Version 3.74: Merge latest updates on dev branch (frontier) (#1621)

* Update version to 3.74

* Add support for Yi Model API (#1635)

* 更新以支持零一万物模型

* 删除newbing

* 修改config

---------

Co-authored-by: binary-husky <qingxu.fu@outlook.com>

* Refactor function signatures in bridge files

* fix qwen api change

* rename and ref functions

* rename and move some cookie functions

* 增加haiku模型，新增endpoint配置说明 (#1626)

* haiku added

* 新增haiku，新增endpoint配置说明

* Haiku added

* 将说明同步至最新Endpoint

---------

Co-authored-by: binary-husky <qingxu.fu@outlook.com>

* private_upload目录下进行文件鉴权 (#1596)

* private_upload目录下进行文件鉴权

* minor fastapi adjustment

* Add logging functionality to enable saving
conversation records

* waiting to fix username retrieve

* support 2rd web path

* allow accessing default user dir

---------

Co-authored-by: binary-husky <qingxu.fu@outlook.com>

* remove yaml deps

* fix favicon

* fix abs path auth problem

* forget to write a return

* add `dashscope` to deps

* fix GHSA-v9q9-xj86-953p

* 用户名重叠越权访问patch (#1681)

* add cohere model api access

* cohere + can_multi_thread

* fix block user access(fail)

* fix fastapi bug

* change cohere api endpoint

* explain version

---------

Co-authored-by: Menghuan1918 <menghuan2003@outlook.com>
Co-authored-by: Skyzayre <120616113+Skyzayre@users.noreply.github.com>
Co-authored-by: XIao <46100050+Kilig947@users.noreply.github.com>

request_llms/bridge_all.py

@@ -34,6 +34,9 @@ from .bridge_google_gemini import predict_no_ui_long_connection  as genai_noui
 from .bridge_zhipu import predict_no_ui_long_connection as zhipu_noui
 from .bridge_zhipu import predict as zhipu_ui
 
+from .bridge_cohere import predict as cohere_ui
+from .bridge_cohere import predict_no_ui_long_connection as cohere_noui
+
 colors = ['#FF00FF', '#00FFFF', '#FF0000', '#990099', '#009999', '#990044']
 
 class LazyloadTiktoken(object):
@@ -64,6 +67,7 @@ newbing_endpoint = "wss://sydney.bing.com/sydney/ChatHub"
 gemini_endpoint = "https://generativelanguage.googleapis.com/v1beta/models"
 claude_endpoint = "https://api.anthropic.com/v1/messages"
 yimodel_endpoint = "https://api.lingyiwanwu.com/v1/chat/completions"
+cohere_endpoint = 'https://api.cohere.ai/v1/chat'
 
 if not AZURE_ENDPOINT.endswith('/'): AZURE_ENDPOINT += '/'
 azure_endpoint = AZURE_ENDPOINT + f'openai/deployments/{AZURE_ENGINE}/chat/completions?api-version=2023-05-15'
@@ -82,6 +86,7 @@ if newbing_endpoint in API_URL_REDIRECT: newbing_endpoint = API_URL_REDIRECT[new
 if gemini_endpoint in API_URL_REDIRECT: gemini_endpoint = API_URL_REDIRECT[gemini_endpoint]
 if claude_endpoint in API_URL_REDIRECT: claude_endpoint = API_URL_REDIRECT[claude_endpoint]
 if yimodel_endpoint in API_URL_REDIRECT: yimodel_endpoint = API_URL_REDIRECT[yimodel_endpoint]
+if cohere_endpoint in API_URL_REDIRECT: cohere_endpoint = API_URL_REDIRECT[cohere_endpoint]
 
 # 获取tokenizer
 tokenizer_gpt35 = LazyloadTiktoken("gpt-3.5-turbo")
@@ -310,6 +315,18 @@ model_info = {
         "tokenizer": tokenizer_gpt35,
         "token_cnt": get_token_num_gpt35,
     },
+
+    # cohere
+    "cohere-command-r-plus": {
+        "fn_with_ui": cohere_ui,
+        "fn_without_ui": cohere_noui,
+        "can_multi_thread": True,
+        "endpoint": cohere_endpoint,
+        "max_token": 1024 * 4,
+        "tokenizer": tokenizer_gpt35,
+        "token_cnt": get_token_num_gpt35,
+    },
+
 }
 # -=-=-=-=-=-=- 月之暗面 -=-=-=-=-=-=-
 from request_llms.bridge_moonshot import predict as moonshot_ui
@@ -359,7 +376,7 @@ for model in AVAIL_LLM_MODELS:
 
 # -=-=-=-=-=-=- 以下部分是新加入的模型，可能附带额外依赖 -=-=-=-=-=-=-
 # claude家族
-claude_models = ["claude-instant-1.2","claude-2.0","claude-2.1","claude-3-sonnet-20240229","claude-3-opus-20240229"]
+claude_models = ["claude-instant-1.2","claude-2.0","claude-2.1","claude-3-haiku-20240307","claude-3-sonnet-20240229","claude-3-opus-20240229"]
 if any(item in claude_models for item in AVAIL_LLM_MODELS):
     from .bridge_claude import predict_no_ui_long_connection as claude_noui
     from .bridge_claude import predict as claude_ui
@@ -393,6 +410,16 @@ if any(item in claude_models for item in AVAIL_LLM_MODELS):
             "token_cnt": get_token_num_gpt35,
         },
     })
+    model_info.update({
+        "claude-3-haiku-20240307": {
+            "fn_with_ui": claude_ui,
+            "fn_without_ui": claude_noui,
+            "endpoint": claude_endpoint,
+            "max_token": 200000,
+            "tokenizer": tokenizer_gpt35,
+            "token_cnt": get_token_num_gpt35,
+        },
+    })
     model_info.update({
         "claude-3-sonnet-20240229": {
             "fn_with_ui": claude_ui,
@@ -789,7 +816,7 @@ def LLM_CATCH_EXCEPTION(f):
     """
     装饰器函数，将错误显示出来
     """
-    def decorated(inputs, llm_kwargs, history, sys_prompt, observe_window, console_slience):
+    def decorated(inputs:str, llm_kwargs:dict, history:list, sys_prompt:str, observe_window:list, console_slience:bool):
         try:
             return f(inputs, llm_kwargs, history, sys_prompt, observe_window, console_slience)
         except Exception as e:
@@ -799,9 +826,9 @@ def LLM_CATCH_EXCEPTION(f):
     return decorated
 
 
-def predict_no_ui_long_connection(inputs, llm_kwargs, history, sys_prompt, observe_window=[], console_slience=False):
+def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list, sys_prompt:str, observe_window:list=[], console_slience:bool=False):
     """
-    发送至LLM，等待回复，一次性完成，不显示中间过程。但内部用stream的方法避免中途网线被掐。
+    发送至LLM，等待回复，一次性完成，不显示中间过程。但内部（尽可能地）用stream的方法避免中途网线被掐。
     inputs：
         是本次问询的输入
     sys_prompt:
@@ -819,7 +846,6 @@ def predict_no_ui_long_connection(inputs, llm_kwargs, history, sys_prompt, obser
     model = llm_kwargs['llm_model']
     n_model = 1
     if '&' not in model:
-        assert not model.startswith("tgui"), "TGUI不支持函数插件的实现"
 
         # 如果只询问1个大语言模型：
         method = model_info[model]["fn_without_ui"]
@@ -880,15 +906,22 @@ def predict_no_ui_long_connection(inputs, llm_kwargs, history, sys_prompt, obser
         return res
 
 
-def predict(inputs, llm_kwargs, *args, **kwargs):
+def predict(inputs:str, llm_kwargs:dict, *args, **kwargs):
     """
     发送至LLM，流式获取输出。
     用于基础的对话功能。
-    inputs 是本次问询的输入
-    top_p, temperature是LLM的内部调优参数
-    history 是之前的对话列表（注意无论是inputs还是history，内容太长了都会触发token数量溢出的错误）
-    chatbot 为WebUI中显示的对话列表，修改它，然后yeild出去，可以直接修改对话界面内容
-    additional_fn代表点击的哪个按钮，按钮见functional.py
+
+    完整参数列表：
+        predict(
+            inputs:str,                     # 是本次问询的输入
+            llm_kwargs:dict,                # 是LLM的内部调优参数
+            plugin_kwargs:dict,             # 是插件的内部参数
+            chatbot:ChatBotWithCookies,     # 原样传递，负责向用户前端展示对话，兼顾前端状态的功能
+            history:list=[],                # 是之前的对话列表
+            system_prompt:str='',           # 系统静默prompt
+            stream:bool=True,               # 是否流式输出（已弃用）
+            additional_fn:str=None          # 基础功能区按钮的附加功能
+        ):
     """
 
     inputs = apply_gpt_academic_string_mask(inputs, mode="show_llm")

request_llms/bridge_chatglmft.py

@@ -137,7 +137,8 @@ class GetGLMFTHandle(Process):
 global glmft_handle
 glmft_handle = None
 #################################################################################
-def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=[], console_slience=False):
+def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="",
+                                  observe_window:list=[], console_slience:bool=False):
     """
         多线程方法
         函数的说明请见 request_llms/bridge_all.py

request_llms/bridge_chatgpt.py

@@ -23,6 +23,7 @@ import random
 # 读取时首先看是否存在私密的config_private配置文件（不受git管控），如果有，则覆盖原config文件
 from toolbox import get_conf, update_ui, is_any_api_key, select_api_key, what_keys, clip_history
 from toolbox import trimmed_format_exc, is_the_upload_folder, read_one_api_model_name, log_chat
+from toolbox import ChatBotWithCookies
 proxies, TIMEOUT_SECONDS, MAX_RETRY, API_ORG, AZURE_CFG_ARRAY = \
     get_conf('proxies', 'TIMEOUT_SECONDS', 'MAX_RETRY', 'API_ORG', 'AZURE_CFG_ARRAY')
 
@@ -69,7 +70,7 @@ def verify_endpoint(endpoint):
         raise ValueError("Endpoint不正确, 请检查AZURE_ENDPOINT的配置! 当前的Endpoint为:" + endpoint)
     return endpoint
 
-def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=None, console_slience=False):
+def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="", observe_window:list=None, console_slience:bool=False):
     """
     发送至chatGPT，等待回复，一次性完成，不显示中间过程。但内部用stream的方法避免中途网线被掐。
     inputs：
@@ -147,7 +148,8 @@ def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="",
     return result
 
 
-def predict(inputs, llm_kwargs, plugin_kwargs, chatbot, history=[], system_prompt='', stream = True, additional_fn=None):
+def predict(inputs:str, llm_kwargs:dict, plugin_kwargs:dict, chatbot:ChatBotWithCookies,
+            history:list=[], system_prompt:str='', stream:bool=True, additional_fn:str=None):
     """
     发送至chatGPT，流式获取输出。
     用于基础的对话功能。

request_llms/bridge_claude.py

@@ -13,11 +13,11 @@ import logging
 import os
 import time
 import traceback
-from toolbox import get_conf, update_ui, trimmed_format_exc, encode_image, every_image_file_in_path
 import json
 import requests
+from toolbox import get_conf, update_ui, trimmed_format_exc, encode_image, every_image_file_in_path, log_chat
 picture_system_prompt = "\n当回复图像时,必须说明正在回复哪张图像。所有图像仅在最后一个问题中提供,即使它们在历史记录中被提及。请使用'这是第X张图像:'的格式来指明您正在描述的是哪张图像。"
-Claude_3_Models = ["claude-3-sonnet-20240229", "claude-3-opus-20240229"]
+Claude_3_Models = ["claude-3-haiku-20240307", "claude-3-sonnet-20240229", "claude-3-opus-20240229"]
 
 # config_private.py放自己的秘密如API和代理网址
 # 读取时首先看是否存在私密的config_private配置文件（不受git管控），如果有，则覆盖原config文件
@@ -95,7 +95,7 @@ def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="",
             # make a POST request to the API endpoint, stream=False
             from .bridge_all import model_info
             endpoint = model_info[llm_kwargs['llm_model']]['endpoint']
-            response = requests.post(endpoint, headers=headers, json=message, 
+            response = requests.post(endpoint, headers=headers, json=message,
                                      proxies=proxies, stream=True, timeout=TIMEOUT_SECONDS);break
         except requests.exceptions.ReadTimeout as e:
             retry += 1
@@ -116,7 +116,7 @@ def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="",
                 if need_to_pass:
                     pass
                 elif is_last_chunk:
-                    logging.info(f'[response] {result}')
+                    # logging.info(f'[response] {result}')
                     break
                 else:
                     if chunkjson and chunkjson['type'] == 'content_block_delta':
@@ -194,7 +194,7 @@ def predict(inputs, llm_kwargs, plugin_kwargs, chatbot, history=[], system_promp
             # make a POST request to the API endpoint, stream=True
             from .bridge_all import model_info
             endpoint = model_info[llm_kwargs['llm_model']]['endpoint']
-            response = requests.post(endpoint, headers=headers, json=message, 
+            response = requests.post(endpoint, headers=headers, json=message,
                                      proxies=proxies, stream=True, timeout=TIMEOUT_SECONDS);break
         except requests.exceptions.ReadTimeout as e:
             retry += 1
@@ -216,7 +216,8 @@ def predict(inputs, llm_kwargs, plugin_kwargs, chatbot, history=[], system_promp
                 if need_to_pass:
                     pass
                 elif is_last_chunk:
-                    logging.info(f'[response] {gpt_replying_buffer}')
+                    log_chat(llm_model=llm_kwargs["llm_model"], input_str=inputs, output_str=gpt_replying_buffer)
+                    # logging.info(f'[response] {gpt_replying_buffer}')
                     break
                 else:
                     if chunkjson and chunkjson['type'] == 'content_block_delta':
@@ -305,4 +306,4 @@ def generate_payload(inputs, llm_kwargs, history, system_prompt, image_paths):
         'stream': True,
         'system': system_prompt
     }
-    return headers, payload
+    return headers, payload

request_llms/bridge_cohere.py

@@ -0,0 +1,328 @@
+# 借鉴了 https://github.com/GaiZhenbiao/ChuanhuChatGPT 项目
+
+"""
+    该文件中主要包含三个函数
+
+    不具备多线程能力的函数：
+    1. predict: 正常对话时使用，具备完备的交互功能，不可多线程
+
+    具备多线程调用能力的函数
+    2. predict_no_ui_long_connection：支持多线程
+"""
+
+import json
+import time
+import gradio as gr
+import logging
+import traceback
+import requests
+import importlib
+import random
+
+# config_private.py放自己的秘密如API和代理网址
+# 读取时首先看是否存在私密的config_private配置文件（不受git管控），如果有，则覆盖原config文件
+from toolbox import get_conf, update_ui, is_any_api_key, select_api_key, what_keys, clip_history
+from toolbox import trimmed_format_exc, is_the_upload_folder, read_one_api_model_name, log_chat
+from toolbox import ChatBotWithCookies
+proxies, TIMEOUT_SECONDS, MAX_RETRY, API_ORG, AZURE_CFG_ARRAY = \
+    get_conf('proxies', 'TIMEOUT_SECONDS', 'MAX_RETRY', 'API_ORG', 'AZURE_CFG_ARRAY')
+
+timeout_bot_msg = '[Local Message] Request timeout. Network error. Please check proxy settings in config.py.' + \
+                  '网络错误，检查代理服务器是否可用，以及代理设置的格式是否正确，格式须是[协议]://[地址]:[端口]，缺一不可。'
+
+def get_full_error(chunk, stream_response):
+    """
+        获取完整的从Cohere返回的报错
+    """
+    while True:
+        try:
+            chunk += next(stream_response)
+        except:
+            break
+    return chunk
+
+def decode_chunk(chunk):
+    # 提前读取一些信息 （用于判断异常）
+    chunk_decoded = chunk.decode()
+    chunkjson = None
+    has_choices = False
+    choice_valid = False
+    has_content = False
+    has_role = False
+    try:
+        chunkjson = json.loads(chunk_decoded)
+        has_choices = 'choices' in chunkjson
+        if has_choices: choice_valid = (len(chunkjson['choices']) > 0)
+        if has_choices and choice_valid: has_content = ("content" in chunkjson['choices'][0]["delta"])
+        if has_content: has_content = (chunkjson['choices'][0]["delta"]["content"] is not None)
+        if has_choices and choice_valid: has_role = "role" in chunkjson['choices'][0]["delta"]
+    except:
+        pass
+    return chunk_decoded, chunkjson, has_choices, choice_valid, has_content, has_role
+
+from functools import lru_cache
+@lru_cache(maxsize=32)
+def verify_endpoint(endpoint):
+    """
+        检查endpoint是否可用
+    """
+    if "你亲手写的api名称" in endpoint:
+        raise ValueError("Endpoint不正确, 请检查AZURE_ENDPOINT的配置! 当前的Endpoint为:" + endpoint)
+    return endpoint
+
+def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="", observe_window:list=None, console_slience:bool=False):
+    """
+    发送，等待回复，一次性完成，不显示中间过程。但内部用stream的方法避免中途网线被掐。
+    inputs：
+        是本次问询的输入
+    sys_prompt:
+        系统静默prompt
+    llm_kwargs：
+        内部调优参数
+    history：
+        是之前的对话列表
+    observe_window = None：
+        用于负责跨越线程传递已经输出的部分，大部分时候仅仅为了fancy的视觉效果，留空即可。observe_window[0]：观测窗。observe_window[1]：看门狗
+    """
+    watch_dog_patience = 5 # 看门狗的耐心, 设置5秒即可
+    headers, payload = generate_payload(inputs, llm_kwargs, history, system_prompt=sys_prompt, stream=True)
+    retry = 0
+    while True:
+        try:
+            # make a POST request to the API endpoint, stream=False
+            from .bridge_all import model_info
+            endpoint = verify_endpoint(model_info[llm_kwargs['llm_model']]['endpoint'])
+            response = requests.post(endpoint, headers=headers, proxies=proxies,
+                                    json=payload, stream=True, timeout=TIMEOUT_SECONDS); break
+        except requests.exceptions.ReadTimeout as e:
+            retry += 1
+            traceback.print_exc()
+            if retry > MAX_RETRY: raise TimeoutError
+            if MAX_RETRY!=0: print(f'请求超时，正在重试 ({retry}/{MAX_RETRY}) ……')
+
+    stream_response = response.iter_lines()
+    result = ''
+    json_data = None
+    while True:
+        try: chunk = next(stream_response)
+        except StopIteration:
+            break
+        except requests.exceptions.ConnectionError:
+            chunk = next(stream_response) # 失败了，重试一次？再失败就没办法了。
+        chunk_decoded, chunkjson, has_choices, choice_valid, has_content, has_role = decode_chunk(chunk)
+        if chunkjson['event_type'] == 'stream-start': continue
+        if chunkjson['event_type'] == 'text-generation':
+            result += chunkjson["text"]
+            if not console_slience: print(chunkjson["text"], end='')
+            if observe_window is not None:
+                # 观测窗，把已经获取的数据显示出去
+                if len(observe_window) >= 1:
+                    observe_window[0] += chunkjson["text"]
+                # 看门狗，如果超过期限没有喂狗，则终止
+                if len(observe_window) >= 2:
+                    if (time.time()-observe_window[1]) > watch_dog_patience:
+                        raise RuntimeError("用户取消了程序。")
+        if chunkjson['event_type'] == 'stream-end': break
+    return result
+
+
+def predict(inputs:str, llm_kwargs:dict, plugin_kwargs:dict, chatbot:ChatBotWithCookies,
+            history:list=[], system_prompt:str='', stream:bool=True, additional_fn:str=None):
+    """
+    发送至chatGPT，流式获取输出。
+    用于基础的对话功能。
+    inputs 是本次问询的输入
+    top_p, temperature是chatGPT的内部调优参数
+    history 是之前的对话列表（注意无论是inputs还是history，内容太长了都会触发token数量溢出的错误）
+    chatbot 为WebUI中显示的对话列表，修改它，然后yeild出去，可以直接修改对话界面内容
+    additional_fn代表点击的哪个按钮，按钮见functional.py
+    """
+    # if is_any_api_key(inputs):
+    #     chatbot._cookies['api_key'] = inputs
+    #     chatbot.append(("输入已识别为Cohere的api_key", what_keys(inputs)))
+    #     yield from update_ui(chatbot=chatbot, history=history, msg="api_key已导入") # 刷新界面
+    #     return
+    # elif not is_any_api_key(chatbot._cookies['api_key']):
+    #     chatbot.append((inputs, "缺少api_key。\n\n1. 临时解决方案：直接在输入区键入api_key，然后回车提交。\n\n2. 长效解决方案：在config.py中配置。"))
+    #     yield from update_ui(chatbot=chatbot, history=history, msg="缺少api_key") # 刷新界面
+    #     return
+
+    user_input = inputs
+    if additional_fn is not None:
+        from core_functional import handle_core_functionality
+        inputs, history = handle_core_functionality(additional_fn, inputs, history, chatbot)
+
+    raw_input = inputs
+    # logging.info(f'[raw_input] {raw_input}')
+    chatbot.append((inputs, ""))
+    yield from update_ui(chatbot=chatbot, history=history, msg="等待响应") # 刷新界面
+
+    # check mis-behavior
+    if is_the_upload_folder(user_input):
+        chatbot[-1] = (inputs, f"[Local Message] 检测到操作错误！当您上传文档之后，需点击“**函数插件区**”按钮进行处理，请勿点击“提交”按钮或者“基础功能区”按钮。")
+        yield from update_ui(chatbot=chatbot, history=history, msg="正常") # 刷新界面
+        time.sleep(2)
+
+    try:
+        headers, payload = generate_payload(inputs, llm_kwargs, history, system_prompt, stream)
+    except RuntimeError as e:
+        chatbot[-1] = (inputs, f"您提供的api-key不满足要求，不包含任何可用于{llm_kwargs['llm_model']}的api-key。您可能选择了错误的模型或请求源。")
+        yield from update_ui(chatbot=chatbot, history=history, msg="api-key不满足要求") # 刷新界面
+        return
+
+    # 检查endpoint是否合法
+    try:
+        from .bridge_all import model_info
+        endpoint = verify_endpoint(model_info[llm_kwargs['llm_model']]['endpoint'])
+    except:
+        tb_str = '```\n' + trimmed_format_exc() + '```'
+        chatbot[-1] = (inputs, tb_str)
+        yield from update_ui(chatbot=chatbot, history=history, msg="Endpoint不满足要求") # 刷新界面
+        return
+
+    history.append(inputs); history.append("")
+
+    retry = 0
+    while True:
+        try:
+            # make a POST request to the API endpoint, stream=True
+            response = requests.post(endpoint, headers=headers, proxies=proxies,
+                                    json=payload, stream=True, timeout=TIMEOUT_SECONDS);break
+        except:
+            retry += 1
+            chatbot[-1] = ((chatbot[-1][0], timeout_bot_msg))
+            retry_msg = f"，正在重试 ({retry}/{MAX_RETRY}) ……" if MAX_RETRY > 0 else ""
+            yield from update_ui(chatbot=chatbot, history=history, msg="请求超时"+retry_msg) # 刷新界面
+            if retry > MAX_RETRY: raise TimeoutError
+
+    gpt_replying_buffer = ""
+
+    is_head_of_the_stream = True
+    if stream:
+        stream_response =  response.iter_lines()
+        while True:
+            try:
+                chunk = next(stream_response)
+            except StopIteration:
+                # 非Cohere官方接口的出现这样的报错，Cohere和API2D不会走这里
+                chunk_decoded = chunk.decode()
+                error_msg = chunk_decoded
+                # 其他情况，直接返回报错
+                chatbot, history = handle_error(inputs, llm_kwargs, chatbot, history, chunk_decoded, error_msg)
+                yield from update_ui(chatbot=chatbot, history=history, msg="非Cohere官方接口返回了错误:" + chunk.decode()) # 刷新界面
+                return
+
+            # 提前读取一些信息 （用于判断异常）
+            chunk_decoded, chunkjson, has_choices, choice_valid, has_content, has_role = decode_chunk(chunk)
+
+            if chunkjson:
+                try:
+                    if chunkjson['event_type'] == 'stream-start':
+                        continue
+                    if chunkjson['event_type'] == 'text-generation':
+                        gpt_replying_buffer = gpt_replying_buffer + chunkjson["text"]
+                        history[-1] = gpt_replying_buffer
+                        chatbot[-1] = (history[-2], history[-1])
+                        yield from update_ui(chatbot=chatbot, history=history, msg="正常") # 刷新界面
+                    if chunkjson['event_type'] == 'stream-end':
+                        log_chat(llm_model=llm_kwargs["llm_model"], input_str=inputs, output_str=gpt_replying_buffer)
+                        history[-1] = gpt_replying_buffer
+                        chatbot[-1] = (history[-2], history[-1])
+                        yield from update_ui(chatbot=chatbot, history=history, msg="正常") # 刷新界面
+                        break
+                except Exception as e:
+                    yield from update_ui(chatbot=chatbot, history=history, msg="Json解析不合常规") # 刷新界面
+                    chunk = get_full_error(chunk, stream_response)
+                    chunk_decoded = chunk.decode()
+                    error_msg = chunk_decoded
+                    chatbot, history = handle_error(inputs, llm_kwargs, chatbot, history, chunk_decoded, error_msg)
+                    yield from update_ui(chatbot=chatbot, history=history, msg="Json异常" + error_msg) # 刷新界面
+                    print(error_msg)
+                    return
+
+def handle_error(inputs, llm_kwargs, chatbot, history, chunk_decoded, error_msg):
+    from .bridge_all import model_info
+    Cohere_website = ' 请登录Cohere查看详情 https://platform.Cohere.com/signup'
+    if "reduce the length" in error_msg:
+        if len(history) >= 2: history[-1] = ""; history[-2] = "" # 清除当前溢出的输入：history[-2] 是本次输入, history[-1] 是本次输出
+        history = clip_history(inputs=inputs, history=history, tokenizer=model_info[llm_kwargs['llm_model']]['tokenizer'],
+                                               max_token_limit=(model_info[llm_kwargs['llm_model']]['max_token'])) # history至少释放二分之一
+        chatbot[-1] = (chatbot[-1][0], "[Local Message] Reduce the length. 本次输入过长, 或历史数据过长. 历史缓存数据已部分释放, 您可以请再次尝试. (若再次失败则更可能是因为输入过长.)")
+    elif "does not exist" in error_msg:
+        chatbot[-1] = (chatbot[-1][0], f"[Local Message] Model {llm_kwargs['llm_model']} does not exist. 模型不存在, 或者您没有获得体验资格.")
+    elif "Incorrect API key" in error_msg:
+        chatbot[-1] = (chatbot[-1][0], "[Local Message] Incorrect API key. Cohere以提供了不正确的API_KEY为由, 拒绝服务. " + Cohere_website)
+    elif "exceeded your current quota" in error_msg:
+        chatbot[-1] = (chatbot[-1][0], "[Local Message] You exceeded your current quota. Cohere以账户额度不足为由, 拒绝服务." + Cohere_website)
+    elif "account is not active" in error_msg:
+        chatbot[-1] = (chatbot[-1][0], "[Local Message] Your account is not active. Cohere以账户失效为由, 拒绝服务." + Cohere_website)
+    elif "associated with a deactivated account" in error_msg:
+        chatbot[-1] = (chatbot[-1][0], "[Local Message] You are associated with a deactivated account. Cohere以账户失效为由, 拒绝服务." + Cohere_website)
+    elif "API key has been deactivated" in error_msg:
+        chatbot[-1] = (chatbot[-1][0], "[Local Message] API key has been deactivated. Cohere以账户失效为由, 拒绝服务." + Cohere_website)
+    elif "bad forward key" in error_msg:
+        chatbot[-1] = (chatbot[-1][0], "[Local Message] Bad forward key. API2D账户额度不足.")
+    elif "Not enough point" in error_msg:
+        chatbot[-1] = (chatbot[-1][0], "[Local Message] Not enough point. API2D账户点数不足.")
+    else:
+        from toolbox import regular_txt_to_markdown
+        tb_str = '```\n' + trimmed_format_exc() + '```'
+        chatbot[-1] = (chatbot[-1][0], f"[Local Message] 异常 \n\n{tb_str} \n\n{regular_txt_to_markdown(chunk_decoded)}")
+    return chatbot, history
+
+def generate_payload(inputs, llm_kwargs, history, system_prompt, stream):
+    """
+    整合所有信息，选择LLM模型，生成http请求，为发送请求做准备
+    """
+    # if not is_any_api_key(llm_kwargs['api_key']):
+    #     raise AssertionError("你提供了错误的API_KEY。\n\n1. 临时解决方案：直接在输入区键入api_key，然后回车提交。\n\n2. 长效解决方案：在config.py中配置。")
+
+    api_key = select_api_key(llm_kwargs['api_key'], llm_kwargs['llm_model'])
+
+    headers = {
+        "Content-Type": "application/json",
+        "Authorization": f"Bearer {api_key}"
+    }
+    if API_ORG.startswith('org-'): headers.update({"Cohere-Organization": API_ORG})
+    if llm_kwargs['llm_model'].startswith('azure-'):
+        headers.update({"api-key": api_key})
+        if llm_kwargs['llm_model'] in AZURE_CFG_ARRAY.keys():
+            azure_api_key_unshared = AZURE_CFG_ARRAY[llm_kwargs['llm_model']]["AZURE_API_KEY"]
+            headers.update({"api-key": azure_api_key_unshared})
+
+    conversation_cnt = len(history) // 2
+
+    messages = [{"role": "SYSTEM", "message": system_prompt}]
+    if conversation_cnt:
+        for index in range(0, 2*conversation_cnt, 2):
+            what_i_have_asked = {}
+            what_i_have_asked["role"] = "USER"
+            what_i_have_asked["message"] = history[index]
+            what_gpt_answer = {}
+            what_gpt_answer["role"] = "CHATBOT"
+            what_gpt_answer["message"] = history[index+1]
+            if what_i_have_asked["message"] != "":
+                if what_gpt_answer["message"] == "": continue
+                if what_gpt_answer["message"] == timeout_bot_msg: continue
+                messages.append(what_i_have_asked)
+                messages.append(what_gpt_answer)
+            else:
+                messages[-1]['message'] = what_gpt_answer['message']
+
+    model = llm_kwargs['llm_model']
+    if model.startswith('cohere-'): model = model[len('cohere-'):]
+    payload = {
+        "model": model,
+        "message": inputs,
+        "chat_history": messages,
+        "temperature": llm_kwargs['temperature'],  # 1.0,
+        "top_p": llm_kwargs['top_p'],  # 1.0,
+        "n": 1,
+        "stream": stream,
+        "presence_penalty": 0,
+        "frequency_penalty": 0,
+    }
+
+    return headers,payload
+
+

request_llms/bridge_google_gemini.py

@@ -7,6 +7,7 @@ import re
 import os
 import time
 from request_llms.com_google import GoogleChatInit
+from toolbox import ChatBotWithCookies
 from toolbox import get_conf, update_ui, update_ui_lastest_msg, have_any_recent_upload_image_files, trimmed_format_exc
 
 proxies, TIMEOUT_SECONDS, MAX_RETRY = get_conf('proxies', 'TIMEOUT_SECONDS', 'MAX_RETRY')
@@ -44,7 +45,8 @@ def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="",
     return gpt_replying_buffer
 
 
-def predict(inputs, llm_kwargs, plugin_kwargs, chatbot, history=[], system_prompt='', stream=True, additional_fn=None):
+def predict(inputs:str, llm_kwargs:dict, plugin_kwargs:dict, chatbot:ChatBotWithCookies,
+            history:list=[], system_prompt:str='', stream:bool=True, additional_fn:str=None):
     # 检查API_KEY
     if get_conf("GEMINI_API_KEY") == "":
         yield from update_ui_lastest_msg(f"请配置 GEMINI_API_KEY。", chatbot=chatbot, history=history, delay=0)

request_llms/bridge_jittorllms_llama.py

@@ -1,10 +1,10 @@
 
-from transformers import AutoModel, AutoTokenizer
 import time
 import threading
 import importlib
 from toolbox import update_ui, get_conf
 from multiprocessing import Process, Pipe
+from transformers import AutoModel, AutoTokenizer
 
 load_message = "jittorllms尚未加载，加载需要一段时间。注意，请避免混用多种jittor模型，否则可能导致显存溢出而造成卡顿，取决于`config.py`的配置，jittorllms消耗大量的内存（CPU）或显存（GPU），也许会导致低配计算机卡死 ……"
 
@@ -106,7 +106,8 @@ class GetGLMHandle(Process):
 global llama_glm_handle
 llama_glm_handle = None
 #################################################################################
-def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=[], console_slience=False):
+def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="",
+                                  observe_window:list=[], console_slience:bool=False):
     """
         多线程方法
         函数的说明请见 request_llms/bridge_all.py

request_llms/bridge_jittorllms_pangualpha.py

@@ -1,10 +1,10 @@
 
-from transformers import AutoModel, AutoTokenizer
 import time
 import threading
 import importlib
 from toolbox import update_ui, get_conf
 from multiprocessing import Process, Pipe
+from transformers import AutoModel, AutoTokenizer
 
 load_message = "jittorllms尚未加载，加载需要一段时间。注意，请避免混用多种jittor模型，否则可能导致显存溢出而造成卡顿，取决于`config.py`的配置，jittorllms消耗大量的内存（CPU）或显存（GPU），也许会导致低配计算机卡死 ……"
 
@@ -106,7 +106,8 @@ class GetGLMHandle(Process):
 global pangu_glm_handle
 pangu_glm_handle = None
 #################################################################################
-def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=[], console_slience=False):
+def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="",
+                                  observe_window:list=[], console_slience:bool=False):
     """
         多线程方法
         函数的说明请见 request_llms/bridge_all.py

request_llms/bridge_jittorllms_rwkv.py

@@ -106,7 +106,8 @@ class GetGLMHandle(Process):
 global rwkv_glm_handle
 rwkv_glm_handle = None
 #################################################################################
-def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=[], console_slience=False):
+def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="",
+                                  observe_window:list=[], console_slience:bool=False):
     """
         多线程方法
         函数的说明请见 request_llms/bridge_all.py

request_llms/bridge_moonshot.py

@@ -8,6 +8,7 @@ import time
 import logging
 
 from toolbox import get_conf, update_ui, log_chat
+from toolbox import ChatBotWithCookies
 
 import requests
 
@@ -146,7 +147,8 @@ def msg_handle_error(llm_kwargs, chunk_decoded):
     return error_msg
 
 
-def predict(inputs, llm_kwargs, plugin_kwargs, chatbot, history=[], system_prompt='', stream=True, additional_fn=None):
+def predict(inputs:str, llm_kwargs:dict, plugin_kwargs:dict, chatbot:ChatBotWithCookies,
+            history:list=[], system_prompt:str='', stream:bool=True, additional_fn:str=None):
     chatbot.append([inputs, ""])
 
     if additional_fn is not None:

request_llms/bridge_moss.py

@@ -171,7 +171,8 @@ class GetGLMHandle(Process):
 global moss_handle
 moss_handle = None
 #################################################################################
-def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=[], console_slience=False):
+def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="",
+                                  observe_window:list=[], console_slience:bool=False):
     """
         多线程方法
         函数的说明请见 request_llms/bridge_all.py

request_llms/bridge_qianfan.py

@@ -117,7 +117,8 @@ def generate_from_baidu_qianfan(inputs, llm_kwargs, history, system_prompt):
                 raise RuntimeError(dec['error_msg'])
 
 
-def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=[], console_slience=False):
+def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="",
+                                  observe_window:list=[], console_slience:bool=False):
     """
         ⭐多线程方法
         函数的说明请见 request_llms/bridge_all.py

request_llms/bridge_qwen.py

@@ -5,7 +5,8 @@ from toolbox import check_packages, report_exception
 
 model_name = 'Qwen'
 
-def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=[], console_slience=False):
+def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="",
+                                  observe_window:list=[], console_slience:bool=False):
     """
         ⭐多线程方法
         函数的说明请见 request_llms/bridge_all.py
@@ -47,6 +48,8 @@ def predict(inputs, llm_kwargs, plugin_kwargs, chatbot, history=[], system_promp
     if additional_fn is not None:
         from core_functional import handle_core_functionality
         inputs, history = handle_core_functionality(additional_fn, inputs, history, chatbot)
+        chatbot[-1] = (inputs, "")
+        yield from update_ui(chatbot=chatbot, history=history)
 
     # 开始接收回复
     from .com_qwenapi import QwenRequestInstance

request_llms/bridge_skylark2.py

@@ -9,7 +9,8 @@ def validate_key():
     if YUNQUE_SECRET_KEY == '': return False
     return True
 
-def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=[], console_slience=False):
+def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="",
+                                  observe_window:list=[], console_slience:bool=False):
     """
         ⭐ 多线程方法
         函数的说明请见 request_llms/bridge_all.py

request_llms/bridge_spark.py

@@ -13,7 +13,8 @@ def validate_key():
         return False
     return True
 
-def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=[], console_slience=False):
+def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="",
+                                  observe_window:list=[], console_slience:bool=False):
     """
         ⭐多线程方法
         函数的说明请见 request_llms/bridge_all.py

request_llms/bridge_zhipu.py

@@ -1,7 +1,8 @@
 import time
 import os
-from toolbox import update_ui, get_conf, update_ui_lastest_msg
+from toolbox import update_ui, get_conf, update_ui_lastest_msg, log_chat
 from toolbox import check_packages, report_exception, have_any_recent_upload_image_files
+from toolbox import ChatBotWithCookies
 
 model_name = '智谱AI大模型'
 zhipuai_default_model = 'glm-4'
@@ -16,7 +17,8 @@ def make_media_input(inputs, image_paths):
         inputs = inputs + f'<br/><br/><div align="center"><img src="file={os.path.abspath(image_path)}"></div>'
     return inputs
 
-def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=[], console_slience=False):
+def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="",
+                                  observe_window:list=[], console_slience:bool=False):
     """
         ⭐多线程方法
         函数的说明请见 request_llms/bridge_all.py
@@ -42,7 +44,8 @@ def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="",
     return response
 
 
-def predict(inputs, llm_kwargs, plugin_kwargs, chatbot, history=[], system_prompt='', stream=True, additional_fn=None):
+def predict(inputs:str, llm_kwargs:dict, plugin_kwargs:dict, chatbot:ChatBotWithCookies,
+            history:list=[], system_prompt:str='', stream:bool=True, additional_fn:str=None):
     """
         ⭐单线程方法
         函数的说明请见 request_llms/bridge_all.py
@@ -90,4 +93,5 @@ def predict(inputs, llm_kwargs, plugin_kwargs, chatbot, history=[], system_promp
         chatbot[-1] = [inputs, response]
         yield from update_ui(chatbot=chatbot, history=history)
     history.extend([inputs, response])
+    log_chat(llm_model=llm_kwargs["llm_model"], input_str=inputs, output_str=response)
     yield from update_ui(chatbot=chatbot, history=history)

request_llms/local_llm_class.py

@@ -1,6 +1,7 @@
 import time
 import threading
 from toolbox import update_ui, Singleton
+from toolbox import ChatBotWithCookies
 from multiprocessing import Process, Pipe
 from contextlib import redirect_stdout
 from request_llms.queued_pipe import create_queue_pipe
@@ -214,7 +215,7 @@ class LocalLLMHandle(Process):
 def get_local_llm_predict_fns(LLMSingletonClass, model_name, history_format='classic'):
     load_message = f"{model_name}尚未加载，加载需要一段时间。注意，取决于`config.py`的配置，{model_name}消耗大量的内存（CPU）或显存（GPU），也许会导致低配计算机卡死 ……"
 
-    def predict_no_ui_long_connection(inputs, llm_kwargs, history=[], sys_prompt="", observe_window=[], console_slience=False):
+    def predict_no_ui_long_connection(inputs:str, llm_kwargs:dict, history:list=[], sys_prompt:str="", observe_window:list=[], console_slience:bool=False):
         """
             refer to request_llms/bridge_all.py
         """
@@ -260,7 +261,8 @@ def get_local_llm_predict_fns(LLMSingletonClass, model_name, history_format='cla
                     raise RuntimeError("程序终止。")
         return response
 
-    def predict(inputs, llm_kwargs, plugin_kwargs, chatbot, history=[], system_prompt='', stream=True, additional_fn=None):
+    def predict(inputs:str, llm_kwargs:dict, plugin_kwargs:dict, chatbot:ChatBotWithCookies,
+                history:list=[], system_prompt:str='', stream:bool=True, additional_fn:str=None):
         """
             refer to request_llms/bridge_all.py
         """