JPGD_BAD_DHT_MARKER, JPGD_BAD_DQT_MARKER, JPGD_BAD_DQT_TABLE, 
-    JPGD_BAD_PRECISION, JPGD_BAD_HEIGHT, JPGD_BAD_WIDTH, JPGD_TOO_MANY_COMPONENTS, 
-    JPGD_BAD_SOF_LENGTH, JPGD_BAD_VARIABLE_MARKER, JPGD_BAD_DRI_LENGTH, JPGD_BAD_SOS_LENGTH,
-    JPGD_BAD_SOS_COMP_ID, JPGD_W_EXTRA_BYTES_BEFORE_MARKER, JPGD_NO_ARITHMITIC_SUPPORT, JPGD_UNEXPECTED_MARKER,
-    JPGD_NOT_JPEG, JPGD_UNSUPPORTED_MARKER, JPGD_BAD_DQT_LENGTH, JPGD_TOO_MANY_BLOCKS,
-    JPGD_UNDEFINED_QUANT_TABLE, JPGD_UNDEFINED_HUFF_TABLE, JPGD_NOT_SINGLE_SCAN, JPGD_UNSUPPORTED_COLORSPACE,
-    JPGD_UNSUPPORTED_SAMP_FACTORS, JPGD_DECODE_ERROR, JPGD_BAD_RESTART_MARKER, JPGD_ASSERTION_ERROR,
-    JPGD_BAD_SOS_SPECTRAL, JPGD_BAD_SOS_SUCCESSIVE, JPGD_STREAM_READ, JPGD_NOTENOUGHMEM
-  };
-    
-  // Input stream interface.
-  // Derive from this class to read input data from sources other than files or memory. Set m_eof_flag to true when no more data is available.
-  // The decoder is rather greedy: it will keep on calling this method until its internal input buffer is full, or until the EOF flag is set.
-  // It the input stream contains data after the JPEG stream's EOI (end of image) marker it will probably be pulled into the internal buffer.
-  // Call the get_total_bytes_read() method to determine the actual size of the JPEG stream after successful decoding.
-  class jpeg_decoder_stream
-  {
-  public:
-    jpeg_decoder_stream() { }
-    virtual ~jpeg_decoder_stream() { }
-
-    // The read() method is called when the internal input buffer is empty.
-    // Parameters:
-    // pBuf - input buffer
-    // max_bytes_to_read - maximum bytes that can be written to pBuf
-    // pEOF_flag - set this to true if at end of stream (no more bytes remaining)
-    // Returns -1 on error, otherwise return the number of bytes actually written to the buffer (which may be 0).
-    // Notes: This method will be called in a loop until you set *pEOF_flag to true or the internal buffer is full.
-    virtual int read(uint8 *pBuf, int max_bytes_to_read, bool *pEOF_flag) = 0;
-  };
-
-  // stdio FILE stream class.
-  class jpeg_decoder_file_stream : public jpeg_decoder_stream
-  {
-    jpeg_decoder_file_stream(const jpeg_decoder_file_stream &);
-    jpeg_decoder_file_stream &operator =(const jpeg_decoder_file_stream &);
-
-    FILE *m_pFile;
-    bool m_eof_flag, m_error_flag;
-
-  public:
-    jpeg_decoder_file_stream();
-    virtual ~jpeg_decoder_file_stream();
-    
-    bool open(const char *Pfilename);
-    void close();
-
-    virtual int read(uint8 *pBuf, int max_bytes_to_read, bool *pEOF_flag);
-  };
-
-  // Memory stream class.
-  class jpeg_decoder_mem_stream : public jpeg_decoder_stream
-  {
-    const uint8 *m_pSrc_data;
-    uint m_ofs, m_size;
-
-  public:
-    jpeg_decoder_mem_stream() : m_pSrc_data(NULL), m_ofs(0), m_size(0) { }
-    jpeg_decoder_mem_stream(const uint8 *pSrc_data, uint size) : m_pSrc_data(pSrc_data), m_ofs(0), m_size(size) { }
-
-    virtual ~jpeg_decoder_mem_stream() { }
-
-    bool open(const uint8 *pSrc_data, uint size);
-    void close() { m_pSrc_data = NULL; m_ofs = 0; m_size = 0; }
-    
-    virtual int read(uint8 *pBuf, int max_bytes_to_read, bool *pEOF_flag);
-  };
-
-  // Loads JPEG file from a jpeg_decoder_stream.
-  unsigned char *decompress_jpeg_image_from_stream(jpeg_decoder_stream *pStream, int *width, int *height, int *actual_comps, int req_comps);
-
-  enum 
-  { 
-    JPGD_IN_BUF_SIZE = 8192, JPGD_MAX_BLOCKS_PER_MCU = 10, JPGD_MAX_HUFF_TABLES = 8, JPGD_MAX_QUANT_TABLES = 4, 
-    JPGD_MAX_COMPONENTS = 4, JPGD_MAX_COMPS_IN_SCAN = 4, JPGD_MAX_BLOCKS_PER_ROW = 8192, JPGD_MAX_HEIGHT = 16384, JPGD_MAX_WIDTH = 16384 
-  };
-          
-  typedef int16 jpgd_quant_t;
-  typedef int16 jpgd_block_t;
-
-  class jpeg_decoder
-  {
-  public:
-    // Call get_error_code() after constructing to determine if the stream is valid or not. You may call the get_width(), get_height(), etc.
-    // methods after the constructor is called. You may then either destruct the object, or begin decoding the image by calling begin_decoding(), then decode() on each scanline.
-    jpeg_decoder(jpeg_decoder_stream *pStream);
-
-    ~jpeg_decoder();
-
-    // Call this method after constructing the object to begin decompression.
-    // If JPGD_SUCCESS is returned you may then call decode() on each scanline.
-    int begin_decoding();
-
-    // Returns the next scan line.
-    // For grayscale images, pScan_line will point to a buffer containing 8-bit pixels (get_bytes_per_pixel() will return 1). 
-    // Otherwise, it will always point to a buffer containing 32-bit RGBA pixels (A will always be 255, and get_bytes_per_pixel() will return 4).
-    // Returns JPGD_SUCCESS if a scan line has been returned.
-    // Returns JPGD_DONE if all scan lines have been returned.
-    // Returns JPGD_FAILED if an error occurred. Call get_error_code() for a more info.
-    int decode(const void** pScan_line, uint* pScan_line_len);
-    
-    inline jpgd_status get_error_code() const { return m_error_code; }
-
-    inline int get_width() const { return m_image_x_size; }
-    inline int get_height() const { return m_image_y_size; }
-
-    inline int get_num_components() const { return m_comps_in_frame; }
-
-    inline int get_bytes_per_pixel() const { return m_dest_bytes_per_pixel; }
-    inline int get_bytes_per_scan_line() const { return m_image_x_size * get_bytes_per_pixel(); }
-
-    // Returns the total number of bytes actually consumed by the decoder (which should equal the actual size of the JPEG file).
-    inline int get_total_bytes_read() const { return m_total_bytes_read; }
-    
-  private:
-    jpeg_decoder(const jpeg_decoder &);
-    jpeg_decoder &operator =(const jpeg_decoder &);
-
-    typedef void (*pDecode_block_func)(jpeg_decoder *, int, int, int);
-
-    struct huff_tables
-    {
-      bool ac_table;
-      uint  look_up[256];
-      uint  look_up2[256];
-      uint8 code_size[256];
-      uint  tree[512];
-    };
-
-    struct coeff_buf
-    {
-      uint8 *pData;
-      int block_num_x, block_num_y;
-      int block_len_x, block_len_y;
-      int block_size;
-    };
-
-    struct mem_block
-    {
-      mem_block *m_pNext;
-      size_t m_used_count;
-      size_t m_size;
-      char m_data[1];
-    };
-
-    jmp_buf m_jmp_state;
-    mem_block *m_pMem_blocks;
-    int m_image_x_size;
-    int m_image_y_size;
-    jpeg_decoder_stream *m_pStream;
-    int m_progressive_flag;
-    uint8 m_huff_ac[JPGD_MAX_HUFF_TABLES];
-    uint8* m_huff_num[JPGD_MAX_HUFF_TABLES];      // pointer to number of Huffman codes per bit size
-    uint8* m_huff_val[JPGD_MAX_HUFF_TABLES];      // pointer to Huffman codes per bit size
-    jpgd_quant_t* m_quant[JPGD_MAX_QUANT_TABLES]; // pointer to quantization tables
-    int m_scan_type;                              // Gray, Yh1v1, Yh1v2, Yh2v1, Yh2v2 (CMYK111, CMYK4114 no longer supported)
-    int m_comps_in_frame;                         // # of components in frame
-    int m_comp_h_samp[JPGD_MAX_COMPONENTS];       // component's horizontal sampling factor
-    int m_comp_v_samp[JPGD_MAX_COMPONENTS];       // component's vertical sampling factor
-    int m_comp_quant[JPGD_MAX_COMPONENTS];        // component's quantization table selector
-    int m_comp_ident[JPGD_MAX_COMPONENTS];        // component's ID
-    int m_comp_h_blocks[JPGD_MAX_COMPONENTS];
-    int m_comp_v_blocks[JPGD_MAX_COMPONENTS];
-    int m_comps_in_scan;                          // # of components in scan
-    int m_comp_list[JPGD_MAX_COMPS_IN_SCAN];      // components in this scan
-    int m_comp_dc_tab[JPGD_MAX_COMPONENTS];       // component's DC Huffman coding table selector
-    int m_comp_ac_tab[JPGD_MAX_COMPONENTS];       // component's AC Huffman coding table selector
-    int m_spectral_start;                         // spectral selection start
-    int m_spectral_end;                           // spectral selection end
-    int m_successive_low;                         // successive approximation low
-    int m_successive_high;                        // successive approximation high
-    int m_max_mcu_x_size;                         // MCU's max. X size in pixels
-    int m_max_mcu_y_size;                         // MCU's max. Y size in pixels
-    int m_blocks_per_mcu;
-    int m_max_blocks_per_row;
-    int m_mcus_per_row, m_mcus_per_col;
-    int m_mcu_org[JPGD_MAX_BLOCKS_PER_MCU];
-    int m_total_lines_left;                       // total # lines left in image
-    int m_mcu_lines_left;                         // total # lines left in this MCU
-    int m_real_dest_bytes_per_scan_line;
-    int m_dest_bytes_per_scan_line;               // rounded up
-    int m_dest_bytes_per_pixel;                   // 4 (RGB) or 1 (Y)
-    huff_tables* m_pHuff_tabs[JPGD_MAX_HUFF_TABLES];
-    coeff_buf* m_dc_coeffs[JPGD_MAX_COMPONENTS];
-    coeff_buf* m_ac_coeffs[JPGD_MAX_COMPONENTS];
-    int m_eob_run;
-    int m_block_y_mcu[JPGD_MAX_COMPONENTS];
-    uint8* m_pIn_buf_ofs;
-    int m_in_buf_left;
-    int m_tem_flag;
-    bool m_eof_flag;
-    uint8 m_in_buf_pad_start[128];
-    uint8 m_in_buf[JPGD_IN_BUF_SIZE + 128];
-    uint8 m_in_buf_pad_end[128];
-    int m_bits_left;
-    uint m_bit_buf;
-    int m_restart_interval;
-    int m_restarts_left;
-    int m_next_restart_num;
-    int m_max_mcus_per_row;
-    int m_max_blocks_per_mcu;
-    int m_expanded_blocks_per_mcu;
-    int m_expanded_blocks_per_row;
-    int m_expanded_blocks_per_component;
-    bool  m_freq_domain_chroma_upsample;
-    int m_max_mcus_per_col;
-    uint m_last_dc_val[JPGD_MAX_COMPONENTS];
-    jpgd_block_t* m_pMCU_coefficients;
-    int m_mcu_block_max_zag[JPGD_MAX_BLOCKS_PER_MCU];
-    uint8* m_pSample_buf;
-    int m_crr[256];
-    int m_cbb[256];
-    int m_crg[256];
-    int m_cbg[256];
-    uint8* m_pScan_line_0;
-    uint8* m_pScan_line_1;
-    jpgd_status m_error_code;
-    bool m_ready_flag;
-    int m_total_bytes_read;
-
-    void free_all_blocks();
-    // BEGIN EPIC MOD
-    UE_NORETURN void stop_decoding(jpgd_status status);
-    // END EPIC MOD
-    void *alloc(size_t n, bool zero = false);
-    void word_clear(void *p, uint16 c, uint n);
-    void prep_in_buffer();
-    void read_dht_marker();
-    void read_dqt_marker();
-    void read_sof_marker();
-    void skip_variable_marker();
-    void read_dri_marker();
-    void read_sos_marker();
-    int next_marker();
-    int process_markers();
-    void locate_soi_marker();
-    void locate_sof_marker();
-    int locate_sos_marker();
-    void init(jpeg_decoder_stream * pStream);
-    void create_look_ups();
-    void fix_in_buffer();
-    void transform_mcu(int mcu_row);
-    void transform_mcu_expand(int mcu_row);
-    coeff_buf* coeff_buf_open(int block_num_x, int block_num_y, int block_len_x, int block_len_y);
-    inline jpgd_block_t *coeff_buf_getp(coeff_buf *cb, int block_x, int block_y);
-    void load_next_row();
-    void decode_next_row();
-    void make_huff_table(int index, huff_tables *pH);
-    void check_quant_tables();
-    void check_huff_tables();
-    void calc_mcu_block_order();
-    int init_scan();
-    void init_frame();
-    void process_restart();
-    void decode_scan(pDecode_block_func decode_block_func);
-    void init_progressive();
-    void init_sequential();
-    void decode_start();
-    void decode_init(jpeg_decoder_stream * pStream);
-    void H2V2Convert();
-    void H2V1Convert();
-    void H1V2Convert();
-    void H1V1Convert();
-    void gray_convert();
-    void expanded_convert();
-    void find_eoi();
-    inline uint get_char();
-    inline uint get_char(bool *pPadding_flag);
-    inline void stuff_char(uint8 q);
-    inline uint8 get_octet();
-    inline uint get_bits(int num_bits);
-    inline uint get_bits_no_markers(int numbits);
-    inline int huff_decode(huff_tables *pH);
-    inline int huff_decode(huff_tables *pH, int& extrabits);
-    static inline uint8 clamp(int i);
-    static void decode_block_dc_first(jpeg_decoder *pD, int component_id, int block_x, int block_y);
-    static void decode_block_dc_refine(jpeg_decoder *pD, int component_id, int block_x, int block_y);
-    static void decode_block_ac_first(jpeg_decoder *pD, int component_id, int block_x, int block_y);
-    static void decode_block_ac_refine(jpeg_decoder *pD, int component_id, int block_x, int block_y);
-  };
-  
-} // namespace jpgd
-
-#endif // JPEG_DECODER_H

crazy_functions/test_project/cpp/libJPG/jpge.h

@@ -1,172 +0,0 @@
-
-// jpge.h - C++ class for JPEG compression.
-// Public domain, Rich Geldreich <richgel99@gmail.com>
-// Alex Evans: Added RGBA support, linear memory allocator.
-#ifndef JPEG_ENCODER_H
-#define JPEG_ENCODER_H
-
-#include <stdint.h>
-
-namespace jpge
-{
-  typedef unsigned char  uint8;
-  typedef signed short   int16;
-  typedef signed int     int32;
-  typedef unsigned short uint16;
-  typedef unsigned int   uint32;
-  typedef unsigned int   uint;
-
-  // JPEG chroma subsampling factors. Y_ONLY (grayscale images) and H2V2 (color images) are the most common.
-  enum subsampling_t { Y_ONLY = 0, H1V1 = 1, H2V1 = 2, H2V2 = 3 };
-
-  // JPEG compression parameters structure.
-  struct params
-  {
-    inline params() : m_quality(85), m_subsampling(H2V2), m_no_chroma_discrim_flag(false), m_two_pass_flag(false) { }
-
-    inline bool check_valid() const
-    {
-      if ((m_quality < 1) || (m_quality > 100)) return false;
-      if ((uint)m_subsampling > (uint)H2V2) return false;
-      return true;
-    }
-
-    // Quality: 1-100, higher is better. Typical values are around 50-95.
-    int m_quality;
-
-    // m_subsampling:
-    // 0 = Y (grayscale) only
-    // 1 = YCbCr, no subsampling (H1V1, YCbCr 1x1x1, 3 blocks per MCU)
-    // 2 = YCbCr, H2V1 subsampling (YCbCr 2x1x1, 4 blocks per MCU)
-    // 3 = YCbCr, H2V2 subsampling (YCbCr 4x1x1, 6 blocks per MCU-- very common)
-    subsampling_t m_subsampling;
-
-    // Disables CbCr discrimination - only intended for testing.
-    // If true, the Y quantization table is also used for the CbCr channels.
-    bool m_no_chroma_discrim_flag;
-
-    bool m_two_pass_flag;
-  };
-  
-  // Writes JPEG image to a file. 
-  // num_channels must be 1 (Y) or 3 (RGB), image pitch must be width*num_channels.
-  bool compress_image_to_jpeg_file(const char *pFilename, int64_t width, int64_t height, int64_t num_channels, const uint8 *pImage_data, const params &comp_params = params());
-
-  // Writes JPEG image to memory buffer. 
-  // On entry, buf_size is the size of the output buffer pointed at by pBuf, which should be at least ~1024 bytes. 
-  // If return value is true, buf_size will be set to the size of the compressed data.
-  bool compress_image_to_jpeg_file_in_memory(void *pBuf, int64_t &buf_size, int64_t width, int64_t height, int64_t num_channels, const uint8 *pImage_data, const params &comp_params = params());
-    
-  // Output stream abstract class - used by the jpeg_encoder class to write to the output stream. 
-  // put_buf() is generally called with len==JPGE_OUT_BUF_SIZE bytes, but for headers it'll be called with smaller amounts.
-  class output_stream
-  {
-  public:
-    virtual ~output_stream() { };
-    virtual bool put_buf(const void* Pbuf, int64_t len) = 0;
-    template<class T> inline bool put_obj(const T& obj) { return put_buf(&obj, sizeof(T)); }
-  };
-    
-  // Lower level jpeg_encoder class - useful if more control is needed than the above helper functions.
-  class jpeg_encoder
-  {
-  public:
-    jpeg_encoder();
-    ~jpeg_encoder();
-
-    // Initializes the compressor.
-    // pStream: The stream object to use for writing compressed data.
-    // params - Compression parameters structure, defined above.
-    // width, height  - Image dimensions.
-    // channels - May be 1, or 3. 1 indicates grayscale, 3 indicates RGB source data.
-    // Returns false on out of memory or if a stream write fails.
-    bool init(output_stream *pStream, int64_t width, int64_t height, int64_t src_channels, const params &comp_params = params());
-    
-    const params &get_params() const { return m_params; }
-    
-    // Deinitializes the compressor, freeing any allocated memory. May be called at any time.
-    void deinit();
-
-    uint get_total_passes() const { return m_params.m_two_pass_flag ? 2 : 1; }
-    inline uint get_cur_pass() { return m_pass_num; }
-
-    // Call this method with each source scanline.
-    // width * src_channels bytes per scanline is expected (RGB or Y format).
-    // You must call with NULL after all scanlines are processed to finish compression.
-    // Returns false on out of memory or if a stream write fails.
-    bool process_scanline(const void* pScanline);
-        
-  private:
-    jpeg_encoder(const jpeg_encoder &);
-    jpeg_encoder &operator =(const jpeg_encoder &);
-
-    typedef int32 sample_array_t;
-        
-    output_stream *m_pStream;
-    params m_params;
-    uint8 m_num_components;
-    uint8 m_comp_h_samp[3], m_comp_v_samp[3];
-    int m_image_x, m_image_y, m_image_bpp, m_image_bpl;
-    int m_image_x_mcu, m_image_y_mcu;
-    int m_image_bpl_xlt, m_image_bpl_mcu;
-    int m_mcus_per_row;
-    int m_mcu_x, m_mcu_y;
-    uint8 *m_mcu_lines[16];
-    uint8 m_mcu_y_ofs;
-    sample_array_t m_sample_array[64];
-    int16 m_coefficient_array[64];
-    int32 m_quantization_tables[2][64];
-    uint m_huff_codes[4][256];
-    uint8 m_huff_code_sizes[4][256];
-    uint8 m_huff_bits[4][17];
-    uint8 m_huff_val[4][256];
-    uint32 m_huff_count[4][256];
-    int m_last_dc_val[3];
-    enum { JPGE_OUT_BUF_SIZE = 2048 };
-    uint8 m_out_buf[JPGE_OUT_BUF_SIZE];
-    uint8 *m_pOut_buf;
-    uint m_out_buf_left;
-    uint32 m_bit_buffer;
-    uint m_bits_in;
-    uint8 m_pass_num;
-    bool m_all_stream_writes_succeeded;
-        
-    void optimize_huffman_table(int table_num, int table_len);
-    void emit_byte(uint8 i);
-    void emit_word(uint i);
-    void emit_marker(int marker);
-    void emit_jfif_app0();
-    void emit_dqt();
-    void emit_sof();
-    void emit_dht(uint8 *bits, uint8 *val, int index, bool ac_flag);
-    void emit_dhts();
-    void emit_sos();
-    void emit_markers();
-    void compute_huffman_table(uint *codes, uint8 *code_sizes, uint8 *bits, uint8 *val);
-    void compute_quant_table(int32 *dst, int16 *src);
-    void adjust_quant_table(int32 *dst, int32 *src);
-    void first_pass_init();
-    bool second_pass_init();
-    bool jpg_open(int p_x_res, int p_y_res, int src_channels);
-    void load_block_8_8_grey(int x);
-    void load_block_8_8(int x, int y, int c);
-    void load_block_16_8(int x, int c);
-    void load_block_16_8_8(int x, int c);
-    void load_quantized_coefficients(int component_num);
-    void flush_output_buffer();
-    void put_bits(uint bits, uint len);
-    void code_coefficients_pass_one(int component_num);
-    void code_coefficients_pass_two(int component_num);
-    void code_block(int component_num);
-    void process_mcu_row();
-    bool terminate_pass_one();
-    bool terminate_pass_two();
-    bool process_end_of_image();
-    void load_mcu(const void* src);
-    void clear();
-    void init();
-  };
-
-} // namespace jpge
-
-#endif // JPEG_ENCODER

crazy_functions/test_project/cpp/libJPG/来源

@@ -1,3 +0,0 @@
-jpge.h - C++ class for JPEG compression.
-Public domain, Rich Geldreich <richgel99@gmail.com>
-Alex Evans: Added RGBA support, linear memory allocator.

crazy_functions/test_project/cpp/longcode/prod_cons.h

@@ -1,433 +0,0 @@
-#pragma once
-
-#include <atomic>
-#include <utility>
-#include <cstring>
-#include <type_traits>
-#include <cstdint>
-
-#include "libipc/def.h"
-
-#include "libipc/platform/detail.h"
-#include "libipc/circ/elem_def.h"
-#include "libipc/utility/log.h"
-#include "libipc/utility/utility.h"
-
-namespace ipc {
-
-////////////////////////////////////////////////////////////////
-/// producer-consumer implementation
-////////////////////////////////////////////////////////////////
-
-template <typename Flag>
-struct prod_cons_impl;
-
-template <>
-struct prod_cons_impl<wr<relat::single, relat::single, trans::unicast>> {
-
-    template <std::size_t DataSize, std::size_t AlignSize>
-    struct elem_t {
-        std::aligned_storage_t<DataSize, AlignSize> data_ {};
-    };
-
-    alignas(cache_line_size) std::atomic<circ::u2_t> rd_; // read index
-    alignas(cache_line_size) std::atomic<circ::u2_t> wt_; // write index
-
-    constexpr circ::u2_t cursor() const noexcept {
-        return 0;
-    }
-
-    template <typename W, typename F, typename E>
-    bool push(W* /*wrapper*/, F&& f, E* elems) {
-        auto cur_wt = circ::index_of(wt_.load(std::memory_order_relaxed));
-        if (cur_wt == circ::index_of(rd_.load(std::memory_order_acquire) - 1)) {
-            return false; // full
-        }
-        std::forward<F>(f)(&(elems[cur_wt].data_));
-        wt_.fetch_add(1, std::memory_order_release);
-        return true;
-    }
-
-    /**
-     * In single-single-unicast, 'force_push' means 'no reader' or 'the only one reader is dead'.
-     * So we could just disconnect all connections of receiver, and return false.
-    */
-    template <typename W, typename F, typename E>
-    bool force_push(W* wrapper, F&&, E*) {
-        wrapper->elems()->disconnect_receiver(~static_cast<circ::cc_t>(0u));
-        return false;
-    }
-
-    template <typename W, typename F, typename R, typename E>
-    bool pop(W* /*wrapper*/, circ::u2_t& /*cur*/, F&& f, R&& out, E* elems) {
-        auto cur_rd = circ::index_of(rd_.load(std::memory_order_relaxed));
-        if (cur_rd == circ::index_of(wt_.load(std::memory_order_acquire))) {
-            return false; // empty
-        }
-        std::forward<F>(f)(&(elems[cur_rd].data_));
-        std::forward<R>(out)(true);
-        rd_.fetch_add(1, std::memory_order_release);
-        return true;
-    }
-};
-
-template <>
-struct prod_cons_impl<wr<relat::single, relat::multi , trans::unicast>>
-     : prod_cons_impl<wr<relat::single, relat::single, trans::unicast>> {
-
-    template <typename W, typename F, typename E>
-    bool force_push(W* wrapper, F&&, E*) {
-        wrapper->elems()->disconnect_receiver(1);
-        return false;
-    }
-
-    template <typename W, typename F, typename R, 
-              template <std::size_t, std::size_t> class E, std::size_t DS, std::size_t AS>
-    bool pop(W* /*wrapper*/, circ::u2_t& /*cur*/, F&& f, R&& out, E<DS, AS>* elems) {
-        byte_t buff[DS];
-        for (unsigned k = 0;;) {
-            auto cur_rd = rd_.load(std::memory_order_relaxed);
-            if (circ::index_of(cur_rd) ==
-                circ::index_of(wt_.load(std::memory_order_acquire))) {
-                return false; // empty
-            }
-            std::memcpy(buff, &(elems[circ::index_of(cur_rd)].data_), sizeof(buff));
-            if (rd_.compare_exchange_weak(cur_rd, cur_rd + 1, std::memory_order_release)) {
-                std::forward<F>(f)(buff);
-                std::forward<R>(out)(true);
-                return true;
-            }
-            ipc::yield(k);
-        }
-    }
-};
-
-template <>
-struct prod_cons_impl<wr<relat::multi , relat::multi, trans::unicast>>
-     : prod_cons_impl<wr<relat::single, relat::multi, trans::unicast>> {
-
-    using flag_t = std::uint64_t;
-
-    template <std::size_t DataSize, std::size_t AlignSize>
-    struct elem_t {
-        std::aligned_storage_t<DataSize, AlignSize> data_ {};
-        std::atomic<flag_t> f_ct_ { 0 }; // commit flag
-    };
-
-    alignas(cache_line_size) std::atomic<circ::u2_t> ct_; // commit index
-
-    template <typename W, typename F, typename E>
-    bool push(W* /*wrapper*/, F&& f, E* elems) {
-        circ::u2_t cur_ct, nxt_ct;
-        for (unsigned k = 0;;) {
-            cur_ct = ct_.load(std::memory_order_relaxed);
-            if (circ::index_of(nxt_ct = cur_ct + 1) ==
-                circ::index_of(rd_.load(std::memory_order_acquire))) {
-                return false; // full
-            }
-            if (ct_.compare_exchange_weak(cur_ct, nxt_ct, std::memory_order_acq_rel)) {
-                break;
-            }
-            ipc::yield(k);
-        }
-        auto* el = elems + circ::index_of(cur_ct);
-        std::forward<F>(f)(&(el->data_));
-        // set flag & try update wt
-        el->f_ct_.store(~static_cast<flag_t>(cur_ct), std::memory_order_release);
-        while (1) {
-            auto cac_ct = el->f_ct_.load(std::memory_order_acquire);
-            if (cur_ct != wt_.load(std::memory_order_relaxed)) {
-                return true;
-            }
-            if ((~cac_ct) != cur_ct) {
-                return true;
-            }
-            if (!el->f_ct_.compare_exchange_strong(cac_ct, 0, std::memory_order_relaxed)) {
-                return true;
-            }
-            wt_.store(nxt_ct, std::memory_order_release);
-            cur_ct = nxt_ct;
-            nxt_ct = cur_ct + 1;
-            el = elems + circ::index_of(cur_ct);
-        }
-        return true;
-    }
-
-    template <typename W, typename F, typename E>
-    bool force_push(W* wrapper, F&&, E*) {
-        wrapper->elems()->disconnect_receiver(1);
-        return false;
-    }
-
-    template <typename W, typename F, typename R, 
-              template <std::size_t, std::size_t> class E, std::size_t DS, std::size_t AS>
-    bool pop(W* /*wrapper*/, circ::u2_t& /*cur*/, F&& f, R&& out, E<DS, AS>* elems) {
-        byte_t buff[DS];
-        for (unsigned k = 0;;) {
-            auto cur_rd = rd_.load(std::memory_order_relaxed);
-            auto cur_wt = wt_.load(std::memory_order_acquire);
-            auto id_rd  = circ::index_of(cur_rd);
-            auto id_wt  = circ::index_of(cur_wt);
-            if (id_rd == id_wt) {
-                auto* el = elems + id_wt;
-                auto cac_ct = el->f_ct_.load(std::memory_order_acquire);
-                if ((~cac_ct) != cur_wt) {
-                    return false; // empty
-                }
-                if (el->f_ct_.compare_exchange_weak(cac_ct, 0, std::memory_order_relaxed)) {
-                    wt_.store(cur_wt + 1, std::memory_order_release);
-                }
-                k = 0;
-            }
-            else {
-                std::memcpy(buff, &(elems[circ::index_of(cur_rd)].data_), sizeof(buff));
-                if (rd_.compare_exchange_weak(cur_rd, cur_rd + 1, std::memory_order_release)) {
-                    std::forward<F>(f)(buff);
-                    std::forward<R>(out)(true);
-                    return true;
-                }
-                ipc::yield(k);
-            }
-        }
-    }
-};
-
-template <>
-struct prod_cons_impl<wr<relat::single, relat::multi, trans::broadcast>> {
-
-    using rc_t = std::uint64_t;
-
-    enum : rc_t {
-        ep_mask = 0x00000000ffffffffull,
-        ep_incr = 0x0000000100000000ull
-    };
-
-    template <std::size_t DataSize, std::size_t AlignSize>
-    struct elem_t {
-        std::aligned_storage_t<DataSize, AlignSize> data_ {};
-        std::atomic<rc_t> rc_ { 0 }; // read-counter
-    };
-
-    alignas(cache_line_size) std::atomic<circ::u2_t> wt_;   // write index
-    alignas(cache_line_size) rc_t epoch_ { 0 };             // only one writer
-
-    circ::u2_t cursor() const noexcept {
-        return wt_.load(std::memory_order_acquire);
-    }
-
-    template <typename W, typename F, typename E>
-    bool push(W* wrapper, F&& f, E* elems) {
-        E* el;
-        for (unsigned k = 0;;) {
-            circ::cc_t cc = wrapper->elems()->connections(std::memory_order_relaxed);
-            if (cc == 0) return false; // no reader
-            el = elems + circ::index_of(wt_.load(std::memory_order_relaxed));
-            // check all consumers have finished reading this element
-            auto cur_rc = el->rc_.load(std::memory_order_acquire);
-            circ::cc_t rem_cc = cur_rc & ep_mask;
-            if ((cc & rem_cc) && ((cur_rc & ~ep_mask) == epoch_)) {
-                return false; // has not finished yet
-            }
-            // consider rem_cc to be 0 here
-            if (el->rc_.compare_exchange_weak(
-                        cur_rc, epoch_ | static_cast<rc_t>(cc), std::memory_order_release)) {
-                break;
-            }
-            ipc::yield(k);
-        }
-        std::forward<F>(f)(&(el->data_));
-        wt_.fetch_add(1, std::memory_order_release);
-        return true;
-    }
-
-    template <typename W, typename F, typename E>
-    bool force_push(W* wrapper, F&& f, E* elems) {
-        E* el;
-        epoch_ += ep_incr;
-        for (unsigned k = 0;;) {
-            circ::cc_t cc = wrapper->elems()->connections(std::memory_order_relaxed);
-            if (cc == 0) return false; // no reader
-            el = elems + circ::index_of(wt_.load(std::memory_order_relaxed));
-            // check all consumers have finished reading this element
-            auto cur_rc = el->rc_.load(std::memory_order_acquire);
-            circ::cc_t rem_cc = cur_rc & ep_mask;
-            if (cc & rem_cc) {
-                ipc::log("force_push: k = %u, cc = %u, rem_cc = %u\n", k, cc, rem_cc);
-                cc = wrapper->elems()->disconnect_receiver(rem_cc); // disconnect all invalid readers
-                if (cc == 0) return false; // no reader
-            }
-            // just compare & exchange
-            if (el->rc_.compare_exchange_weak(
-                        cur_rc, epoch_ | static_cast<rc_t>(cc), std::memory_order_release)) {
-                break;
-            }
-            ipc::yield(k);
-        }
-        std::forward<F>(f)(&(el->data_));
-        wt_.fetch_add(1, std::memory_order_release);
-        return true;
-    }
-
-    template <typename W, typename F, typename R, typename E>
-    bool pop(W* wrapper, circ::u2_t& cur, F&& f, R&& out, E* elems) {
-        if (cur == cursor()) return false; // acquire
-        auto* el = elems + circ::index_of(cur++);
-        std::forward<F>(f)(&(el->data_));
-        for (unsigned k = 0;;) {
-            auto cur_rc = el->rc_.load(std::memory_order_acquire);
-            if ((cur_rc & ep_mask) == 0) {
-                std::forward<R>(out)(true);
-                return true;
-            }
-            auto nxt_rc = cur_rc & ~static_cast<rc_t>(wrapper->connected_id());
-            if (el->rc_.compare_exchange_weak(cur_rc, nxt_rc, std::memory_order_release)) {
-                std::forward<R>(out)((nxt_rc & ep_mask) == 0);
-                return true;
-            }
-            ipc::yield(k);
-        }
-    }
-};
-
-template <>
-struct prod_cons_impl<wr<relat::multi, relat::multi, trans::broadcast>> {
-
-    using rc_t   = std::uint64_t;
-    using flag_t = std::uint64_t;
-
-    enum : rc_t {
-        rc_mask = 0x00000000ffffffffull,
-        ep_mask = 0x00ffffffffffffffull,
-        ep_incr = 0x0100000000000000ull,
-        ic_mask = 0xff000000ffffffffull,
-        ic_incr = 0x0000000100000000ull
-    };
-
-    template <std::size_t DataSize, std::size_t AlignSize>
-    struct elem_t {
-        std::aligned_storage_t<DataSize, AlignSize> data_ {};
-        std::atomic<rc_t  > rc_   { 0 }; // read-counter
-        std::atomic<flag_t> f_ct_ { 0 }; // commit flag
-    };
-
-    alignas(cache_line_size) std::atomic<circ::u2_t> ct_;   // commit index
-    alignas(cache_line_size) std::atomic<rc_t> epoch_ { 0 };
-
-    circ::u2_t cursor() const noexcept {
-        return ct_.load(std::memory_order_acquire);
-    }
-
-    constexpr static rc_t inc_rc(rc_t rc) noexcept {
-        return (rc & ic_mask) | ((rc + ic_incr) & ~ic_mask);
-    }
-
-    constexpr static rc_t inc_mask(rc_t rc) noexcept {
-        return inc_rc(rc) & ~rc_mask;
-    }
-
-    template <typename W, typename F, typename E>
-    bool push(W* wrapper, F&& f, E* elems) {
-        E* el;
-        circ::u2_t cur_ct;
-        rc_t epoch = epoch_.load(std::memory_order_acquire);
-        for (unsigned k = 0;;) {
-            circ::cc_t cc = wrapper->elems()->connections(std::memory_order_relaxed);
-            if (cc == 0) return false; // no reader
-            el = elems + circ::index_of(cur_ct = ct_.load(std::memory_order_relaxed));
-            // check all consumers have finished reading this element
-            auto cur_rc = el->rc_.load(std::memory_order_relaxed);
-            circ::cc_t rem_cc = cur_rc & rc_mask;
-            if ((cc & rem_cc) && ((cur_rc & ~ep_mask) == epoch)) {
-                return false; // has not finished yet
-            }
-            else if (!rem_cc) {
-                auto cur_fl = el->f_ct_.load(std::memory_order_acquire);
-                if ((cur_fl != cur_ct) && cur_fl) {
-                    return false; // full
-                }
-            }
-            // consider rem_cc to be 0 here
-            if (el->rc_.compare_exchange_weak(
-                        cur_rc, inc_mask(epoch | (cur_rc & ep_mask)) | static_cast<rc_t>(cc), std::memory_order_relaxed) &&
-                epoch_.compare_exchange_weak(epoch, epoch, std::memory_order_acq_rel)) {
-                break;
-            }
-            ipc::yield(k);
-        }
-        // only one thread/process would touch here at one time
-        ct_.store(cur_ct + 1, std::memory_order_release);
-        std::forward<F>(f)(&(el->data_));
-        // set flag & try update wt
-        el->f_ct_.store(~static_cast<flag_t>(cur_ct), std::memory_order_release);
-        return true;
-    }
-
-    template <typename W, typename F, typename E>
-    bool force_push(W* wrapper, F&& f, E* elems) {
-        E* el;
-        circ::u2_t cur_ct;
-        rc_t epoch = epoch_.fetch_add(ep_incr, std::memory_order_release) + ep_incr;
-        for (unsigned k = 0;;) {
-            circ::cc_t cc = wrapper->elems()->connections(std::memory_order_relaxed);
-            if (cc == 0) return false; // no reader
-            el = elems + circ::index_of(cur_ct = ct_.load(std::memory_order_relaxed));
-            // check all consumers have finished reading this element
-            auto cur_rc = el->rc_.load(std::memory_order_acquire);
-            circ::cc_t rem_cc = cur_rc & rc_mask;
-            if (cc & rem_cc) {
-                ipc::log("force_push: k = %u, cc = %u, rem_cc = %u\n", k, cc, rem_cc);
-                cc = wrapper->elems()->disconnect_receiver(rem_cc); // disconnect all invalid readers
-                if (cc == 0) return false; // no reader
-            }
-            // just compare & exchange
-            if (el->rc_.compare_exchange_weak(
-                        cur_rc, inc_mask(epoch | (cur_rc & ep_mask)) | static_cast<rc_t>(cc), std::memory_order_relaxed)) {
-                if (epoch == epoch_.load(std::memory_order_acquire)) {
-                    break;
-                }
-                else if (push(wrapper, std::forward<F>(f), elems)) {
-                    return true;
-                }
-                epoch = epoch_.fetch_add(ep_incr, std::memory_order_release) + ep_incr;
-            }
-            ipc::yield(k);
-        }
-        // only one thread/process would touch here at one time
-        ct_.store(cur_ct + 1, std::memory_order_release);
-        std::forward<F>(f)(&(el->data_));
-        // set flag & try update wt
-        el->f_ct_.store(~static_cast<flag_t>(cur_ct), std::memory_order_release);
-        return true;
-    }
-
-    template <typename W, typename F, typename R, typename E, std::size_t N>
-    bool pop(W* wrapper, circ::u2_t& cur, F&& f, R&& out, E(& elems)[N]) {
-        auto* el = elems + circ::index_of(cur);
-        auto cur_fl = el->f_ct_.load(std::memory_order_acquire);
-        if (cur_fl != ~static_cast<flag_t>(cur)) {
-            return false; // empty
-        }
-        ++cur;
-        std::forward<F>(f)(&(el->data_));
-        for (unsigned k = 0;;) {
-            auto cur_rc = el->rc_.load(std::memory_order_acquire);
-            if ((cur_rc & rc_mask) == 0) {
-                std::forward<R>(out)(true);
-                el->f_ct_.store(cur + N - 1, std::memory_order_release);
-                return true;
-            }
-            auto nxt_rc = inc_rc(cur_rc) & ~static_cast<rc_t>(wrapper->connected_id());
-            bool last_one = false;
-            if ((last_one = (nxt_rc & rc_mask) == 0)) {
-                el->f_ct_.store(cur + N - 1, std::memory_order_release);
-            }
-            if (el->rc_.compare_exchange_weak(cur_rc, nxt_rc, std::memory_order_release)) {
-                std::forward<R>(out)(last_one);
-                return true;
-            }
-            ipc::yield(k);
-        }
-    }
-};
-
-} // namespace ipc

crazy_functions/test_project/latex/attention/background.tex

@@ -1,58 +0,0 @@
-The goal of reducing sequential computation also forms the foundation of the Extended Neural GPU \citep{extendedngpu}, ByteNet \citep{NalBytenet2017} and ConvS2S \citep{JonasFaceNet2017}, all of which use convolutional neural networks as basic building block, computing hidden representations in parallel for all input and output positions. In these models, the number of operations required to relate signals from two arbitrary input or output positions grows in the distance between positions, linearly for ConvS2S and logarithmically for ByteNet. This makes it more difficult to learn dependencies between distant positions \citep{hochreiter2001gradient}. In the Transformer this is reduced to a constant number of operations, albeit at the cost of reduced effective resolution due to averaging attention-weighted positions, an effect we counteract with Multi-Head Attention as described in section~\ref{sec:attention}. 
-
-Self-attention, sometimes called intra-attention is an attention mechanism relating different positions of a single sequence in order to compute a representation of the sequence. Self-attention has been used successfully in a variety of tasks including reading comprehension, abstractive summarization, textual entailment and learning task-independent sentence representations \citep{cheng2016long, decomposableAttnModel, paulus2017deep, lin2017structured}.
-
-End-to-end memory networks are based on a recurrent attention mechanism instead of sequence-aligned recurrence and have been shown to perform well on simple-language question answering and language modeling tasks \citep{sukhbaatar2015}.
-
-To the best of our knowledge, however, the Transformer is the first transduction model relying entirely on self-attention to compute representations of its input and output without using sequence-aligned RNNs or convolution.
-In the following sections, we will describe the Transformer, motivate self-attention and discuss its advantages over models such as \citep{neural_gpu, NalBytenet2017} and \citep{JonasFaceNet2017}.
-
-
-%\citep{JonasFaceNet2017} report new SOTA on machine translation for English-to-German (EnDe), Enlish-to-French (EnFr) and English-to-Romanian language pairs. 
-
-%For example,! in MT, we must draw information from both input and previous output words to translate an output word accurately. An attention layer \citep{bahdanau2014neural} can connect a very large number of positions at low computation cost, making it an essential ingredient in competitive recurrent models for machine translation.
-
-%A natural question to ask then is, "Could we replace recurrence with attention?". \marginpar{Don't know if it's the most natural question to ask given the previous statements. Also, need to say that the complexity table summarizes these statements} Such a model would be blessed with the computational efficiency of attention and the power of cross-positional communication. In this work, show that pure attention models work remarkably well for MT, achieving new SOTA results on EnDe and EnFr, and can be trained in under $2$ days on xyz architecture. 
-
-%After the seminal models introduced in \citep{sutskever14, bahdanau2014neural, cho2014learning}, recurrent models have become the dominant solution for both sequence modeling and sequence-to-sequence transduction. Many efforts such as \citep{wu2016google,luong2015effective,jozefowicz2016exploring} have pushed the boundaries of machine translation (MT) and language modeling with recurrent endoder-decoder and recurrent language models. Recent effort \citep{shazeer2017outrageously} has successfully combined the power of conditional computation with sequence models to train very large models for MT, pushing SOTA at lower computational cost.
-
-%Recurrent models compute a vector of hidden states $h_t$, for each time step $t$ of computation. $h_t$ is a function of both the input at time $t$ and the previous hidden state $h_t$. This dependence on the previous hidden state precludes processing all timesteps at once, instead requiring long sequences of sequential operations.  In practice, this results in greatly reduced computational efficiency, as on modern computing hardware, a single operation on a large batch is much faster than a large number of operations on small batches.  The problem gets worse at longer sequence lengths. Although sequential computation is not a severe bottleneck at inference time, as autoregressively generating each output requires all previous outputs, the inability to compute scores at all output positions at once hinders us from rapidly training our models over large datasets. Although impressive work such as \citep{Kuchaiev2017Factorization} is able to significantly accelerate the training of LSTMs with factorization tricks, we are still bound by the linear dependence on sequence length.
-
-%If the model could compute hidden states at each time step using only the inputs and outputs,  it would be liberated from the dependence on results from previous time steps during training. This line of thought is the foundation of recent efforts such as the Markovian neural GPU \citep{neural_gpu}, ByteNet \citep{NalBytenet2017} and ConvS2S \citep{JonasFaceNet2017}, all of which use convolutional neural networks as a building block to compute hidden representations simultaneously for all timesteps, resulting in $O(1)$ sequential time complexity. \citep{JonasFaceNet2017} report new SOTA on machine translation for English-to-German (EnDe), Enlish-to-French (EnFr) and English-to-Romanian language pairs. 
-
-%A crucial component for accurate sequence prediction is modeling cross-positional communication. For example, in MT, we must draw information from both input and previous output words to translate an output word accurately. An attention layer \citep{bahdanau2014neural} can connect a very large number of positions at a low computation cost, also $O(1)$ sequential time complexity, making it an essential ingredient in recurrent encoder-decoder architectures for MT. A natural question to ask then is, "Could we replace recurrence with attention?". \marginpar{Don't know if it's the most natural question to ask given the previous statements. Also, need to say that the complexity table summarizes these statements} Such a model would be blessed with the computational efficiency of attention and the power of cross-positional communication. In this work, show that pure attention models work remarkably well for MT, achieving new SOTA results on EnDe and EnFr, and can be trained in under $2$ days on xyz architecture. 
-
-
-
-%Note: Facebook model is no better than RNNs in this regard, since it requires a number of layers proportional to the distance you want to communicate.  Bytenet is more promising, since it requires a logarithmnic number of layers (does bytenet have SOTA results)?   
-
-%Note: An attention  layer can connect a very large number of positions at a low computation cost in O(1) sequential operations.  This is why encoder-decoder attention has been so successful in seq-to-seq models so far.  It is only natural, then, to also use attention to connect the timesteps of the same sequence.
-
-%Note: I wouldn't say that long sequences are not a problem during inference.  It would be great if we could infer with no long sequences.  We could just say later on that, while our training graph is constant-depth, our model still requires sequential operations in the decoder part during inference due to the autoregressive nature of the model.   
-
-%\begin{table}[h!]
-%\caption{Attention models are quite efficient for cross-positional communications when sequence length is smaller than channel depth. $n$ represents the sequence length and $d$ represents the channel depth.}
-%\label{tab:op_complexities}
-%\begin{center}
-%\vspace{-5pt}
-%\scalebox{0.75}{
-
-%\begin{tabular}{l|c|c|c}
-%\hline \hline
-%Layer Type & Receptive & Complexity & Sequential  \\
-%           & Field     &            & Operations  \\
-%\hline
-%Pointwise Feed-Forward & $1$ & $O(n \cdot d^2)$ & $O(1)$ \\
-%\hline
-%Recurrent & $n$ & $O(n \cdot d^2)$ & $O(n)$ \\
-%\hline
-%Convolutional & $r$ & $O(r \cdot n \cdot d^2)$ & $O(1)$ \\
-%\hline
-%Convolutional (separable) & $r$ & $O(r \cdot n \cdot d + n %\cdot d^2)$ & $O(1)$ \\
-%\hline
-%Attention & $r$ & $O(r \cdot n \cdot d)$ & $O(1)$ \\
-%\hline \hline
-%\end{tabular}
-%}
-%\end{center}
-%\end{table}

crazy_functions/test_project/latex/attention/introduction.tex

@@ -1,18 +0,0 @@
-Recurrent neural networks, long short-term memory \citep{hochreiter1997} and gated recurrent \citep{gruEval14} neural networks in particular, have been firmly established as state of the art approaches in sequence modeling and transduction problems such as language modeling and machine translation \citep{sutskever14, bahdanau2014neural, cho2014learning}. Numerous efforts have since continued to push the boundaries of recurrent language models and encoder-decoder architectures \citep{wu2016google,luong2015effective,jozefowicz2016exploring}.
-
-Recurrent models typically factor computation along the symbol positions of the input and output sequences. Aligning the positions to steps in computation time, they generate a sequence of hidden states $h_t$, as a function of the previous hidden state $h_{t-1}$ and the input for position $t$. This inherently sequential nature precludes parallelization within training examples, which becomes critical at longer sequence lengths, as memory constraints limit batching across examples.
-%\marginpar{not sure if the memory constraints are understandable here}
-Recent work has achieved significant improvements in computational efficiency through factorization tricks \citep{Kuchaiev2017Factorization} and conditional computation \citep{shazeer2017outrageously}, while also improving model performance in case of the latter. The fundamental constraint of sequential computation, however, remains.
-
-%\marginpar{@all: there is work on analyzing what attention really does in seq2seq models, couldn't find it right away} 
-
-Attention mechanisms have become an integral part of compelling sequence modeling and transduction models in various tasks, allowing modeling of dependencies without regard to their distance in the input or output sequences \citep{bahdanau2014neural, structuredAttentionNetworks}. In all but a few cases \citep{decomposableAttnModel}, however, such attention mechanisms are used in conjunction with a recurrent network.
-
-%\marginpar{not sure if "cross-positional communication" is understandable without explanation}
-%\marginpar{insert exact training times and stats for the model that reaches sota earliest, maybe even a single GPU model?}
-
-In this work we propose the Transformer, a model architecture eschewing recurrence and instead relying entirely on an attention mechanism to draw global dependencies between input and output. The Transformer allows for significantly more parallelization and can reach a new state of the art in translation quality after being trained for as little as twelve hours on eight P100 GPUs. 
-%\marginpar{you removed the constant number of repetitions part. I wrote it because I wanted to make it clear that the model does not only perform attention once, while it's also not recurrent. I thought that might be important to get across early.}
-
-% Just a standard paragraph with citations, rewrite.
-%After the seminal papers of \citep{sutskever14}, \citep{bahdanau2014neural}, and \citep{cho2014learning}, recurrent models have become the dominant solution for both sequence modeling and sequence-to-sequence transduction. Many efforts such as \citep{wu2016google,luong2015effective,jozefowicz2016exploring} have pushed the boundaries of machine translation and language modeling with recurrent sequence models. Recent effort \citep{shazeer2017outrageously} has combined the power of conditional computation with sequence models to train very large models for machine translation, pushing SOTA at lower computational cost. Recurrent models compute a vector of hidden states $h_t$, for each time step $t$ of computation. $h_t$ is a function of both the input at time $t$ and the previous hidden state $h_t$. This dependence on the previous hidden state encumbers recurrnet models to process multiple inputs at once, and their time complexity is a linear function of the length of the input and output, both during training and inference. [What I want to say here is that although this is fine during decoding, at training time, we are given both input and output and this linear nature does not allow the RNN to process all inputs and outputs simultaneously and haven't been used on datasets that are the of the scale of the web. What's the largest dataset we have ? . Talk about Nividia and possibly other's effors to speed up things, and possibly other efforts that alleviate this, but are still limited by it's comptuational nature]. Rest of the intro: What if you could construct the state based on the actual inputs and outputs, then you could construct them all at once. This has been the foundation of many promising recent efforts, bytenet,facenet (Also talk about quasi rnn here). Now we talk about attention!! Along with cell architectures such as long short-term meory (LSTM) \citep{hochreiter1997}, and gated recurrent units (GRUs) \citep{cho2014learning}, attention has emerged as an essential ingredient in successful sequence models, in particular for machine translation. In recent years, many, if not all, state-of-the-art (SOTA) results in machine translation have been achieved with attention-based sequence models \citep{wu2016google,luong2015effective,jozefowicz2016exploring}. Talk about the neon work on how it played with attention to do self attention! Then talk about what we do.

crazy_functions/test_project/latex/attention/model_architecture.tex

@@ -1,155 +0,0 @@
-
-\begin{figure}
-  \centering
-  \includegraphics[scale=0.6]{Figures/ModalNet-21}
-  \caption{The Transformer - model architecture.}
-  \label{fig:model-arch}
-\end{figure}
-
-% Although the primary workhorse of our model is attention, 
-%Our model maintains the encoder-decoder structure that is common to many so-called sequence-to-sequence models \citep{bahdanau2014neural,sutskever14}.  As in all such architectures, the encoder computes a representation of the input sequence, and the decoder consumes these representations along with the output tokens to autoregressively produce the output sequence.  Where, traditionally, the encoder and decoder contain stacks of recurrent or convolutional layers, our encoder and decoder stacks are composed of attention layers and position-wise feed-forward layers (Figure~\ref{fig:model-arch}).  The following sections describe the gross architecture and these particular components in detail.
-
-Most competitive neural sequence transduction models have an encoder-decoder structure \citep{cho2014learning,bahdanau2014neural,sutskever14}. Here, the encoder maps an input sequence of symbol representations $(x_1, ..., x_n)$ to a sequence of continuous representations $\mathbf{z} = (z_1, ..., z_n)$. Given $\mathbf{z}$, the decoder then generates an output sequence $(y_1,...,y_m)$ of symbols one element at a time. At each step the model is auto-regressive \citep{graves2013generating}, consuming the previously generated symbols as additional input when generating the next.
-
-The Transformer follows this overall architecture using stacked self-attention and point-wise, fully connected layers for both the encoder and decoder, shown in the left and right halves of Figure~\ref{fig:model-arch}, respectively.
-
-\subsection{Encoder and Decoder Stacks}
-
-\paragraph{Encoder:}The encoder is composed of a stack of $N=6$ identical layers. Each layer has two sub-layers. The first is a multi-head self-attention mechanism, and the second is a simple, position-wise fully connected feed-forward network.   We employ a residual connection \citep{he2016deep} around each of the two sub-layers, followed by layer normalization \cite{layernorm2016}.  That is, the output of each sub-layer is $\mathrm{LayerNorm}(x + \mathrm{Sublayer}(x))$, where $\mathrm{Sublayer}(x)$ is the function implemented by the sub-layer itself.  To facilitate these residual connections, all sub-layers in the model, as well as the embedding layers, produce outputs of dimension $\dmodel=512$.
-
-\paragraph{Decoder:}The decoder is also composed of a stack of $N=6$ identical layers.  In addition to the two sub-layers in each encoder layer, the decoder inserts a third sub-layer, which performs multi-head attention over the output of the encoder stack.  Similar to the encoder, we employ residual connections around each of the sub-layers, followed by layer normalization.  We also modify the self-attention sub-layer in the decoder stack to prevent positions from attending to subsequent positions.  This masking, combined with fact that the output embeddings are offset by one position, ensures that the predictions for position $i$ can depend only on the known outputs at positions less than $i$.
-
-% In our model (Figure~\ref{fig:model-arch}), the encoder and decoder are composed of stacks of alternating self-attention layers (for cross-positional communication) and position-wise feed-forward layers (for in-place computation).  In addition, the decoder stack contains encoder-decoder attention layers.  Since attention is agnostic to the distances between words, our model requires a "positional encoding" to be added to the encoder and decoder input. The following sections describe all of these components in detail.
-
-\subsection{Attention} \label{sec:attention}
-An attention function can be described as mapping a query and a set of key-value pairs to an output, where the query, keys, values, and output are all vectors.  The output is computed as a weighted sum of the values, where the weight assigned to each value is computed by a compatibility function of the query with the corresponding key.
-
-\subsubsection{Scaled Dot-Product Attention} \label{sec:scaled-dot-prod}
-
-% \begin{figure}
-%   \centering
-%   \includegraphics[scale=0.6]{Figures/ModalNet-19}
-%   \caption{Scaled Dot-Product Attention.}
-%   \label{fig:multi-head-att}
-% \end{figure}
-
-We call our particular attention "Scaled Dot-Product Attention" (Figure~\ref{fig:multi-head-att}).   The input consists of queries and keys of dimension $d_k$, and values of dimension $d_v$.  We compute the dot products of the query with all keys, divide each by $\sqrt{d_k}$, and apply a softmax function to obtain the weights on the values.
-
-In practice, we compute the attention function on a set of queries simultaneously, packed together into a matrix $Q$.   The keys and values are also packed together into matrices $K$ and $V$.  We compute the matrix of outputs as:
-
-\begin{equation}
-   \mathrm{Attention}(Q, K, V) = \mathrm{softmax}(\frac{QK^T}{\sqrt{d_k}})V
-\end{equation}
-
-The two most commonly used attention functions are additive attention \citep{bahdanau2014neural}, and dot-product (multiplicative) attention.  Dot-product attention is identical to our algorithm, except for the scaling factor of $\frac{1}{\sqrt{d_k}}$. Additive attention computes the compatibility function using a feed-forward network with a single hidden layer.  While the two are similar in theoretical complexity, dot-product attention is much faster and more space-efficient in practice, since it can be implemented using highly optimized matrix multiplication code. 
-
-%We scale the dot products by $1/\sqrt{d_k}$ to limit the magnitude of the dot products, which works well in practice. Otherwise, we found applying the softmax to often result in weights very close to 0 or 1, and hence minuscule gradients.
-
-% Already described in the subsequent section
-%When used as part of decoder self-attention, an optional mask function is applied just before the softmax to prevent positions from attending to subsequent positions.   This mask simply sets the logits corresponding to all illegal connections (those outside of the lower triangle) to $-\infty$.
-
-%\paragraph{Comparison to Additive Attention: } We choose dot product attention over additive attention \citep{bahdanau2014neural} since it can be computed using highly optimized matrix multiplication code.  This optimization is particularly important to us, as we employ many attention layers in our model.
-
-While for small values of $d_k$ the two mechanisms perform similarly, additive attention outperforms dot product attention without scaling for larger values of $d_k$ \citep{DBLP:journals/corr/BritzGLL17}. We suspect that for large values of $d_k$, the dot products grow large in magnitude, pushing the softmax function into regions where it has extremely small gradients  \footnote{To illustrate why the dot products get large, assume that the components of $q$ and $k$ are independent random variables with mean $0$ and variance $1$.  Then their dot product, $q \cdot k = \sum_{i=1}^{d_k} q_ik_i$, has mean $0$ and variance $d_k$.}. To counteract this effect, we scale the dot products by $\frac{1}{\sqrt{d_k}}$.
-
-
-%We suspect this to be caused by the dot products growing too large in magnitude to result in useful gradients after applying the softmax function.  To counteract this, we scale the dot product by $1/\sqrt{d_k}$.
-
-
-\subsubsection{Multi-Head Attention} \label{sec:multihead}
-
-\begin{figure}
-\begin{minipage}[t]{0.5\textwidth}
-  \centering
-  Scaled Dot-Product Attention \\
-  \vspace{0.5cm}
-  \includegraphics[scale=0.6]{Figures/ModalNet-19}
-\end{minipage}
-\begin{minipage}[t]{0.5\textwidth}
-  \centering 
-  Multi-Head Attention \\
-  \vspace{0.1cm}
-  \includegraphics[scale=0.6]{Figures/ModalNet-20}  
-\end{minipage}
-
-
-  % \centering
-
-  \caption{(left) Scaled Dot-Product Attention. (right) Multi-Head Attention consists of several attention layers running in parallel.}
-  \label{fig:multi-head-att}
-\end{figure}
-
-Instead of performing a single attention function with $\dmodel$-dimensional keys, values and queries, we found it beneficial to linearly project the queries, keys and values $h$ times with different, learned linear projections to $d_k$, $d_k$ and $d_v$ dimensions, respectively.
-On each of these projected versions of queries, keys and values we then perform the attention function in parallel, yielding $d_v$-dimensional output values. These are concatenated and once again projected, resulting in the final values, as depicted in Figure~\ref{fig:multi-head-att}.
-
-Multi-head attention allows the model to jointly attend to information from different representation subspaces at different positions. With a single attention head, averaging inhibits this.
-
-\begin{align*}
-    \mathrm{MultiHead}(Q, K, V) &= \mathrm{Concat}(\mathrm{head_1}, ..., \mathrm{head_h})W^O\\
-%    \mathrm{where} \mathrm{head_i} &= \mathrm{Attention}(QW_Q_i^{\dmodel \times d_q}, KW_K_i^{\dmodel \times d_k}, VW^V_i^{\dmodel \times d_v})\\
-    \text{where}~\mathrm{head_i} &= \mathrm{Attention}(QW^Q_i, KW^K_i, VW^V_i)\\
-\end{align*}
-
-Where the projections are parameter matrices $W^Q_i \in \mathbb{R}^{\dmodel \times d_k}$, $W^K_i \in \mathbb{R}^{\dmodel \times d_k}$, $W^V_i \in \mathbb{R}^{\dmodel \times d_v}$ and $W^O \in \mathbb{R}^{hd_v \times \dmodel}$.
-
-
-%find it better (and no more expensive) to have multiple parallel attention layers (each over the full set of positions) with proportionally lower-dimensional keys, values and queries.  We call this "Multi-Head Attention" (Figure~\ref{fig:multi-head-att}).  The keys, values, and queries for each of these parallel attention layers are computed by learned linear transformations of the inputs to the multi-head attention.  We use different linear transformations across different parallel attention layers.  The output of the parallel attention layers are concatenated, and then passed through a final learned linear transformation. 
-
-In this work we employ $h=8$ parallel attention layers, or heads. For each of these we use $d_k=d_v=\dmodel/h=64$.
-Due to the reduced dimension of each head, the total computational cost is similar to that of single-head attention with full dimensionality.
-
-\subsubsection{Applications of Attention in our Model}
-
-The Transformer uses multi-head attention in three different ways: 
-\begin{itemize}
- \item In "encoder-decoder attention" layers, the queries come from the previous decoder layer, and the memory keys and values come from the output of the encoder.   This allows every position in the decoder to attend over all positions in the input sequence.  This mimics the typical encoder-decoder attention mechanisms in sequence-to-sequence models such as \citep{wu2016google, bahdanau2014neural,JonasFaceNet2017}.
-
- \item The encoder contains self-attention layers.  In a self-attention layer all of the keys, values and queries come from the same place, in this case, the output of the previous layer in the encoder.   Each position in the encoder can attend to all positions in the previous layer of the encoder.
-
- \item Similarly, self-attention layers in the decoder allow each position in the decoder to attend to all positions in the decoder up to and including that position.  We need to prevent leftward information flow in the decoder to preserve the auto-regressive property.  We implement this inside of scaled dot-product attention by masking out (setting to $-\infty$) all values in the input of the softmax which correspond to illegal connections.  See Figure~\ref{fig:multi-head-att}.
-
-\end{itemize}
-
-\subsection{Position-wise Feed-Forward Networks}\label{sec:ffn}
-
-In addition to attention sub-layers, each of the layers in our encoder and decoder contains a fully connected feed-forward network, which is applied to each position separately and identically.  This consists of two linear transformations with a ReLU activation in between.
-
-\begin{equation}
-   \mathrm{FFN}(x)=\max(0, xW_1 + b_1) W_2 + b_2
-\end{equation}
-
-While the linear transformations are the same across different positions, they use different parameters from layer to layer. Another way of describing this is as two convolutions with kernel size 1.  The dimensionality of input and output is $\dmodel=512$, and the inner-layer has dimensionality $d_{ff}=2048$.
-
-
-
-%In the appendix, we describe how the position-wise feed-forward network can also be seen as a form of attention.
-
-%from Jakob: The number of operations required for the model to relate signals from two arbitrary input or output positions grows in the distance between positions in input or output, linearly for ConvS2S and logarithmically for ByteNet, making it harder to learn dependencies between these positions \citep{hochreiter2001gradient}. In the transformer this is reduced to a constant number of operations, albeit at the cost of effective resolution caused by averaging attention-weighted positions, an effect we aim to counteract with multi-headed attention.
-
-
-%Figure~\ref{fig:simple-att} presents a simple attention function, $A$, with a single head, that forms the basis of our multi-head attention. $A$ takes a query key vector $\kq$, matrices of memory keys $\km$ and memory values $\vm$ ,and produces a query value vector $\vq$ as 
-%\begin{equation*} \label{eq:attention}
-%    A(\kq, \km, \vm) = {\vm}^T (Softmax(\km \kq).
-%\end{equation*}
-%We linearly transform $\kq,\,\km$, and $\vm$ with learned matrices ${\Wkq \text{,} \, \Wkm}$, and ${\Wvm}$ before calling the attention function, and transform the output query with $\Wvq$ before handing it to the feed forward layer. Each attention layer has it's own set of transformation matrices, which are shared across all query positions. $A$ is applied in parallel for each query position, and is implemented very efficiently as a batch of matrix multiplies. The self-attention and encoder-decoder attention layers use $A$, but with different arguments. For example, in encdoder self-attention, queries in encoder layer $i$ attention to memories in encoder layer $i-1$. To ensure that decoder self-attention layers do not look at future words, we add $- \inf$ to the softmax logits in positions $j+1$ to query length for query position $l$.  
-
-%In simple attention, the query value is a weighted combination of the memory values where the attention weights sum to one. Although this function performs well in practice, the constraint on attention weights can restrict the amount of information that flows from memories to queries because the query cannot focus on multiple memory positions at once, which might be desirable when translating long sequences. \marginpar{@usz, could you think of an example of this ?} We remedy this by maintaining multiple attention heads at each query position that attend to all memory positions in parallel, with a different set of parameters  per attention head $h$. 
-%\marginpar{}
-
-\subsection{Embeddings and Softmax}
-Similarly to other sequence transduction models, we use learned embeddings to convert the input tokens and output tokens to vectors of dimension $\dmodel$.  We also use the usual learned linear transformation and softmax function to convert the decoder output to predicted next-token probabilities.  In our model, we share the same weight matrix between the two embedding layers and the pre-softmax linear transformation, similar to \citep{press2016using}.   In the embedding layers, we multiply those weights by $\sqrt{\dmodel}$.
-
-
-\subsection{Positional Encoding}
-Since our model contains no recurrence and no convolution, in order for the model to make use of the order of the sequence, we must inject some information about the relative or absolute position of the tokens in the sequence.  To this end, we add "positional encodings" to the input embeddings at the bottoms of the encoder and decoder stacks.  The positional encodings have the same dimension $\dmodel$ as the embeddings, so that the two can be summed.   There are many choices of positional encodings, learned and fixed \citep{JonasFaceNet2017}.
-
-In this work, we use sine and cosine functions of different frequencies:
-
-\begin{align*}
-    PE_{(pos,2i)} = sin(pos / 10000^{2i/\dmodel}) \\
-    PE_{(pos,2i+1)} = cos(pos / 10000^{2i/\dmodel})
-\end{align*}
-
-where $pos$ is the position and $i$ is the dimension.  That is, each dimension of the positional encoding corresponds to a sinusoid.  The wavelengths form a geometric progression from $2\pi$ to $10000 \cdot 2\pi$.  We chose this function because we hypothesized it would allow the model to easily learn to attend by relative positions, since for any fixed offset $k$, $PE_{pos+k}$ can be represented as a linear function of $PE_{pos}$.
-
-We also experimented with using learned positional embeddings \citep{JonasFaceNet2017} instead, and found that the two versions produced nearly identical results (see Table~\ref{tab:variations} row (E)).  We chose the sinusoidal version because it may allow the model to extrapolate to sequence lengths longer than the ones encountered during training.

crazy_functions/test_project/latex/attention/parameter_attention.tex

@@ -1,45 +0,0 @@
-\pagebreak
-\section*{Two Feed-Forward Layers = Attention over Parameters}\label{sec:parameter_attention}
-
-In addition to attention layers, our model contains position-wise feed-forward networks (Section \ref{sec:ffn}), which consist of two linear transformations with a ReLU activation in between.  In fact, these networks too can be seen as a form of attention.  Compare the formula for such a network with the formula for a simple dot-product attention layer (biases and scaling factors omitted):
-
-\begin{align*}
-    FFN(x, W_1, W_2) = ReLU(xW_1)W_2 \\
-    A(q, K, V) = Softmax(qK^T)V
-\end{align*}
-
-Based on the similarity of these formulae, the two-layer feed-forward network can be seen as a kind of attention, where the keys and values are the rows of the trainable parameter matrices $W_1$ and $W_2$, and where we use ReLU instead of Softmax in the compatibility function.
-
-%the compatablity function is $compat(q, k_i) = ReLU(q \cdot k_i)$ instead of $Softmax(qK_T)_i$.
-
-Given this similarity, we experimented with replacing the position-wise feed-forward networks with attention layers similar to the ones we use everywhere else our model. The multi-head-attention-over-parameters sublayer is identical to the multi-head attention described in \ref{sec:multihead}, except that the "keys" and "values" inputs to each attention head are trainable model parameters, as opposed to being linear projections of a previous layer.  These parameters are scaled up by a factor of $\sqrt{d_{model}}$ in order to be more similar to activations.
-
-In our first experiment, we replaced each position-wise feed-forward network with a multi-head-attention-over-parameters sublayer with $h_p=8$ heads, key-dimensionality $d_{pk}=64$, and value-dimensionality $d_{pv}=64$, using $n_p=1536$ key-value pairs for each attention head.  The sublayer has a total of $2097152$ parameters, including the parameters in the query projection and the output projection.  This matches the number of parameters in the position-wise feed-forward network that we replaced.  While the theoretical amount of computation is also the same, in practice, the attention version caused the step times to be about 30\% longer.
-
-In our second experiment, we used $h_p=8$ heads, and $n_p=512$ key-value pairs for each attention head, again matching the total number of parameters in the base model.
-
-Results for the first experiment were slightly worse than for the base model, and results for the second experiment were slightly better, see Table~\ref{tab:parameter_attention}.
-
-\begin{table}[h]
-\caption{Replacing the position-wise feed-forward networks with multihead-attention-over-parameters produces similar results to the base model.  All metrics are on the English-to-German translation development set, newstest2013.}
-\label{tab:parameter_attention}
-\begin{center}
-\vspace{-2mm}
-%\scalebox{1.0}{
-\begin{tabular}{c|cccccc|cccc}
-\hline\rule{0pt}{2.0ex}
- & \multirow{2}{*}{$\dmodel$} & \multirow{2}{*}{$\dff$} &
-\multirow{2}{*}{$h_p$} & \multirow{2}{*}{$d_{pk}$} & \multirow{2}{*}{$d_{pv}$} &
- \multirow{2}{*}{$n_p$} &
- PPL & BLEU & params & training\\
- & & & & & &  & (dev) & (dev) & $\times10^6$ & time \\
-\hline\rule{0pt}{2.0ex}
-base & 512 & 2048 & & & & & 4.92 & 25.8 & 65 & 12 hours\\
-\hline\rule{0pt}{2.0ex}
-AOP$_1$ & 512 & & 8 & 64 & 64 & 1536 & 4.92& 25.5  & 65 & 16 hours\\
-AOP$_2$ & 512 & & 16 & 64 & 64 & 512 & \textbf{4.86} & \textbf{25.9}  & 65 & 16 hours \\
-\hline
-\end{tabular}
-%}
-\end{center}
-\end{table}

crazy_functions/test_project/latex/attention/来源

@@ -1,8 +0,0 @@
-chatgpt的老祖宗《Attention is all you need》
-
-Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, Illia Polosukhin
-
-真实的摘要如下
-The dominant sequence transduction models are based on complex recurrent or convolutional neural networks in an encoder-decoder configuration. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely. Experiments on two machine translation tasks show these models to be superior in quality while being more parallelizable and requiring significantly less time to train. Our model achieves 28.4 BLEU on the WMT 2014 English-to-German translation task, improving over the existing best results, including ensembles by over 2 BLEU. On the WMT 2014 English-to-French translation task, our model establishes a new single-model state-of-the-art BLEU score of 41.8 after training for 3.5 days on eight GPUs, a small fraction of the training costs of the best models from the literature. We show that the Transformer generalizes well to other tasks by applying it successfully to English constituency parsing both with large and limited training data.
-
-https://arxiv.org/abs/1706.03762

crazy_functions/test_project/python/dqn/__init__.py

@@ -1,2 +0,0 @@
-from stable_baselines3.dqn.dqn import DQN
-from stable_baselines3.dqn.policies import CnnPolicy, MlpPolicy

crazy_functions/test_project/python/dqn/dqn.py

@@ -1,245 +0,0 @@
-from typing import Any, Dict, List, Optional, Tuple, Type, Union
-
-import gym
-import numpy as np
-import torch as th
-from torch.nn import functional as F
-
-from stable_baselines3.common import logger
-from stable_baselines3.common.off_policy_algorithm import OffPolicyAlgorithm
-from stable_baselines3.common.preprocessing import maybe_transpose
-from stable_baselines3.common.type_aliases import GymEnv, MaybeCallback, Schedule
-from stable_baselines3.common.utils import get_linear_fn, is_vectorized_observation, polyak_update
-from stable_baselines3.dqn.policies import DQNPolicy
-
-
-class DQN(OffPolicyAlgorithm):
-    """
-    Deep Q-Network (DQN)
-
-    Paper: https://arxiv.org/abs/1312.5602, https://www.nature.com/articles/nature14236
-    Default hyperparameters are taken from the nature paper,
-    except for the optimizer and learning rate that were taken from Stable Baselines defaults.
-
-    :param policy: The policy model to use (MlpPolicy, CnnPolicy, ...)
-    :param env: The environment to learn from (if registered in Gym, can be str)
-    :param learning_rate: The learning rate, it can be a function
-        of the current progress remaining (from 1 to 0)
-    :param buffer_size: size of the replay buffer
-    :param learning_starts: how many steps of the model to collect transitions for before learning starts
-    :param batch_size: Minibatch size for each gradient update
-    :param tau: the soft update coefficient ("Polyak update", between 0 and 1) default 1 for hard update
-    :param gamma: the discount factor
-    :param train_freq: Update the model every ``train_freq`` steps. Alternatively pass a tuple of frequency and unit
-        like ``(5, "step")`` or ``(2, "episode")``.
-    :param gradient_steps: How many gradient steps to do after each rollout (see ``train_freq``)
-        Set to ``-1`` means to do as many gradient steps as steps done in the environment
-        during the rollout.
-    :param optimize_memory_usage: Enable a memory efficient variant of the replay buffer
-        at a cost of more complexity.
-        See https://github.com/DLR-RM/stable-baselines3/issues/37#issuecomment-637501195
-    :param target_update_interval: update the target network every ``target_update_interval``
-        environment steps.
-    :param exploration_fraction: fraction of entire training period over which the exploration rate is reduced
-    :param exploration_initial_eps: initial value of random action probability
-    :param exploration_final_eps: final value of random action probability
-    :param max_grad_norm: The maximum value for the gradient clipping
-    :param tensorboard_log: the log location for tensorboard (if None, no logging)
-    :param create_eval_env: Whether to create a second environment that will be
-        used for evaluating the agent periodically. (Only available when passing string for the environment)
-    :param policy_kwargs: additional arguments to be passed to the policy on creation
-    :param verbose: the verbosity level: 0 no output, 1 info, 2 debug
-    :param seed: Seed for the pseudo random generators
-    :param device: Device (cpu, cuda, ...) on which the code should be run.
-        Setting it to auto, the code will be run on the GPU if possible.
-    :param _init_setup_model: Whether or not to build the network at the creation of the instance
-    """
-
-    def __init__(
-        self,
-        policy: Union[str, Type[DQNPolicy]],
-        env: Union[GymEnv, str],
-        learning_rate: Union[float, Schedule] = 1e-4,
-        buffer_size: int = 1000000,
-        learning_starts: int = 50000,
-        batch_size: Optional[int] = 32,
-        tau: float = 1.0,
-        gamma: float = 0.99,
-        train_freq: Union[int, Tuple[int, str]] = 4,
-        gradient_steps: int = 1,
-        optimize_memory_usage: bool = False,
-        target_update_interval: int = 10000,
-        exploration_fraction: float = 0.1,
-        exploration_initial_eps: float = 1.0,
-        exploration_final_eps: float = 0.05,
-        max_grad_norm: float = 10,
-        tensorboard_log: Optional[str] = None,
-        create_eval_env: bool = False,
-        policy_kwargs: Optional[Dict[str, Any]] = None,
-        verbose: int = 0,
-        seed: Optional[int] = None,
-        device: Union[th.device, str] = "auto",
-        _init_setup_model: bool = True,
-    ):
-
-        super(DQN, self).__init__(
-            policy,
-            env,
-            DQNPolicy,
-            learning_rate,
-            buffer_size,
-            learning_starts,
-            batch_size,
-            tau,
-            gamma,
-            train_freq,
-            gradient_steps,
-            action_noise=None,  # No action noise
-            policy_kwargs=policy_kwargs,
-            tensorboard_log=tensorboard_log,
-            verbose=verbose,
-            device=device,
-            create_eval_env=create_eval_env,
-            seed=seed,
-            sde_support=False,
-            optimize_memory_usage=optimize_memory_usage,
-            supported_action_spaces=(gym.spaces.Discrete,),
-        )
-
-        self.exploration_initial_eps = exploration_initial_eps
-        self.exploration_final_eps = exploration_final_eps
-        self.exploration_fraction = exploration_fraction
-        self.target_update_interval = target_update_interval
-        self.max_grad_norm = max_grad_norm
-        # "epsilon" for the epsilon-greedy exploration
-        self.exploration_rate = 0.0
-        # Linear schedule will be defined in `_setup_model()`
-        self.exploration_schedule = None
-        self.q_net, self.q_net_target = None, None
-
-        if _init_setup_model:
-            self._setup_model()
-
-    def _setup_model(self) -> None:
-        super(DQN, self)._setup_model()
-        self._create_aliases()
-        self.exploration_schedule = get_linear_fn(
-            self.exploration_initial_eps, self.exploration_final_eps, self.exploration_fraction
-        )
-
-    def _create_aliases(self) -> None:
-        self.q_net = self.policy.q_net
-        self.q_net_target = self.policy.q_net_target
-
-    def _on_step(self) -> None:
-        """
-        Update the exploration rate and target network if needed.
-        This method is called in ``collect_rollouts()`` after each step in the environment.
-        """
-        if self.num_timesteps % self.target_update_interval == 0:
-            polyak_update(self.q_net.parameters(), self.q_net_target.parameters(), self.tau)
-
-        self.exploration_rate = self.exploration_schedule(self._current_progress_remaining)
-        logger.record("rollout/exploration rate", self.exploration_rate)
-
-    def train(self, gradient_steps: int, batch_size: int = 100) -> None:
-        # Update learning rate according to schedule
-        self._update_learning_rate(self.policy.optimizer)
-
-        losses = []
-        for _ in range(gradient_steps):
-            # Sample replay buffer
-            replay_data = self.replay_buffer.sample(batch_size, env=self._vec_normalize_env)
-
-            with th.no_grad():
-                # Compute the next Q-values using the target network
-                next_q_values = self.q_net_target(replay_data.next_observations)
-                # Follow greedy policy: use the one with the highest value
-                next_q_values, _ = next_q_values.max(dim=1)
-                # Avoid potential broadcast issue
-                next_q_values = next_q_values.reshape(-1, 1)
-                # 1-step TD target
-                target_q_values = replay_data.rewards + (1 - replay_data.dones) * self.gamma * next_q_values
-
-            # Get current Q-values estimates
-            current_q_values = self.q_net(replay_data.observations)
-
-            # Retrieve the q-values for the actions from the replay buffer
-            current_q_values = th.gather(current_q_values, dim=1, index=replay_data.actions.long())
-
-            # Compute Huber loss (less sensitive to outliers)
-            loss = F.smooth_l1_loss(current_q_values, target_q_values)
-            losses.append(loss.item())
-
-            # Optimize the policy
-            self.policy.optimizer.zero_grad()
-            loss.backward()
-            # Clip gradient norm
-            th.nn.utils.clip_grad_norm_(self.policy.parameters(), self.max_grad_norm)
-            self.policy.optimizer.step()
-
-        # Increase update counter
-        self._n_updates += gradient_steps
-
-        logger.record("train/n_updates", self._n_updates, exclude="tensorboard")
-        logger.record("train/loss", np.mean(losses))
-
-    def predict(
-        self,
-        observation: np.ndarray,
-        state: Optional[np.ndarray] = None,
-        mask: Optional[np.ndarray] = None,
-        deterministic: bool = False,
-    ) -> Tuple[np.ndarray, Optional[np.ndarray]]:
-        """
-        Overrides the base_class predict function to include epsilon-greedy exploration.
-
-        :param observation: the input observation
-        :param state: The last states (can be None, used in recurrent policies)
-        :param mask: The last masks (can be None, used in recurrent policies)
-        :param deterministic: Whether or not to return deterministic actions.
-        :return: the model's action and the next state
-            (used in recurrent policies)
-        """
-        if not deterministic and np.random.rand() < self.exploration_rate:
-            if is_vectorized_observation(maybe_transpose(observation, self.observation_space), self.observation_space):
-                n_batch = observation.shape[0]
-                action = np.array([self.action_space.sample() for _ in range(n_batch)])
-            else:
-                action = np.array(self.action_space.sample())
-        else:
-            action, state = self.policy.predict(observation, state, mask, deterministic)
-        return action, state
-
-    def learn(
-        self,
-        total_timesteps: int,
-        callback: MaybeCallback = None,
-        log_interval: int = 4,
-        eval_env: Optional[GymEnv] = None,
-        eval_freq: int = -1,
-        n_eval_episodes: int = 5,
-        tb_log_name: str = "DQN",
-        eval_log_path: Optional[str] = None,
-        reset_num_timesteps: bool = True,
-    ) -> OffPolicyAlgorithm:
-
-        return super(DQN, self).learn(
-            total_timesteps=total_timesteps,
-            callback=callback,
-            log_interval=log_interval,
-            eval_env=eval_env,
-            eval_freq=eval_freq,
-            n_eval_episodes=n_eval_episodes,
-            tb_log_name=tb_log_name,
-            eval_log_path=eval_log_path,
-            reset_num_timesteps=reset_num_timesteps,
-        )
-
-    def _excluded_save_params(self) -> List[str]:
-        return super(DQN, self)._excluded_save_params() + ["q_net", "q_net_target"]
-
-    def _get_torch_save_params(self) -> Tuple[List[str], List[str]]:
-        state_dicts = ["policy", "policy.optimizer"]
-
-        return state_dicts, []

crazy_functions/test_project/python/dqn/policies.py

@@ -1,237 +0,0 @@
-from typing import Any, Dict, List, Optional, Type
-
-import gym
-import torch as th
-from torch import nn
-
-from stable_baselines3.common.policies import BasePolicy, register_policy
-from stable_baselines3.common.torch_layers import BaseFeaturesExtractor, FlattenExtractor, NatureCNN, create_mlp
-from stable_baselines3.common.type_aliases import Schedule
-
-
-class QNetwork(BasePolicy):
-    """
-    Action-Value (Q-Value) network for DQN
-
-    :param observation_space: Observation space
-    :param action_space: Action space
-    :param net_arch: The specification of the policy and value networks.
-    :param activation_fn: Activation function
-    :param normalize_images: Whether to normalize images or not,
-         dividing by 255.0 (True by default)
-    """
-
-    def __init__(
-        self,
-        observation_space: gym.spaces.Space,
-        action_space: gym.spaces.Space,
-        features_extractor: nn.Module,
-        features_dim: int,
-        net_arch: Optional[List[int]] = None,
-        activation_fn: Type[nn.Module] = nn.ReLU,
-        normalize_images: bool = True,
-    ):
-        super(QNetwork, self).__init__(
-            observation_space,
-            action_space,
-            features_extractor=features_extractor,
-            normalize_images=normalize_images,
-        )
-
-        if net_arch is None:
-            net_arch = [64, 64]
-
-        self.net_arch = net_arch
-        self.activation_fn = activation_fn
-        self.features_extractor = features_extractor
-        self.features_dim = features_dim
-        self.normalize_images = normalize_images
-        action_dim = self.action_space.n  # number of actions
-        q_net = create_mlp(self.features_dim, action_dim, self.net_arch, self.activation_fn)
-        self.q_net = nn.Sequential(*q_net)
-
-    def forward(self, obs: th.Tensor) -> th.Tensor:
-        """
-        Predict the q-values.
-
-        :param obs: Observation
-        :return: The estimated Q-Value for each action.
-        """
-        return self.q_net(self.extract_features(obs))
-
-    def _predict(self, observation: th.Tensor, deterministic: bool = True) -> th.Tensor:
-        q_values = self.forward(observation)
-        # Greedy action
-        action = q_values.argmax(dim=1).reshape(-1)
-        return action
-
-    def _get_constructor_parameters(self) -> Dict[str, Any]:
-        data = super()._get_constructor_parameters()
-
-        data.update(
-            dict(
-                net_arch=self.net_arch,
-                features_dim=self.features_dim,
-                activation_fn=self.activation_fn,
-                features_extractor=self.features_extractor,
-            )
-        )
-        return data
-
-
-class DQNPolicy(BasePolicy):
-    """
-    Policy class with Q-Value Net and target net for DQN
-
-    :param observation_space: Observation space
-    :param action_space: Action space
-    :param lr_schedule: Learning rate schedule (could be constant)
-    :param net_arch: The specification of the policy and value networks.
-    :param activation_fn: Activation function
-    :param features_extractor_class: Features extractor to use.
-    :param features_extractor_kwargs: Keyword arguments
-        to pass to the features extractor.
-    :param normalize_images: Whether to normalize images or not,
-         dividing by 255.0 (True by default)
-    :param optimizer_class: The optimizer to use,
-        ``th.optim.Adam`` by default
-    :param optimizer_kwargs: Additional keyword arguments,
-        excluding the learning rate, to pass to the optimizer
-    """
-
-    def __init__(
-        self,
-        observation_space: gym.spaces.Space,
-        action_space: gym.spaces.Space,
-        lr_schedule: Schedule,
-        net_arch: Optional[List[int]] = None,
-        activation_fn: Type[nn.Module] = nn.ReLU,
-        features_extractor_class: Type[BaseFeaturesExtractor] = FlattenExtractor,
-        features_extractor_kwargs: Optional[Dict[str, Any]] = None,
-        normalize_images: bool = True,
-        optimizer_class: Type[th.optim.Optimizer] = th.optim.Adam,
-        optimizer_kwargs: Optional[Dict[str, Any]] = None,
-    ):
-        super(DQNPolicy, self).__init__(
-            observation_space,
-            action_space,
-            features_extractor_class,
-            features_extractor_kwargs,
-            optimizer_class=optimizer_class,
-            optimizer_kwargs=optimizer_kwargs,
-        )
-
-        if net_arch is None:
-            if features_extractor_class == FlattenExtractor:
-                net_arch = [64, 64]
-            else:
-                net_arch = []
-
-        self.net_arch = net_arch
-        self.activation_fn = activation_fn
-        self.normalize_images = normalize_images
-
-        self.net_args = {
-            "observation_space": self.observation_space,
-            "action_space": self.action_space,
-            "net_arch": self.net_arch,
-            "activation_fn": self.activation_fn,
-            "normalize_images": normalize_images,
-        }
-
-        self.q_net, self.q_net_target = None, None
-        self._build(lr_schedule)
-
-    def _build(self, lr_schedule: Schedule) -> None:
-        """
-        Create the network and the optimizer.
-
-        :param lr_schedule: Learning rate schedule
-            lr_schedule(1) is the initial learning rate
-        """
-
-        self.q_net = self.make_q_net()
-        self.q_net_target = self.make_q_net()
-        self.q_net_target.load_state_dict(self.q_net.state_dict())
-
-        # Setup optimizer with initial learning rate
-        self.optimizer = self.optimizer_class(self.parameters(), lr=lr_schedule(1), **self.optimizer_kwargs)
-
-    def make_q_net(self) -> QNetwork:
-        # Make sure we always have separate networks for features extractors etc
-        net_args = self._update_features_extractor(self.net_args, features_extractor=None)
-        return QNetwork(**net_args).to(self.device)
-
-    def forward(self, obs: th.Tensor, deterministic: bool = True) -> th.Tensor:
-        return self._predict(obs, deterministic=deterministic)
-
-    def _predict(self, obs: th.Tensor, deterministic: bool = True) -> th.Tensor:
-        return self.q_net._predict(obs, deterministic=deterministic)
-
-    def _get_constructor_parameters(self) -> Dict[str, Any]:
-        data = super()._get_constructor_parameters()
-
-        data.update(
-            dict(
-                net_arch=self.net_args["net_arch"],
-                activation_fn=self.net_args["activation_fn"],
-                lr_schedule=self._dummy_schedule,  # dummy lr schedule, not needed for loading policy alone
-                optimizer_class=self.optimizer_class,
-                optimizer_kwargs=self.optimizer_kwargs,
-                features_extractor_class=self.features_extractor_class,
-                features_extractor_kwargs=self.features_extractor_kwargs,
-            )
-        )
-        return data
-
-
-MlpPolicy = DQNPolicy
-
-
-class CnnPolicy(DQNPolicy):
-    """
-    Policy class for DQN when using images as input.
-
-    :param observation_space: Observation space
-    :param action_space: Action space
-    :param lr_schedule: Learning rate schedule (could be constant)
-    :param net_arch: The specification of the policy and value networks.
-    :param activation_fn: Activation function
-    :param features_extractor_class: Features extractor to use.
-    :param normalize_images: Whether to normalize images or not,
-         dividing by 255.0 (True by default)
-    :param optimizer_class: The optimizer to use,
-        ``th.optim.Adam`` by default
-    :param optimizer_kwargs: Additional keyword arguments,
-        excluding the learning rate, to pass to the optimizer
-    """
-
-    def __init__(
-        self,
-        observation_space: gym.spaces.Space,
-        action_space: gym.spaces.Space,
-        lr_schedule: Schedule,
-        net_arch: Optional[List[int]] = None,
-        activation_fn: Type[nn.Module] = nn.ReLU,
-        features_extractor_class: Type[BaseFeaturesExtractor] = NatureCNN,
-        features_extractor_kwargs: Optional[Dict[str, Any]] = None,
-        normalize_images: bool = True,
-        optimizer_class: Type[th.optim.Optimizer] = th.optim.Adam,
-        optimizer_kwargs: Optional[Dict[str, Any]] = None,
-    ):
-        super(CnnPolicy, self).__init__(
-            observation_space,
-            action_space,
-            lr_schedule,
-            net_arch,
-            activation_fn,
-            features_extractor_class,
-            features_extractor_kwargs,
-            normalize_images,
-            optimizer_class,
-            optimizer_kwargs,
-        )
-
-
-register_policy("MlpPolicy", MlpPolicy)
-register_policy("CnnPolicy", CnnPolicy)

crazy_functions/test_project/python/dqn/来源

@@ -1,2 +0,0 @@
-github stablebaseline3
-https://github.com/DLR-RM/stable-baselines3

crazy_functions/test_project/其他测试

@@ -1,27 +0,0 @@
-"In practice, we found that a high-entropy initial state is more likely to increase the speed of training.
-The entropy is calculated by:
-$$H=-\sum_{k= 1}^{n_k} p(k) \cdot \log p(k), p(k)=\frac{|A_k|}{|\mathcal{A}|}$$
-where $H$ is the entropy, $|A_k|$ is the number of agent nodes in $k$-th cluster, $|\mathcal{A}|$ is the total number of agents.
-To ensure the Cooperation Graph initialization has higher entropy,
-we will randomly generate multiple initial states, 
-rank by their entropy and then pick the one with maximum $H$."
-
-```
-FROM ubuntu:latest
-
-RUN apt-get update && \
-    apt-get install -y python3 python3-pip && \
-    rm -rf /var/lib/apt/lists/*
-
-RUN echo '[global]' > /etc/pip.conf && \
-    echo 'index-url = https://mirrors.aliyun.com/pypi/simple/' >> /etc/pip.conf && \
-    echo 'trusted-host = mirrors.aliyun.com' >> /etc/pip.conf
-
-RUN pip3 install gradio requests[socks] mdtex2html
-
-COPY . /gpt
-WORKDIR /gpt
-
-
-CMD ["python3", "main.py"]
-```

crazy_functions/vector_fns/vector_database.py

@@ -28,7 +28,7 @@ EMBEDDING_DEVICE = "cpu"
 
 # 基于上下文的prompt模版，请务必保留"{question}"和"{context}"
 PROMPT_TEMPLATE = """已知信息：
-{context} 
+{context}
 
 根据上述已知信息，简洁和专业的来回答用户的问题。如果无法从中得到答案，请说 “根据已知信息无法回答该问题” 或 “没有提供足够的相关信息”，不允许在答案中添加编造成分，答案请使用中文。 问题是：{question}"""
 
@@ -58,7 +58,7 @@ OPEN_CROSS_DOMAIN = False
 def similarity_search_with_score_by_vector(
         self, embedding: List[float], k: int = 4
 ) -> List[Tuple[Document, float]]:
-    
+
     def seperate_list(ls: List[int]) -> List[List[int]]:
         lists = []
         ls1 = [ls[0]]
@@ -200,7 +200,7 @@ class LocalDocQA:
             return vs_path, loaded_files
         else:
             raise RuntimeError("文件加载失败，请检查文件格式是否正确")
-        
+
     def get_loaded_file(self, vs_path):
         ds = self.vector_store.docstore
         return set([ds._dict[k].metadata['source'].split(vs_path)[-1] for k in ds._dict])
@@ -290,10 +290,10 @@ class knowledge_archive_interface():
         self.threadLock.acquire()
         # import uuid
         self.current_id = id
-        self.qa_handle, self.kai_path = construct_vector_store(   
-            vs_id=self.current_id, 
+        self.qa_handle, self.kai_path = construct_vector_store(
+            vs_id=self.current_id,
             vs_path=vs_path,
-            files=file_manifest, 
+            files=file_manifest,
             sentence_size=100,
             history=[],
             one_conent="",
@@ -304,7 +304,7 @@ class knowledge_archive_interface():
 
     def get_current_archive_id(self):
         return self.current_id
-    
+
     def get_loaded_file(self, vs_path):
         return self.qa_handle.get_loaded_file(vs_path)
 
@@ -312,10 +312,10 @@ class knowledge_archive_interface():
         self.threadLock.acquire()
         if not self.current_id == id:
             self.current_id = id
-            self.qa_handle, self.kai_path = construct_vector_store(   
-                vs_id=self.current_id, 
+            self.qa_handle, self.kai_path = construct_vector_store(
+                vs_id=self.current_id,
                 vs_path=vs_path,
-                files=[], 
+                files=[],
                 sentence_size=100,
                 history=[],
                 one_conent="",
@@ -329,7 +329,7 @@ class knowledge_archive_interface():
             query = txt,
             vs_path = self.kai_path,
             score_threshold=VECTOR_SEARCH_SCORE_THRESHOLD,
-            vector_search_top_k=VECTOR_SEARCH_TOP_K, 
+            vector_search_top_k=VECTOR_SEARCH_TOP_K,
             chunk_conent=True,
             chunk_size=CHUNK_SIZE,
             text2vec = self.get_chinese_text2vec(),

crazy_functions/vt_fns/vt_call_plugin.py

@@ -35,9 +35,9 @@ def get_recent_file_prompt_support(chatbot):
     most_recent_uploaded = chatbot._cookies.get("most_recent_uploaded", None)
     path = most_recent_uploaded['path']
     prompt =   "\nAdditional Information:\n"
-    prompt =   "In case that this plugin requires a path or a file as argument," 
-    prompt += f"it is important for you to know that the user has recently uploaded a file, located at: `{path}`" 
-    prompt += f"Only use it when necessary, otherwise, you can ignore this file." 
+    prompt =   "In case that this plugin requires a path or a file as argument,"
+    prompt += f"it is important for you to know that the user has recently uploaded a file, located at: `{path}`"
+    prompt += f"Only use it when necessary, otherwise, you can ignore this file."
     return prompt
 
 def get_inputs_show_user(inputs, plugin_arr_enum_prompt):
@@ -82,7 +82,7 @@ def execute_plugin(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prom
         msg += "\n但您可以尝试再试一次\n"
         yield from update_ui_lastest_msg(lastmsg=msg, chatbot=chatbot, history=history, delay=2)
         return
-    
+
     # ⭐ ⭐ ⭐ 确认插件参数
     if not have_any_recent_upload_files(chatbot):
         appendix_info = ""
@@ -99,7 +99,7 @@ def execute_plugin(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prom
     inputs = f"A plugin named {plugin_sel.plugin_selection} is selected, " + \
              "you should extract plugin_arg from the user requirement, the user requirement is: \n\n" + \
              ">> " + (txt + appendix_info).rstrip('\n').replace('\n','\n>> ') + '\n\n' + \
-             gpt_json_io.format_instructions 
+             gpt_json_io.format_instructions
     run_gpt_fn = lambda inputs, sys_prompt: predict_no_ui_long_connection(
         inputs=inputs, llm_kwargs=llm_kwargs, history=[], sys_prompt=sys_prompt, observe_window=[])
     plugin_sel = gpt_json_io.generate_output_auto_repair(run_gpt_fn(inputs, ""), run_gpt_fn)

crazy_functions/vt_fns/vt_modify_config.py

@@ -10,7 +10,7 @@ def modify_configuration_hot(txt, llm_kwargs, plugin_kwargs, chatbot, history, s
     ALLOW_RESET_CONFIG = get_conf('ALLOW_RESET_CONFIG')
     if not ALLOW_RESET_CONFIG:
         yield from update_ui_lastest_msg(
-            lastmsg=f"当前配置不允许被修改！如需激活本功能，请在config.py中设置ALLOW_RESET_CONFIG=True后重启软件。", 
+            lastmsg=f"当前配置不允许被修改！如需激活本功能，请在config.py中设置ALLOW_RESET_CONFIG=True后重启软件。",
             chatbot=chatbot, history=history, delay=2
         )
         return
@@ -35,7 +35,7 @@ def modify_configuration_hot(txt, llm_kwargs, plugin_kwargs, chatbot, history, s
     inputs = "Analyze how to change configuration according to following user input, answer me with json: \n\n" + \
              ">> " + txt.rstrip('\n').replace('\n','\n>> ') + '\n\n' + \
              gpt_json_io.format_instructions
-    
+
     run_gpt_fn = lambda inputs, sys_prompt: predict_no_ui_long_connection(
         inputs=inputs, llm_kwargs=llm_kwargs, history=[], sys_prompt=sys_prompt, observe_window=[])
     user_intention = gpt_json_io.generate_output_auto_repair(run_gpt_fn(inputs, ""), run_gpt_fn)
@@ -45,11 +45,11 @@ def modify_configuration_hot(txt, llm_kwargs, plugin_kwargs, chatbot, history, s
     ok = (explicit_conf in txt)
     if ok:
         yield from update_ui_lastest_msg(
-            lastmsg=f"正在执行任务: {txt}\n\n新配置{explicit_conf}={user_intention.new_option_value}", 
+            lastmsg=f"正在执行任务: {txt}\n\n新配置{explicit_conf}={user_intention.new_option_value}",
             chatbot=chatbot, history=history, delay=1
         )
         yield from update_ui_lastest_msg(
-            lastmsg=f"正在执行任务: {txt}\n\n新配置{explicit_conf}={user_intention.new_option_value}\n\n正在修改配置中", 
+            lastmsg=f"正在执行任务: {txt}\n\n新配置{explicit_conf}={user_intention.new_option_value}\n\n正在修改配置中",
             chatbot=chatbot, history=history, delay=2
         )
 
@@ -69,7 +69,7 @@ def modify_configuration_reboot(txt, llm_kwargs, plugin_kwargs, chatbot, history
     ALLOW_RESET_CONFIG = get_conf('ALLOW_RESET_CONFIG')
     if not ALLOW_RESET_CONFIG:
         yield from update_ui_lastest_msg(
-            lastmsg=f"当前配置不允许被修改！如需激活本功能，请在config.py中设置ALLOW_RESET_CONFIG=True后重启软件。", 
+            lastmsg=f"当前配置不允许被修改！如需激活本功能，请在config.py中设置ALLOW_RESET_CONFIG=True后重启软件。",
             chatbot=chatbot, history=history, delay=2
         )
         return

crazy_functions/vt_fns/vt_state.py

@@ -6,7 +6,7 @@ class VoidTerminalState():
 
     def reset_state(self):
         self.has_provided_explaination = False
- 
+
     def lock_plugin(self, chatbot):
         chatbot._cookies['lock_plugin'] = 'crazy_functions.虚空终端->虚空终端'
         chatbot._cookies['plugin_state'] = pickle.dumps(self)

crazy_functions/下载arxiv论文翻译摘要.py

@@ -144,8 +144,8 @@ def 下载arxiv论文并翻译摘要(txt, llm_kwargs, plugin_kwargs, chatbot, hi
     try:
         import bs4
     except:
-        report_exception(chatbot, history, 
-            a = f"解析项目: {txt}", 
+        report_exception(chatbot, history,
+            a = f"解析项目: {txt}",
             b = f"导入软件依赖失败。使用该模块需要额外依赖，安装方法```pip install --upgrade beautifulsoup4```。")
         yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
         return
@@ -157,12 +157,12 @@ def 下载arxiv论文并翻译摘要(txt, llm_kwargs, plugin_kwargs, chatbot, hi
     try:
         pdf_path, info = download_arxiv_(txt)
     except:
-        report_exception(chatbot, history, 
-            a = f"解析项目: {txt}", 
+        report_exception(chatbot, history,
+            a = f"解析项目: {txt}",
             b = f"下载pdf文件未成功")
         yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
         return
-    
+
     # 翻译摘要等
     i_say =            f"请你阅读以下学术论文相关的材料，提取摘要，翻译为中文。材料如下：{str(info)}"
     i_say_show_user =  f'请你阅读以下学术论文相关的材料，提取摘要，翻译为中文。论文：{pdf_path}'

crazy_functions/互动小游戏.py

@@ -12,9 +12,9 @@ def 随机小游戏(prompt, llm_kwargs, plugin_kwargs, chatbot, history, system_
     # 选择游戏
     cls = MiniGame_ResumeStory
     # 如果之前已经初始化了游戏实例，则继续该实例；否则重新初始化
-    state = cls.sync_state(chatbot, 
-                           llm_kwargs, 
-                           cls, 
+    state = cls.sync_state(chatbot,
+                           llm_kwargs,
+                           cls,
                            plugin_name='MiniGame_ResumeStory',
                            callback_fn='crazy_functions.互动小游戏->随机小游戏',
                            lock_plugin=True
@@ -30,9 +30,9 @@ def 随机小游戏1(prompt, llm_kwargs, plugin_kwargs, chatbot, history, system
     # 选择游戏
     cls = MiniGame_ASCII_Art
     # 如果之前已经初始化了游戏实例，则继续该实例；否则重新初始化
-    state = cls.sync_state(chatbot, 
-                           llm_kwargs, 
-                           cls, 
+    state = cls.sync_state(chatbot,
+                           llm_kwargs,
+                           cls,
                            plugin_name='MiniGame_ASCII_Art',
                            callback_fn='crazy_functions.互动小游戏->随机小游戏1',
                            lock_plugin=True

crazy_functions/交互功能函数模板.py

@@ -38,7 +38,7 @@ def 交互功能模板函数(txt, llm_kwargs, plugin_kwargs, chatbot, history, s
         inputs=inputs_show_user=f"Extract all image urls in this html page, pick the first 5 images and show them with markdown format: \n\n {page_return}"
         gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
             inputs=inputs, inputs_show_user=inputs_show_user,
-            llm_kwargs=llm_kwargs, chatbot=chatbot, history=[], 
+            llm_kwargs=llm_kwargs, chatbot=chatbot, history=[],
             sys_prompt="When you want to show an image, use markdown format. e.g. ![image_description](image_url). If there are no image url provided, answer 'no image url provided'"
         )
         chatbot[-1] = [chatbot[-1][0], gpt_say]

crazy_functions/代码重写为全英文_多线程.py

@@ -1,138 +0,0 @@
-import threading
-from request_llm.bridge_all import predict_no_ui_long_connection
-from toolbox import update_ui
-from toolbox import CatchException, write_results_to_file, report_execption
-from .crazy_utils import breakdown_txt_to_satisfy_token_limit
-
-def extract_code_block_carefully(txt):
-    splitted = txt.split('```')
-    n_code_block_seg = len(splitted) - 1
-    if n_code_block_seg <= 1: return txt
-    # 剩下的情况都开头除去 ``` 结尾除去一次 ```
-    txt_out = '```'.join(splitted[1:-1])
-    return txt_out
-
-
-
-def break_txt_into_half_at_some_linebreak(txt):
-    lines = txt.split('\n')
-    n_lines = len(lines)
-    pre = lines[:(n_lines//2)]
-    post = lines[(n_lines//2):]
-    return "\n".join(pre), "\n".join(post)
-
-
-@CatchException
-def 全项目切换英文(txt, llm_kwargs, plugin_kwargs, chatbot, history, sys_prompt, web_port):
-    # 第1步：清空历史，以免输入溢出
-    history = []
-
-    # 第2步：尝试导入依赖，如果缺少依赖，则给出安装建议
-    try:
-        import tiktoken
-    except:
-        report_execption(chatbot, history, 
-            a = f"解析项目: {txt}", 
-            b = f"导入软件依赖失败。使用该模块需要额外依赖，安装方法```pip install --upgrade tiktoken```。")
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        return
-
-    # 第3步：集合文件
-    import time, glob, os, shutil, re
-    os.makedirs('gpt_log/generated_english_version', exist_ok=True)
-    os.makedirs('gpt_log/generated_english_version/crazy_functions', exist_ok=True)
-    file_manifest = [f for f in glob.glob('./*.py') if ('test_project' not in f) and ('gpt_log' not in f)] + \
-                    [f for f in glob.glob('./crazy_functions/*.py') if ('test_project' not in f) and ('gpt_log' not in f)]
-    # file_manifest = ['./toolbox.py']
-    i_say_show_user_buffer = []
-
-    # 第4步：随便显示点什么防止卡顿的感觉
-    for index, fp in enumerate(file_manifest):
-        # if 'test_project' in fp: continue
-        with open(fp, 'r', encoding='utf-8', errors='replace') as f:
-            file_content = f.read()
-        i_say_show_user =f'[{index}/{len(file_manifest)}] 接下来请将以下代码中包含的所有中文转化为英文，只输出转化后的英文代码，请用代码块输出代码: {os.path.abspath(fp)}'
-        i_say_show_user_buffer.append(i_say_show_user)
-        chatbot.append((i_say_show_user, "[Local Message] 等待多线程操作，中间过程不予显示."))
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-
-
-    # 第5步：Token限制下的截断与处理
-    MAX_TOKEN = 3000
-    from request_llm.bridge_all import model_info
-    enc = model_info["gpt-3.5-turbo"]['tokenizer']
-    def get_token_fn(txt): return len(enc.encode(txt, disallowed_special=()))
-
-
-    # 第6步：任务函数
-    mutable_return = [None for _ in file_manifest]
-    observe_window = [[""] for _ in file_manifest]
-    def thread_worker(fp,index):
-        if index > 10: 
-            time.sleep(60)
-            print('Openai 限制免费用户每分钟20次请求，降低请求频率中。')
-        with open(fp, 'r', encoding='utf-8', errors='replace') as f:
-            file_content = f.read()
-        i_say_template = lambda fp, file_content: f'接下来请将以下代码中包含的所有中文转化为英文，只输出代码，文件名是{fp}，文件代码是 ```{file_content}```'
-        try:
-            gpt_say = ""
-            # 分解代码文件
-            file_content_breakdown = breakdown_txt_to_satisfy_token_limit(file_content, get_token_fn, MAX_TOKEN)
-            for file_content_partial in file_content_breakdown:
-                i_say = i_say_template(fp, file_content_partial)
-                # # ** gpt request **
-                gpt_say_partial = predict_no_ui_long_connection(inputs=i_say, llm_kwargs=llm_kwargs, history=[], sys_prompt=sys_prompt, observe_window=observe_window[index])
-                gpt_say_partial = extract_code_block_carefully(gpt_say_partial)
-                gpt_say += gpt_say_partial
-            mutable_return[index] = gpt_say
-        except ConnectionAbortedError as token_exceed_err:
-            print('至少一个线程任务Token溢出而失败', e)
-        except Exception as e:
-            print('至少一个线程任务意外失败', e)
-
-    # 第7步：所有线程同时开始执行任务函数
-    handles = [threading.Thread(target=thread_worker, args=(fp,index)) for index, fp in enumerate(file_manifest)]
-    for h in handles:
-        h.daemon = True
-        h.start()
-    chatbot.append(('开始了吗？', f'多线程操作已经开始'))
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-
-    # 第8步：循环轮询各个线程是否执行完毕
-    cnt = 0
-    while True:
-        cnt += 1
-        time.sleep(0.2)
-        th_alive = [h.is_alive() for h in handles]
-        if not any(th_alive): break
-        # 更好的UI视觉效果
-        observe_win = []
-        for thread_index, alive in enumerate(th_alive): 
-            observe_win.append("[ ..."+observe_window[thread_index][0][-60:].replace('\n','').replace('```','...').replace(' ','.').replace('<br/>','.....').replace('$','.')+"... ]")
-        stat = [f'执行中: {obs}\n\n' if alive else '已完成\n\n' for alive, obs in zip(th_alive, observe_win)]
-        stat_str = ''.join(stat)
-        chatbot[-1] = (chatbot[-1][0], f'多线程操作已经开始，完成情况: \n\n{stat_str}' + ''.join(['.']*(cnt%10+1)))
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-
-    # 第9步：把结果写入文件
-    for index, h in enumerate(handles):
-        h.join() # 这里其实不需要join了，肯定已经都结束了
-        fp = file_manifest[index]
-        gpt_say = mutable_return[index]
-        i_say_show_user = i_say_show_user_buffer[index]
-
-        where_to_relocate = f'gpt_log/generated_english_version/{fp}'
-        if gpt_say is not None:
-            with open(where_to_relocate, 'w+', encoding='utf-8') as f:  
-                f.write(gpt_say)
-        else:  # 失败
-            shutil.copyfile(file_manifest[index], where_to_relocate)
-        chatbot.append((i_say_show_user, f'[Local Message] 已完成{os.path.abspath(fp)}的转化，\n\n存入{os.path.abspath(where_to_relocate)}'))
-        history.append(i_say_show_user); history.append(gpt_say)
-        yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-        time.sleep(1)
-
-    # 第10步：备份一个文件
-    res = write_results_to_file(history)
-    chatbot.append(("生成一份任务执行报告", res))
-    yield from update_ui(chatbot=chatbot, history=history) # 刷新界面

crazy_functions/函数动态生成.py

@@ -6,10 +6,10 @@
     - 将图像转为灰度图像
     - 将csv文件转excel表格
 
-Testing: 
-    - Crop the image, keeping the bottom half. 
-    - Swap the blue channel and red channel of the image. 
-    - Convert the image to grayscale. 
+Testing:
+    - Crop the image, keeping the bottom half.
+    - Swap the blue channel and red channel of the image.
+    - Convert the image to grayscale.
     - Convert the CSV file to an Excel spreadsheet.
 """
 
@@ -29,12 +29,12 @@ import multiprocessing
 
 templete = """
 ```python
-import ...  # Put dependencies here, e.g. import numpy as np. 
+import ...  # Put dependencies here, e.g. import numpy as np.
 
 class TerminalFunction(object): # Do not change the name of the class, The name of the class must be `TerminalFunction`
 
     def run(self, path):    # The name of the function must be `run`, it takes only a positional argument.
-        # rewrite the function you have just written here 
+        # rewrite the function you have just written here
         ...
         return generated_file_path
 ```
@@ -48,7 +48,7 @@ def get_code_block(reply):
     import re
     pattern = r"```([\s\S]*?)```" # regex pattern to match code blocks
     matches = re.findall(pattern, reply) # find all code blocks in text
-    if len(matches) == 1: 
+    if len(matches) == 1:
         return matches[0].strip('python') #  code block
     for match in matches:
         if 'class TerminalFunction' in match:
@@ -68,8 +68,8 @@ def gpt_interact_multi_step(txt, file_type, llm_kwargs, chatbot, history):
 
     # 第一步
     gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-        inputs=i_say, inputs_show_user=i_say, 
-        llm_kwargs=llm_kwargs, chatbot=chatbot, history=demo, 
+        inputs=i_say, inputs_show_user=i_say,
+        llm_kwargs=llm_kwargs, chatbot=chatbot, history=demo,
         sys_prompt= r"You are a world-class programmer."
     )
     history.extend([i_say, gpt_say])
@@ -82,33 +82,33 @@ def gpt_interact_multi_step(txt, file_type, llm_kwargs, chatbot, history):
     ]
     i_say = "".join(prompt_compose); inputs_show_user = "If previous stage is successful, rewrite the function you have just written to satisfy executable templete. "
     gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-        inputs=i_say, inputs_show_user=inputs_show_user, 
-        llm_kwargs=llm_kwargs, chatbot=chatbot, history=history, 
+        inputs=i_say, inputs_show_user=inputs_show_user,
+        llm_kwargs=llm_kwargs, chatbot=chatbot, history=history,
         sys_prompt= r"You are a programmer. You need to replace `...` with valid packages, do not give `...` in your answer!"
     )
     code_to_return = gpt_say
     history.extend([i_say, gpt_say])
     yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 界面更新
-    
+
     # # 第三步
     # i_say = "Please list to packages to install to run the code above. Then show me how to use `try_install_deps` function to install them."
     # i_say += 'For instance. `try_install_deps(["opencv-python", "scipy", "numpy"])`'
     # installation_advance = yield from request_gpt_model_in_new_thread_with_ui_alive(
-    #     inputs=i_say, inputs_show_user=inputs_show_user, 
-    #     llm_kwargs=llm_kwargs, chatbot=chatbot, history=history, 
+    #     inputs=i_say, inputs_show_user=inputs_show_user,
+    #     llm_kwargs=llm_kwargs, chatbot=chatbot, history=history,
     #     sys_prompt= r"You are a programmer."
     # )
 
-    # # # 第三步  
+    # # # 第三步
     # i_say = "Show me how to use `pip` to install packages to run the code above. "
     # i_say += 'For instance. `pip install -r opencv-python scipy numpy`'
     # installation_advance = yield from request_gpt_model_in_new_thread_with_ui_alive(
-    #     inputs=i_say, inputs_show_user=i_say, 
-    #     llm_kwargs=llm_kwargs, chatbot=chatbot, history=history, 
+    #     inputs=i_say, inputs_show_user=i_say,
+    #     llm_kwargs=llm_kwargs, chatbot=chatbot, history=history,
     #     sys_prompt= r"You are a programmer."
     # )
     installation_advance = ""
-    
+
     return code_to_return, installation_advance, txt, file_type, llm_kwargs, chatbot, history
 
 
@@ -117,7 +117,7 @@ def gpt_interact_multi_step(txt, file_type, llm_kwargs, chatbot, history):
 def for_immediate_show_off_when_possible(file_type, fp, chatbot):
     if file_type in ['png', 'jpg']:
         image_path = os.path.abspath(fp)
-        chatbot.append(['这是一张图片, 展示如下:',  
+        chatbot.append(['这是一张图片, 展示如下:',
             f'本地文件地址: <br/>`{image_path}`<br/>'+
             f'本地文件预览: <br/><div align="center"><img src="file={image_path}"></div>'
         ])
@@ -177,7 +177,7 @@ def 函数动态生成(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_
         chatbot.append(["文件检索", "没有发现任何近期上传的文件。"])
         yield from update_ui_lastest_msg("没有发现任何近期上传的文件。", chatbot, history, 1)
         return  # 2. 如果没有文件
-    
+
     # 读取文件
     file_type = file_list[0].split('.')[-1]
 
@@ -185,7 +185,7 @@ def 函数动态生成(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_
     if is_the_upload_folder(txt):
         yield from update_ui_lastest_msg(f"请在输入框内填写需求, 然后再次点击该插件! 至于您的文件，不用担心, 文件路径 {txt} 已经被记忆. ", chatbot, history, 1)
         return
-    
+
     # 开始干正事
     MAX_TRY = 3
     for j in range(MAX_TRY):  # 最多重试5次
@@ -238,7 +238,7 @@ def 函数动态生成(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_
             # chatbot.append(["如果是缺乏依赖，请参考以下建议", installation_advance])
             yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
             return
-        
+
         # 顺利完成，收尾
         res = str(res)
         if os.path.exists(res):
@@ -248,5 +248,5 @@ def 函数动态生成(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_
             yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 界面更新
         else:
             chatbot.append(["执行成功了，结果是一个字符串", "结果：" + res])
-            yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 界面更新   
+            yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 界面更新
 

crazy_functions/命令行助手.py

@@ -21,8 +21,8 @@ def 命令行助手(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_pro
     i_say = "请写bash命令实现以下功能：" + txt
     # 开始
     gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-        inputs=i_say, inputs_show_user=txt, 
-        llm_kwargs=llm_kwargs, chatbot=chatbot, history=[], 
+        inputs=i_say, inputs_show_user=txt,
+        llm_kwargs=llm_kwargs, chatbot=chatbot, history=[],
         sys_prompt="你是一个Linux大师级用户。注意，当我要求你写bash命令时，尽可能地仅用一行命令解决我的要求。"
     )
     yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 界面更新

crazy_functions/图片生成.py

@@ -7,7 +7,7 @@ def gen_image(llm_kwargs, prompt, resolution="1024x1024", model="dall-e-2", qual
     from request_llms.bridge_all import model_info
 
     proxies = get_conf('proxies')
-    # Set up OpenAI API key and model 
+    # Set up OpenAI API key and model
     api_key = select_api_key(llm_kwargs['api_key'], llm_kwargs['llm_model'])
     chat_endpoint = model_info[llm_kwargs['llm_model']]['endpoint']
     # 'https://api.openai.com/v1/chat/completions'
@@ -113,7 +113,7 @@ def 图片生成_DALLE2(prompt, llm_kwargs, plugin_kwargs, chatbot, history, sys
     if ("advanced_arg" in plugin_kwargs) and (plugin_kwargs["advanced_arg"] == ""): plugin_kwargs.pop("advanced_arg")
     resolution = plugin_kwargs.get("advanced_arg", '1024x1024')
     image_url, image_path = gen_image(llm_kwargs, prompt, resolution)
-    chatbot.append([prompt,  
+    chatbot.append([prompt,
         f'图像中转网址: <br/>`{image_url}`<br/>'+
         f'中转网址预览: <br/><div align="center"><img src="{image_url}"></div>'
         f'本地文件地址: <br/>`{image_path}`<br/>'+
@@ -144,7 +144,7 @@ def 图片生成_DALLE3(prompt, llm_kwargs, plugin_kwargs, chatbot, history, sys
         elif part in ['vivid', 'natural']:
             style = part
     image_url, image_path = gen_image(llm_kwargs, prompt, resolution, model="dall-e-3", quality=quality, style=style)
-    chatbot.append([prompt,  
+    chatbot.append([prompt,
         f'图像中转网址: <br/>`{image_url}`<br/>'+
         f'中转网址预览: <br/><div align="center"><img src="{image_url}"></div>'
         f'本地文件地址: <br/>`{image_path}`<br/>'+
@@ -164,7 +164,7 @@ class ImageEditState(GptAcademicState):
         confirm = (len(file_manifest) >= 1 and file_manifest[0].endswith('.png') and os.path.exists(file_manifest[0]))
         file = None if not confirm else file_manifest[0]
         return confirm, file
-    
+
     def lock_plugin(self, chatbot):
         chatbot._cookies['lock_plugin'] = 'crazy_functions.图片生成->图片修改_DALLE2'
         self.dump_state(chatbot)

crazy_functions/多智能体.py

@@ -57,11 +57,11 @@ def 多智能体终端(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_
         if get_conf("AUTOGEN_USE_DOCKER"):
             import docker
     except:
-        chatbot.append([ f"处理任务: {txt}", 
+        chatbot.append([ f"处理任务: {txt}",
             f"导入软件依赖失败。使用该模块需要额外依赖，安装方法```pip install --upgrade pyautogen docker```。"])
         yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
         return
-    
+
     # 尝试导入依赖，如果缺少依赖，则给出安装建议
     try:
         import autogen
@@ -72,7 +72,7 @@ def 多智能体终端(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_
         chatbot.append([f"处理任务: {txt}", f"缺少docker运行环境！"])
         yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
         return
-    
+
     # 解锁插件
     chatbot.get_cookies()['lock_plugin'] = None
     persistent_class_multi_user_manager = GradioMultiuserManagerForPersistentClasses()

crazy_functions/对话历史存档.py

@@ -66,7 +66,7 @@ def read_file_to_chat(chatbot, history, file_name):
         i_say, gpt_say = h.split('<hr style="border-top: dotted 3px #ccc;">')
         chatbot.append([i_say, gpt_say])
     chatbot.append([f"存档文件详情？", f"[Local Message] 载入对话{len(html)}条，上下文{len(history)}条。"])
-    return chatbot, history    
+    return chatbot, history
 
 @CatchException
 def 对话历史存档(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, user_request):
@@ -80,7 +80,7 @@ def 对话历史存档(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_
     user_request    当前用户的请求信息（IP地址等）
     """
 
-    chatbot.append(("保存当前对话", 
+    chatbot.append(("保存当前对话",
         f"[Local Message] {write_chat_to_file(chatbot, history)}，您可以调用下拉菜单中的“载入对话历史存档”还原当下的对话。"))
     yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 由于请求gpt需要一段时间，我们先及时地做一次界面更新
 
@@ -108,9 +108,9 @@ def 载入对话历史存档(txt, llm_kwargs, plugin_kwargs, chatbot, history, s
         if txt == "": txt = '空空如也的输入栏'
         import glob
         local_history = "<br/>".join([
-            "`"+hide_cwd(f)+f" ({gen_file_preview(f)})"+"`" 
+            "`"+hide_cwd(f)+f" ({gen_file_preview(f)})"+"`"
             for f in glob.glob(
-                f'{get_log_folder(get_user(chatbot), plugin_name="chat_history")}/**/{f_prefix}*.html', 
+                f'{get_log_folder(get_user(chatbot), plugin_name="chat_history")}/**/{f_prefix}*.html',
                 recursive=True
             )])
         chatbot.append([f"正在查找对话历史文件（html格式）: {txt}", f"找不到任何html文件: {txt}。但本地存储了以下历史文件，您可以将任意一个文件路径粘贴到输入区，然后重试：<br/>{local_history}"])
@@ -139,7 +139,7 @@ def 删除所有本地对话历史记录(txt, llm_kwargs, plugin_kwargs, chatbot
 
     import glob, os
     local_history = "<br/>".join([
-        "`"+hide_cwd(f)+"`" 
+        "`"+hide_cwd(f)+"`"
         for f in glob.glob(
             f'{get_log_folder(get_user(chatbot), plugin_name="chat_history")}/**/{f_prefix}*.html', recursive=True
         )])

crazy_functions/总结word文档.py

@@ -40,10 +40,10 @@ def 解析docx(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot
             i_say = f'请对下面的文章片段用中文做概述，文件名是{os.path.relpath(fp, project_folder)}，文章内容是 ```{paper_frag}```'
             i_say_show_user = f'请对下面的文章片段做概述: {os.path.abspath(fp)}的第{i+1}/{len(paper_fragments)}个片段。'
             gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-                inputs=i_say, 
-                inputs_show_user=i_say_show_user, 
+                inputs=i_say,
+                inputs_show_user=i_say_show_user,
                 llm_kwargs=llm_kwargs,
-                chatbot=chatbot, 
+                chatbot=chatbot,
                 history=[],
                 sys_prompt="总结文章。"
             )
@@ -56,10 +56,10 @@ def 解析docx(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot
         if len(paper_fragments) > 1:
             i_say = f"根据以上的对话，总结文章{os.path.abspath(fp)}的主要内容。"
             gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-                inputs=i_say, 
-                inputs_show_user=i_say, 
+                inputs=i_say,
+                inputs_show_user=i_say,
                 llm_kwargs=llm_kwargs,
-                chatbot=chatbot, 
+                chatbot=chatbot,
                 history=this_paper_history,
                 sys_prompt="总结文章。"
             )

crazy_functions/批量Markdown翻译.py

@@ -53,7 +53,7 @@ class PaperFileGroup():
 def 多文件翻译(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, language='en'):
     from .crazy_utils import request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency
 
-    #  <-------- 读取Markdown文件，删除其中的所有注释 ----------> 
+    #  <-------- 读取Markdown文件，删除其中的所有注释 ---------->
     pfg = PaperFileGroup()
 
     for index, fp in enumerate(file_manifest):
@@ -63,23 +63,23 @@ def 多文件翻译(file_manifest, project_folder, llm_kwargs, plugin_kwargs, ch
             pfg.file_paths.append(fp)
             pfg.file_contents.append(file_content)
 
-    #  <-------- 拆分过长的Markdown文件 ----------> 
+    #  <-------- 拆分过长的Markdown文件 ---------->
     pfg.run_file_split(max_token_limit=1500)
     n_split = len(pfg.sp_file_contents)
 
-    #  <-------- 多线程翻译开始 ----------> 
+    #  <-------- 多线程翻译开始 ---------->
     if language == 'en->zh':
-        inputs_array = ["This is a Markdown file, translate it into Chinese, do not modify any existing Markdown commands:" + 
+        inputs_array = ["This is a Markdown file, translate it into Chinese, do not modify any existing Markdown commands:" +
                         f"\n\n{frag}" for frag in pfg.sp_file_contents]
         inputs_show_user_array = [f"翻译 {f}" for f in pfg.sp_file_tag]
         sys_prompt_array = ["You are a professional academic paper translator." for _ in range(n_split)]
     elif language == 'zh->en':
-        inputs_array = [f"This is a Markdown file, translate it into English, do not modify any existing Markdown commands:" + 
+        inputs_array = [f"This is a Markdown file, translate it into English, do not modify any existing Markdown commands:" +
                         f"\n\n{frag}" for frag in pfg.sp_file_contents]
         inputs_show_user_array = [f"翻译 {f}" for f in pfg.sp_file_tag]
         sys_prompt_array = ["You are a professional academic paper translator." for _ in range(n_split)]
     else:
-        inputs_array = [f"This is a Markdown file, translate it into {language}, do not modify any existing Markdown commands, only answer me with translated results:" + 
+        inputs_array = [f"This is a Markdown file, translate it into {language}, do not modify any existing Markdown commands, only answer me with translated results:" +
                         f"\n\n{frag}" for frag in pfg.sp_file_contents]
         inputs_show_user_array = [f"翻译 {f}" for f in pfg.sp_file_tag]
         sys_prompt_array = ["You are a professional academic paper translator." for _ in range(n_split)]
@@ -103,7 +103,7 @@ def 多文件翻译(file_manifest, project_folder, llm_kwargs, plugin_kwargs, ch
     except:
         logging.error(trimmed_format_exc())
 
-    #  <-------- 整理结果，退出 ----------> 
+    #  <-------- 整理结果，退出 ---------->
     create_report_file_name = gen_time_str() + f"-chatgpt.md"
     res = write_history_to_file(gpt_response_collection, file_basename=create_report_file_name)
     promote_file_to_downloadzone(res, chatbot=chatbot)
@@ -255,7 +255,7 @@ def Markdown翻译指定语言(txt, llm_kwargs, plugin_kwargs, chatbot, history,
         report_exception(chatbot, history, a = f"解析项目: {txt}", b = f"找不到任何.md文件: {txt}")
         yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
         return
-    
+
     if ("advanced_arg" in plugin_kwargs) and (plugin_kwargs["advanced_arg"] == ""): plugin_kwargs.pop("advanced_arg")
     language = plugin_kwargs.get("advanced_arg", 'Chinese')
     yield from 多文件翻译(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, language=language)

crazy_functions/批量总结PDF文档.py

@@ -17,7 +17,7 @@ def 解析PDF(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot,
         file_content, page_one = read_and_clean_pdf_text(file_name) # （尝试）按照章节切割PDF
         file_content = file_content.encode('utf-8', 'ignore').decode()   # avoid reading non-utf8 chars
         page_one = str(page_one).encode('utf-8', 'ignore').decode()  # avoid reading non-utf8 chars
-        
+
         TOKEN_LIMIT_PER_FRAGMENT = 2500
 
         from crazy_functions.pdf_fns.breakdown_txt import breakdown_text_to_satisfy_token_limit
@@ -25,7 +25,7 @@ def 解析PDF(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot,
         page_one_fragments = breakdown_text_to_satisfy_token_limit(txt=str(page_one), limit=TOKEN_LIMIT_PER_FRAGMENT//4, llm_model=llm_kwargs['llm_model'])
         # 为了更好的效果，我们剥离Introduction之后的部分（如果有）
         paper_meta = page_one_fragments[0].split('introduction')[0].split('Introduction')[0].split('INTRODUCTION')[0]
-        
+
         ############################## <第 1 步，从摘要中提取高价值信息，放到history中> ##################################
         final_results = []
         final_results.append(paper_meta)
@@ -44,10 +44,10 @@ def 解析PDF(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot,
             i_say = f"Read this section, recapitulate the content of this section with less than {NUM_OF_WORD} Chinese characters: {paper_fragments[i]}"
             i_say_show_user = f"[{i+1}/{n_fragment}] Read this section, recapitulate the content of this section with less than {NUM_OF_WORD} Chinese characters: {paper_fragments[i][:200]}"
             gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(i_say, i_say_show_user,  # i_say=真正给chatgpt的提问， i_say_show_user=给用户看的提问
-                                                                                llm_kwargs, chatbot, 
+                                                                                llm_kwargs, chatbot,
                                                                                 history=["The main idea of the previous section is?", last_iteration_result], # 迭代上一次的结果
                                                                                 sys_prompt="Extract the main idea of this section with Chinese."  # 提示
-                                                                                ) 
+                                                                                )
             iteration_results.append(gpt_say)
             last_iteration_result = gpt_say
 
@@ -67,15 +67,15 @@ def 解析PDF(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot,
     - (2):What are the past methods? What are the problems with them? Is the approach well motivated?
     - (3):What is the research methodology proposed in this paper?
     - (4):On what task and what performance is achieved by the methods in this paper? Can the performance support their goals?
-Follow the format of the output that follows:                  
+Follow the format of the output that follows:
 1. Title: xxx\n\n
 2. Authors: xxx\n\n
 3. Affiliation: xxx\n\n
 4. Keywords: xxx\n\n
 5. Urls: xxx or xxx , xxx \n\n
 6. Summary: \n\n
-    - (1):xxx;\n 
-    - (2):xxx;\n 
+    - (1):xxx;\n
+    - (2):xxx;\n
     - (3):xxx;\n
     - (4):xxx.\n\n
 Be sure to use Chinese answers (proper nouns need to be marked in English), statements as concise and academic as possible,
@@ -85,8 +85,8 @@ do not have too much repetitive information, numerical values using the original
         file_write_buffer.extend(final_results)
         i_say, final_results = input_clipping(i_say, final_results, max_token_limit=2000)
         gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-            inputs=i_say, inputs_show_user='开始最终总结', 
-            llm_kwargs=llm_kwargs, chatbot=chatbot, history=final_results, 
+            inputs=i_say, inputs_show_user='开始最终总结',
+            llm_kwargs=llm_kwargs, chatbot=chatbot, history=final_results,
             sys_prompt= f"Extract the main idea of this paper with less than {NUM_OF_WORD} Chinese characters"
         )
         final_results.append(gpt_say)
@@ -114,8 +114,8 @@ def 批量总结PDF文档(txt, llm_kwargs, plugin_kwargs, chatbot, history, syst
     try:
         import fitz
     except:
-        report_exception(chatbot, history, 
-            a = f"解析项目: {txt}", 
+        report_exception(chatbot, history,
+            a = f"解析项目: {txt}",
             b = f"导入软件依赖失败。使用该模块需要额外依赖，安装方法```pip install --upgrade pymupdf```。")
         yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
         return
@@ -134,7 +134,7 @@ def 批量总结PDF文档(txt, llm_kwargs, plugin_kwargs, chatbot, history, syst
 
     # 搜索需要处理的文件清单
     file_manifest = [f for f in glob.glob(f'{project_folder}/**/*.pdf', recursive=True)]
-    
+
     # 如果没找到任何文件
     if len(file_manifest) == 0:
         report_exception(chatbot, history, a = f"解析项目: {txt}", b = f"找不到任何.tex或.pdf文件: {txt}")

crazy_functions/批量总结PDF文档pdfminer.py

@@ -85,10 +85,10 @@ def 解析Paper(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbo
             msg = '正常'
             # ** gpt request **
             gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-                inputs=i_say, 
-                inputs_show_user=i_say_show_user, 
+                inputs=i_say,
+                inputs_show_user=i_say_show_user,
                 llm_kwargs=llm_kwargs,
-                chatbot=chatbot, 
+                chatbot=chatbot,
                 history=[],
                 sys_prompt="总结文章。"
             )  # 带超时倒计时
@@ -106,10 +106,10 @@ def 解析Paper(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbo
         msg = '正常'
         # ** gpt request **
         gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-            inputs=i_say, 
-            inputs_show_user=i_say, 
+            inputs=i_say,
+            inputs_show_user=i_say,
             llm_kwargs=llm_kwargs,
-            chatbot=chatbot, 
+            chatbot=chatbot,
             history=history,
             sys_prompt="总结文章。"
         )  # 带超时倒计时
@@ -138,8 +138,8 @@ def 批量总结PDF文档pdfminer(txt, llm_kwargs, plugin_kwargs, chatbot, histo
     try:
         import pdfminer, bs4
     except:
-        report_exception(chatbot, history, 
-            a = f"解析项目: {txt}", 
+        report_exception(chatbot, history,
+            a = f"解析项目: {txt}",
             b = f"导入软件依赖失败。使用该模块需要额外依赖，安装方法```pip install --upgrade pdfminer beautifulsoup4```。")
         yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
         return

crazy_functions/批量翻译PDF文档_NOUGAT.py

@@ -76,8 +76,8 @@ def 批量翻译PDF文档(txt, llm_kwargs, plugin_kwargs, chatbot, history, syst
     success_mmd, file_manifest_mmd, _ = get_files_from_everything(txt, type='.mmd')
     success = success or success_mmd
     file_manifest += file_manifest_mmd
-    chatbot.append(["文件列表：", ", ".join([e.split('/')[-1] for e in file_manifest])]); 
-    yield from update_ui(      chatbot=chatbot, history=history) 
+    chatbot.append(["文件列表：", ", ".join([e.split('/')[-1] for e in file_manifest])]);
+    yield from update_ui(      chatbot=chatbot, history=history)
     # 检测输入参数，如没有给定输入参数，直接退出
     if not success:
         if txt == "": txt = '空空如也的输入栏'

crazy_functions/批量翻译PDF文档_多线程.py

@@ -68,7 +68,7 @@ def 解析PDF_基于GROBID(file_manifest, project_folder, llm_kwargs, plugin_kwa
         with open(grobid_json_res, 'w+', encoding='utf8') as f:
             f.write(json.dumps(article_dict, indent=4, ensure_ascii=False))
         promote_file_to_downloadzone(grobid_json_res, chatbot=chatbot)
-        
+
         if article_dict is None: raise RuntimeError("解析PDF失败，请检查PDF是否损坏。")
         yield from translate_pdf(article_dict, llm_kwargs, chatbot, fp, generated_conclusion_files, TOKEN_LIMIT_PER_FRAGMENT, DST_LANG)
     chatbot.append(("给出输出文件清单", str(generated_conclusion_files + generated_html_files)))
@@ -97,7 +97,7 @@ def 解析PDF(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot,
 
         # 为了更好的效果，我们剥离Introduction之后的部分（如果有）
         paper_meta = page_one_fragments[0].split('introduction')[0].split('Introduction')[0].split('INTRODUCTION')[0]
-        
+
         # 单线，获取文章meta信息
         paper_meta_info = yield from request_gpt_model_in_new_thread_with_ui_alive(
             inputs=f"以下是一篇学术论文的基础信息，请从中提取出“标题”、“收录会议或期刊”、“作者”、“摘要”、“编号”、“作者邮箱”这六个部分。请用markdown格式输出，最后用中文翻译摘要部分。请提取：{paper_meta}",
@@ -121,7 +121,7 @@ def 解析PDF(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot,
         )
         gpt_response_collection_md = copy.deepcopy(gpt_response_collection)
         # 整理报告的格式
-        for i,k in enumerate(gpt_response_collection_md): 
+        for i,k in enumerate(gpt_response_collection_md):
             if i%2==0:
                 gpt_response_collection_md[i] = f"\n\n---\n\n ## 原文[{i//2}/{len(gpt_response_collection_md)//2}]： \n\n {paper_fragments[i//2].replace('#', '')}  \n\n---\n\n ## 翻译[{i//2}/{len(gpt_response_collection_md)//2}]：\n "
             else:
@@ -139,18 +139,18 @@ def 解析PDF(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot,
 
         # write html
         try:
-            ch = construct_html() 
+            ch = construct_html()
             orig = ""
             trans = ""
             gpt_response_collection_html = copy.deepcopy(gpt_response_collection)
-            for i,k in enumerate(gpt_response_collection_html): 
+            for i,k in enumerate(gpt_response_collection_html):
                 if i%2==0:
                     gpt_response_collection_html[i] = paper_fragments[i//2].replace('#', '')
                 else:
                     gpt_response_collection_html[i] = gpt_response_collection_html[i]
             final = ["论文概况", paper_meta_info.replace('# ', '### '),  "二、论文翻译",  ""]
             final.extend(gpt_response_collection_html)
-            for i, k in enumerate(final): 
+            for i, k in enumerate(final):
                 if i%2==0:
                     orig = k
                 if i%2==1:

crazy_functions/数学动画生成manim.py

@@ -27,7 +27,7 @@ def eval_manim(code):
 
     class_name = get_class_name(code)
 
-    try: 
+    try:
         time_str = gen_time_str()
         subprocess.check_output([sys.executable, '-c', f"from gpt_log.MyAnimation import {class_name}; {class_name}().render()"])
         shutil.move(f'media/videos/1080p60/{class_name}.mp4', f'gpt_log/{class_name}-{time_str}.mp4')
@@ -36,7 +36,7 @@ def eval_manim(code):
         output = e.output.decode()
         print(f"Command returned non-zero exit status {e.returncode}: {output}.")
         return f"Evaluating python script failed: {e.output}."
-    except: 
+    except:
         print('generating mp4 failed')
         return "Generating mp4 failed."
 
@@ -45,7 +45,7 @@ def get_code_block(reply):
     import re
     pattern = r"```([\s\S]*?)```" # regex pattern to match code blocks
     matches = re.findall(pattern, reply) # find all code blocks in text
-    if len(matches) != 1: 
+    if len(matches) != 1:
         raise RuntimeError("GPT is not generating proper code.")
     return matches[0].strip('python') #  code block
 
@@ -61,7 +61,7 @@ def 动画生成(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt
     user_request    当前用户的请求信息（IP地址等）
     """
     # 清空历史，以免输入溢出
-    history = []    
+    history = []
 
     # 基本信息：功能、贡献者
     chatbot.append([
@@ -73,24 +73,24 @@ def 动画生成(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt
     # 尝试导入依赖, 如果缺少依赖, 则给出安装建议
     dep_ok = yield from inspect_dependency(chatbot=chatbot, history=history) # 刷新界面
     if not dep_ok: return
-    
+
     # 输入
     i_say = f'Generate a animation to show: ' + txt
     demo = ["Here is some examples of manim", examples_of_manim()]
     _, demo = input_clipping(inputs="", history=demo, max_token_limit=2560)
     # 开始
     gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-        inputs=i_say, inputs_show_user=i_say, 
-        llm_kwargs=llm_kwargs, chatbot=chatbot, history=demo, 
+        inputs=i_say, inputs_show_user=i_say,
+        llm_kwargs=llm_kwargs, chatbot=chatbot, history=demo,
         sys_prompt=
         r"Write a animation script with 3blue1brown's manim. "+
-        r"Please begin with `from manim import *`. " + 
+        r"Please begin with `from manim import *`. " +
         r"Answer me with a code block wrapped by ```."
     )
     chatbot.append(["开始生成动画", "..."])
     history.extend([i_say, gpt_say])
     yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 界面更新
-    
+
     # 将代码转为动画
     code = get_code_block(gpt_say)
     res = eval_manim(code)

crazy_functions/理解PDF文档内容.py

@@ -15,7 +15,7 @@ def 解析PDF(file_name, llm_kwargs, plugin_kwargs, chatbot, history, system_pro
     file_content, page_one = read_and_clean_pdf_text(file_name) # （尝试）按照章节切割PDF
     file_content = file_content.encode('utf-8', 'ignore').decode()   # avoid reading non-utf8 chars
     page_one = str(page_one).encode('utf-8', 'ignore').decode()  # avoid reading non-utf8 chars
-    
+
     TOKEN_LIMIT_PER_FRAGMENT = 2500
 
     from crazy_functions.pdf_fns.breakdown_txt import breakdown_text_to_satisfy_token_limit
@@ -23,7 +23,7 @@ def 解析PDF(file_name, llm_kwargs, plugin_kwargs, chatbot, history, system_pro
     page_one_fragments = breakdown_text_to_satisfy_token_limit(txt=str(page_one), limit=TOKEN_LIMIT_PER_FRAGMENT//4, llm_model=llm_kwargs['llm_model'])
     # 为了更好的效果，我们剥离Introduction之后的部分（如果有）
     paper_meta = page_one_fragments[0].split('introduction')[0].split('Introduction')[0].split('INTRODUCTION')[0]
-    
+
     ############################## <第 1 步，从摘要中提取高价值信息，放到history中> ##################################
     final_results = []
     final_results.append(paper_meta)
@@ -42,10 +42,10 @@ def 解析PDF(file_name, llm_kwargs, plugin_kwargs, chatbot, history, system_pro
         i_say = f"Read this section, recapitulate the content of this section with less than {NUM_OF_WORD} words: {paper_fragments[i]}"
         i_say_show_user = f"[{i+1}/{n_fragment}] Read this section, recapitulate the content of this section with less than {NUM_OF_WORD} words: {paper_fragments[i][:200]} ...."
         gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(i_say, i_say_show_user,  # i_say=真正给chatgpt的提问， i_say_show_user=给用户看的提问
-                                                                           llm_kwargs, chatbot, 
+                                                                           llm_kwargs, chatbot,
                                                                            history=["The main idea of the previous section is?", last_iteration_result], # 迭代上一次的结果
                                                                            sys_prompt="Extract the main idea of this section, answer me with Chinese."  # 提示
-                                                                        ) 
+                                                                        )
         iteration_results.append(gpt_say)
         last_iteration_result = gpt_say
 
@@ -76,8 +76,8 @@ def 理解PDF文档内容标准文件输入(txt, llm_kwargs, plugin_kwargs, chat
     try:
         import fitz
     except:
-        report_exception(chatbot, history, 
-            a = f"解析项目: {txt}", 
+        report_exception(chatbot, history,
+            a = f"解析项目: {txt}",
             b = f"导入软件依赖失败。使用该模块需要额外依赖，安装方法```pip install --upgrade pymupdf```。")
         yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
         return

crazy_functions/生成函数注释.py

@@ -16,7 +16,7 @@ def 生成函数注释(file_manifest, project_folder, llm_kwargs, plugin_kwargs,
         chatbot.append((i_say_show_user, "[Local Message] waiting gpt response."))
         yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
 
-        if not fast_debug: 
+        if not fast_debug:
             msg = '正常'
             # ** gpt request **
             gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
@@ -27,7 +27,7 @@ def 生成函数注释(file_manifest, project_folder, llm_kwargs, plugin_kwargs,
             yield from update_ui(chatbot=chatbot, history=history, msg=msg) # 刷新界面
             if not fast_debug: time.sleep(2)
 
-    if not fast_debug: 
+    if not fast_debug:
         res = write_history_to_file(history)
         promote_file_to_downloadzone(res, chatbot=chatbot)
         chatbot.append(("完成了吗？", res))

crazy_functions/生成多种Mermaid图表.py

@@ -179,15 +179,15 @@ def 解析历史输入(history,llm_kwargs,file_manifest,chatbot,plugin_kwargs):
         i_say = f"Read this section, recapitulate the content of this section with less than {NUM_OF_WORD} words in Chinese: {txt[i]}"
         i_say_show_user = f"[{i+1}/{n_txt}] Read this section, recapitulate the content of this section with less than {NUM_OF_WORD} words: {txt[i][:200]} ...."
         gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(i_say, i_say_show_user,  # i_say=真正给chatgpt的提问， i_say_show_user=给用户看的提问
-                                                                           llm_kwargs, chatbot, 
+                                                                           llm_kwargs, chatbot,
                                                                            history=["The main content of the previous section is?", last_iteration_result], # 迭代上一次的结果
                                                                            sys_prompt="Extracts the main content from the text section where it is located for graphing purposes, answer me with Chinese."  # 提示
-                                                                        ) 
+                                                                        )
         results.append(gpt_say)
         last_iteration_result = gpt_say
     ############################## <第 2 步，根据整理的摘要选择图表类型> ##################################
     if ("advanced_arg" in plugin_kwargs) and (plugin_kwargs["advanced_arg"] == ""): plugin_kwargs.pop("advanced_arg")
-    gpt_say = plugin_kwargs.get("advanced_arg", "")     #将图表类型参数赋值为插件参数  
+    gpt_say = plugin_kwargs.get("advanced_arg", "")     #将图表类型参数赋值为插件参数
     results_txt = '\n'.join(results)    #合并摘要
     if gpt_say not in ['1','2','3','4','5','6','7','8','9']:    #如插件参数不正确则使用对话模型判断
         i_say_show_user = f'接下来将判断适合的图表类型,如连续3次判断失败将会使用流程图进行绘制'; gpt_say = "[Local Message] 收到。"   # 用户提示
@@ -198,7 +198,7 @@ def 解析历史输入(history,llm_kwargs,file_manifest,chatbot,plugin_kwargs):
             gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
                 inputs=i_say,
                 inputs_show_user=i_say_show_user,
-                llm_kwargs=llm_kwargs, chatbot=chatbot, history=[], 
+                llm_kwargs=llm_kwargs, chatbot=chatbot, history=[],
                 sys_prompt=""
             )
             if gpt_say in ['1','2','3','4','5','6','7','8','9']:     #判断返回是否正确
@@ -228,12 +228,12 @@ def 解析历史输入(history,llm_kwargs,file_manifest,chatbot,plugin_kwargs):
     gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
         inputs=i_say,
         inputs_show_user=i_say_show_user,
-        llm_kwargs=llm_kwargs, chatbot=chatbot, history=[], 
+        llm_kwargs=llm_kwargs, chatbot=chatbot, history=[],
         sys_prompt=""
     )
     history.append(gpt_say)
     yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 界面更新
-    
+
 @CatchException
 def 生成多种Mermaid图表(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, web_port):
     """
@@ -249,11 +249,11 @@ def 生成多种Mermaid图表(txt, llm_kwargs, plugin_kwargs, chatbot, history,
 
     # 基本信息：功能、贡献者
     chatbot.append([
-        "函数插件功能？", 
+        "函数插件功能？",
         "根据当前聊天历史或指定的路径文件(文件内容优先)绘制多种mermaid图表，将会由对话模型首先判断适合的图表类型，随后绘制图表。\
         \n您也可以使用插件参数指定绘制的图表类型,函数插件贡献者: Menghuan1918"])
     yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
-    
+
     if os.path.exists(txt):     #如输入区无内容则直接解析历史记录
         from crazy_functions.pdf_fns.parse_word import extract_text_from_files
         file_exist, final_result, page_one, file_manifest, excption = extract_text_from_files(txt, chatbot, history)
@@ -264,15 +264,15 @@ def 生成多种Mermaid图表(txt, llm_kwargs, plugin_kwargs, chatbot, history,
 
     if excption != "":
         if excption == "word":
-            report_exception(chatbot, history, 
-                a = f"解析项目: {txt}", 
+            report_exception(chatbot, history,
+                a = f"解析项目: {txt}",
                 b = f"找到了.doc文件，但是该文件格式不被支持，请先转化为.docx格式。")
-            
+
         elif excption == "pdf":
-            report_exception(chatbot, history, 
-                a = f"解析项目: {txt}", 
+            report_exception(chatbot, history,
+                a = f"解析项目: {txt}",
                 b = f"导入软件依赖失败。使用该模块需要额外依赖，安装方法```pip install --upgrade pymupdf```。")
-        
+
         elif excption == "word_pip":
                 report_exception(chatbot, history,
                     a=f"解析项目: {txt}",

crazy_functions/知识库问答.py

@@ -9,7 +9,7 @@ install_msg ="""
 
 3. python -m pip install unstructured[all-docs] --upgrade
 
-4. python -c 'import nltk; nltk.download("punkt")' 
+4. python -c 'import nltk; nltk.download("punkt")'
 """
 
 @CatchException
@@ -56,7 +56,7 @@ def 知识库文件注入(txt, llm_kwargs, plugin_kwargs, chatbot, history, syst
         chatbot.append(["没有找到任何可读取文件", "当前支持的格式包括: txt, md, docx, pptx, pdf, json等"])
         yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
         return
-    
+
     # < -------------------预热文本向量化模组--------------- >
     chatbot.append(['<br/>'.join(file_manifest), "正在预热文本向量化模组, 如果是第一次运行, 将消耗较长时间下载中文向量化模型..."])
     yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
@@ -109,8 +109,8 @@ def 读取知识库作答(txt, llm_kwargs, plugin_kwargs, chatbot, history, syst
     chatbot.append((txt, f'[知识库 {kai_id}] ' + prompt))
     yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 由于请求gpt需要一段时间，我们先及时地做一次界面更新
     gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-        inputs=prompt, inputs_show_user=txt, 
-        llm_kwargs=llm_kwargs, chatbot=chatbot, history=[], 
+        inputs=prompt, inputs_show_user=txt,
+        llm_kwargs=llm_kwargs, chatbot=chatbot, history=[],
         sys_prompt=system_prompt
     )
     history.extend((prompt, gpt_say))

crazy_functions/联网的ChatGPT.py

@@ -40,10 +40,10 @@ def scrape_text(url, proxies) -> str:
         'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/94.0.4606.61 Safari/537.36',
         'Content-Type': 'text/plain',
     }
-    try: 
+    try:
         response = requests.get(url, headers=headers, proxies=proxies, timeout=8)
         if response.encoding == "ISO-8859-1": response.encoding = response.apparent_encoding
-    except: 
+    except:
         return "无法连接到该网页"
     soup = BeautifulSoup(response.text, "html.parser")
     for script in soup(["script", "style"]):
@@ -66,7 +66,7 @@ def 连接网络回答问题(txt, llm_kwargs, plugin_kwargs, chatbot, history, s
     user_request    当前用户的请求信息（IP地址等）
     """
     history = []    # 清空历史，以免输入溢出
-    chatbot.append((f"请结合互联网信息回答以下问题：{txt}", 
+    chatbot.append((f"请结合互联网信息回答以下问题：{txt}",
                     "[Local Message] 请注意，您正在调用一个[函数插件]的模板，该模板可以实现ChatGPT联网信息综合。该函数面向希望实现更多有趣功能的开发者，它可以作为创建新功能函数的模板。您若希望分享新的功能模组，请不吝PR！"))
     yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 由于请求gpt需要一段时间，我们先及时地做一次界面更新
 
@@ -91,13 +91,13 @@ def 连接网络回答问题(txt, llm_kwargs, plugin_kwargs, chatbot, history, s
     # ------------- < 第3步：ChatGPT综合 > -------------
     i_say = f"从以上搜索结果中抽取信息，然后回答问题：{txt}"
     i_say, history = input_clipping(    # 裁剪输入，从最长的条目开始裁剪，防止爆token
-        inputs=i_say, 
-        history=history, 
+        inputs=i_say,
+        history=history,
         max_token_limit=model_info[llm_kwargs['llm_model']]['max_token']*3//4
     )
     gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-        inputs=i_say, inputs_show_user=i_say, 
-        llm_kwargs=llm_kwargs, chatbot=chatbot, history=history, 
+        inputs=i_say, inputs_show_user=i_say,
+        llm_kwargs=llm_kwargs, chatbot=chatbot, history=history,
         sys_prompt="请从给定的若干条搜索结果中抽取信息，对最相关的两个搜索结果进行总结，然后回答问题。"
     )
     chatbot[-1] = (i_say, gpt_say)

crazy_functions/虚空终端.py

@@ -33,7 +33,7 @@ explain_msg = """
     - 「请调用插件，解析python源代码项目，代码我刚刚打包拖到上传区了」
     - 「请问Transformer网络的结构是怎样的？」
 
-2. 您可以打开插件下拉菜单以了解本项目的各种能力。    
+2. 您可以打开插件下拉菜单以了解本项目的各种能力。
 
 3. 如果您使用「调用插件xxx」、「修改配置xxx」、「请问」等关键词，您的意图可以被识别的更准确。
 
@@ -67,7 +67,7 @@ class UserIntention(BaseModel):
 def chat(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt, user_intention):
     gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
         inputs=txt, inputs_show_user=txt,
-        llm_kwargs=llm_kwargs, chatbot=chatbot, history=[], 
+        llm_kwargs=llm_kwargs, chatbot=chatbot, history=[],
         sys_prompt=system_prompt
     )
     chatbot[-1] = [txt, gpt_say]
@@ -115,7 +115,7 @@ def 虚空终端(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt
     if is_the_upload_folder(txt):
         state.set_state(chatbot=chatbot, key='has_provided_explaination', value=False)
         appendix_msg = "\n\n**很好，您已经上传了文件**，现在请您描述您的需求。"
-        
+
     if is_certain or (state.has_provided_explaination):
         # 如果意图明确，跳过提示环节
         state.set_state(chatbot=chatbot, key='has_provided_explaination', value=True)
@@ -152,7 +152,7 @@ def 虚空终端主路由(txt, llm_kwargs, plugin_kwargs, chatbot, history, syst
         analyze_res = run_gpt_fn(inputs, "")
         try:
             user_intention = gpt_json_io.generate_output_auto_repair(analyze_res, run_gpt_fn)
-            lastmsg=f"正在执行任务: {txt}\n\n用户意图理解: 意图={explain_intention_to_user[user_intention.intention_type]}", 
+            lastmsg=f"正在执行任务: {txt}\n\n用户意图理解: 意图={explain_intention_to_user[user_intention.intention_type]}",
         except JsonStringError as e:
             yield from update_ui_lastest_msg(
                 lastmsg=f"正在执行任务: {txt}\n\n用户意图理解: 失败 当前语言模型（{llm_kwargs['llm_model']}）不能理解您的意图", chatbot=chatbot, history=history, delay=0)
@@ -161,7 +161,7 @@ def 虚空终端主路由(txt, llm_kwargs, plugin_kwargs, chatbot, history, syst
         pass
 
     yield from update_ui_lastest_msg(
-        lastmsg=f"正在执行任务: {txt}\n\n用户意图理解: 意图={explain_intention_to_user[user_intention.intention_type]}", 
+        lastmsg=f"正在执行任务: {txt}\n\n用户意图理解: 意图={explain_intention_to_user[user_intention.intention_type]}",
         chatbot=chatbot, history=history, delay=0)
 
     # 用户意图: 修改本项目的配置

crazy_functions/解析项目源代码.py

@@ -82,13 +82,13 @@ def 解析源代码新(file_manifest, project_folder, llm_kwargs, plugin_kwargs,
             inputs=inputs, inputs_show_user=inputs_show_user, llm_kwargs=llm_kwargs, chatbot=chatbot,
             history=this_iteration_history_feed,   # 迭代之前的分析
             sys_prompt="你是一个程序架构分析师，正在分析一个项目的源代码。" + sys_prompt_additional)
-        
+
         diagram_code = make_diagram(this_iteration_files, result, this_iteration_history_feed)
         summary = "请用一句话概括这些文件的整体功能。\n\n" + diagram_code
         summary_result = yield from request_gpt_model_in_new_thread_with_ui_alive(
-            inputs=summary, 
-            inputs_show_user=summary, 
-            llm_kwargs=llm_kwargs, 
+            inputs=summary,
+            inputs_show_user=summary,
+            llm_kwargs=llm_kwargs,
             chatbot=chatbot,
             history=[i_say, result],   # 迭代之前的分析
             sys_prompt="你是一个程序架构分析师，正在分析一个项目的源代码。" + sys_prompt_additional)
@@ -345,9 +345,12 @@ def 解析任意code项目(txt, llm_kwargs, plugin_kwargs, chatbot, history, sys
     pattern_except_suffix = [_.lstrip(" ^*.,").rstrip(" ,") for _ in txt_pattern.split(" ") if _ != "" and _.strip().startswith("^*.")]
     pattern_except_suffix += ['zip', 'rar', '7z', 'tar', 'gz'] # 避免解析压缩文件
     # 将要忽略匹配的文件名(例如: ^README.md)
-    pattern_except_name = [_.lstrip(" ^*,").rstrip(" ,").replace(".", "\.") for _ in txt_pattern.split(" ") if _ != "" and _.strip().startswith("^") and not _.strip().startswith("^*.")]
+    pattern_except_name = [_.lstrip(" ^*,").rstrip(" ,").replace(".", r"\.") # 移除左边通配符，移除右侧逗号，转义点号
+                           for _ in txt_pattern.split(" ") # 以空格分割
+                           if (_ != "" and _.strip().startswith("^") and not _.strip().startswith("^*."))   # ^开始，但不是^*.开始
+                           ]
     # 生成正则表达式
-    pattern_except = '/[^/]+\.(' + "|".join(pattern_except_suffix) + ')$'
+    pattern_except = r'/[^/]+\.(' + "|".join(pattern_except_suffix) + ')$'
     pattern_except += '|/(' + "|".join(pattern_except_name) + ')$' if pattern_except_name != [] else ''
 
     history.clear()

crazy_functions/询问多个大语言模型.py

@@ -20,8 +20,8 @@ def 同时问询(txt, llm_kwargs, plugin_kwargs, chatbot, history, system_prompt
     # llm_kwargs['llm_model'] = 'chatglm&gpt-3.5-turbo&api2d-gpt-3.5-turbo' # 支持任意数量的llm接口，用&符号分隔
     llm_kwargs['llm_model'] = MULTI_QUERY_LLM_MODELS # 支持任意数量的llm接口，用&符号分隔
     gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-        inputs=txt, inputs_show_user=txt, 
-        llm_kwargs=llm_kwargs, chatbot=chatbot, history=history, 
+        inputs=txt, inputs_show_user=txt,
+        llm_kwargs=llm_kwargs, chatbot=chatbot, history=history,
         sys_prompt=system_prompt,
         retry_times_at_unknown_error=0
     )
@@ -52,8 +52,8 @@ def 同时问询_指定模型(txt, llm_kwargs, plugin_kwargs, chatbot, history,
     yield from update_ui(chatbot=chatbot, history=history) # 刷新界面 # 由于请求gpt需要一段时间，我们先及时地做一次界面更新
 
     gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
-        inputs=txt, inputs_show_user=txt, 
-        llm_kwargs=llm_kwargs, chatbot=chatbot, history=history, 
+        inputs=txt, inputs_show_user=txt,
+        llm_kwargs=llm_kwargs, chatbot=chatbot, history=history,
         sys_prompt=system_prompt,
         retry_times_at_unknown_error=0
     )

crazy_functions/语音助手.py

@@ -39,7 +39,7 @@ class AsyncGptTask():
         try:
             MAX_TOKEN_ALLO = 2560
             i_say, history = input_clipping(i_say, history, max_token_limit=MAX_TOKEN_ALLO)
-            gpt_say_partial = predict_no_ui_long_connection(inputs=i_say, llm_kwargs=llm_kwargs, history=history, sys_prompt=sys_prompt, 
+            gpt_say_partial = predict_no_ui_long_connection(inputs=i_say, llm_kwargs=llm_kwargs, history=history, sys_prompt=sys_prompt,
                                                             observe_window=observe_window[index], console_slience=True)
         except ConnectionAbortedError as token_exceed_err:
             print('至少一个线程任务Token溢出而失败', e)
@@ -120,7 +120,7 @@ class InterviewAssistant(AliyunASR):
             yield from update_ui(chatbot=chatbot, history=history)      # 刷新界面
             self.plugin_wd.feed()
 
-            if self.event_on_result_chg.is_set(): 
+            if self.event_on_result_chg.is_set():
                 # called when some words have finished
                 self.event_on_result_chg.clear()
                 chatbot[-1] = list(chatbot[-1])
@@ -151,7 +151,7 @@ class InterviewAssistant(AliyunASR):
                 # add gpt task 创建子线程请求gpt，避免线程阻塞
                 history = chatbot2history(chatbot)
                 self.agt.add_async_gpt_task(self.buffered_sentence, len(chatbot)-1, llm_kwargs, history, system_prompt)
-                
+
                 self.buffered_sentence = ""
                 chatbot.append(["[ 请讲话 ]", "[ 正在等您说完问题 ]"])
                 yield from update_ui(chatbot=chatbot, history=history) # 刷新界面

crazy_functions/谷歌检索小助手.py

@@ -20,10 +20,10 @@ def get_meta_information(url, chatbot, history):
     proxies = get_conf('proxies')
     headers = {
         'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/116.0.0.0 Safari/537.36',
-        'Accept-Encoding': 'gzip, deflate, br', 
+        'Accept-Encoding': 'gzip, deflate, br',
         'Accept-Language': 'en-US,en;q=0.9,zh-CN;q=0.8,zh;q=0.7',
         'Cache-Control':'max-age=0',
-        'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.7', 
+        'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.7',
         'Connection': 'keep-alive'
     }
     try:
@@ -95,7 +95,7 @@ def get_meta_information(url, chatbot, history):
         )
         try: paper = next(search.results())
         except: paper = None
-        
+
         is_match = paper is not None and string_similar(title, paper.title) > 0.90
 
         # 如果在Arxiv上匹配失败，检索文章的历史版本的题目
@@ -146,8 +146,8 @@ def 谷歌检索小助手(txt, llm_kwargs, plugin_kwargs, chatbot, history, syst
         import math
         from bs4 import BeautifulSoup
     except:
-        report_exception(chatbot, history, 
-            a = f"解析项目: {txt}", 
+        report_exception(chatbot, history,
+            a = f"解析项目: {txt}",
             b = f"导入软件依赖失败。使用该模块需要额外依赖，安装方法```pip install --upgrade beautifulsoup4 arxiv```。")
         yield from update_ui(chatbot=chatbot, history=history) # 刷新界面
         return
@@ -163,7 +163,7 @@ def 谷歌检索小助手(txt, llm_kwargs, plugin_kwargs, chatbot, history, syst
         if len(meta_paper_info_list[:batchsize]) > 0:
             i_say = "下面是一些学术文献的数据，提取出以下内容：" + \
             "1、英文题目；2、中文题目翻译；3、作者；4、arxiv公开（is_paper_in_arxiv）；4、引用数量（cite）；5、中文摘要翻译。" + \
-            f"以下是信息源：{str(meta_paper_info_list[:batchsize])}" 
+            f"以下是信息源：{str(meta_paper_info_list[:batchsize])}"
 
             inputs_show_user = f"请分析此页面中出现的所有文章：{txt}，这是第{batch+1}批"
             gpt_say = yield from request_gpt_model_in_new_thread_with_ui_alive(
@@ -175,11 +175,11 @@ def 谷歌检索小助手(txt, llm_kwargs, plugin_kwargs, chatbot, history, syst
             history.extend([ f"第{batch+1}批", gpt_say ])
             meta_paper_info_list = meta_paper_info_list[batchsize:]
 
-    chatbot.append(["状态？", 
+    chatbot.append(["状态？",
         "已经全部完成，您可以试试让AI写一个Related Works，例如您可以继续输入Write a \"Related Works\" section about \"你搜索的研究领域\" for me."])
     msg = '正常'
     yield from update_ui(chatbot=chatbot, history=history, msg=msg) # 刷新界面
     path = write_history_to_file(history)
     promote_file_to_downloadzone(path, chatbot=chatbot)
-    chatbot.append(("完成了吗？", path)); 
+    chatbot.append(("完成了吗？", path));
     yield from update_ui(chatbot=chatbot, history=history, msg=msg) # 刷新界面