Upload 114 files

CITATION.cff

@@ -1,5 +1,5 @@
 cff-version: 1.2.0
-title: ChuanhuChatGPT
+title: Chuanhu Chat
 message: >-
   If you use this software, please cite it using these
   metadata.
@@ -13,8 +13,8 @@ authors:
     orcid: https://orcid.org/0009-0005-0357-272X
 repository-code: 'https://github.com/GaiZhenbiao/ChuanhuChatGPT'
 url: 'https://github.com/GaiZhenbiao/ChuanhuChatGPT'
-abstract: Provided a light and easy to use interface for ChatGPT API
+abstract: This software provides a light and easy-to-use interface for ChatGPT API and many LLMs.
 license: GPL-3.0
-commit: bd0034c37e5af6a90bd9c2f7dd073f6cd27c61af
-version: '20230405'
-date-released: '2023-04-05'
+commit: c6c08bc62ef80e37c8be52f65f9b6051a7eea1fa
+version: '20230709'
+date-released: '2023-07-09'

ChuanhuChatbot.py

@@ -1,8 +1,11 @@
 # -*- coding:utf-8 -*-
-import os
 import logging
-import sys
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s [%(levelname)s] [%(filename)s:%(lineno)d] %(message)s",
+)
 
+import colorama
 import gradio as gr
 
 from modules import config
@@ -10,6 +13,9 @@ from modules.config import *
 from modules.utils import *
 from modules.presets import *
 from modules.overwrites import *
+from modules.webui import *
+from modules.repo import *
+from modules.train_func import *
 from modules.models.models import get_model
 
 logging.getLogger("httpx").setLevel(logging.WARNING)
@@ -17,13 +23,13 @@ logging.getLogger("httpx").setLevel(logging.WARNING)
 gr.Chatbot._postprocess_chat_messages = postprocess_chat_messages
 gr.Chatbot.postprocess = postprocess
 
-with open("assets/custom.css", "r", encoding="utf-8") as f:
-    customCSS = f.read()
+# with open("web_assets/css/ChuanhuChat.css", "r", encoding="utf-8") as f:
+#     ChuanhuChatCSS = f.read()
 
 def create_new_model():
     return get_model(model_name = MODELS[DEFAULT_MODEL], access_key = my_api_key)[0]
 
-with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
+with gr.Blocks(theme=small_and_beautiful_theme) as demo:
     user_name = gr.State("")
     promptTemplates = gr.State(load_template(get_template_names(plain=True)[0], mode=2))
     user_question = gr.State("")
@@ -34,31 +40,45 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
     topic = gr.State(i18n("未命名对话历史记录"))
 
     with gr.Row():
-        gr.HTML(CHUANHU_TITLE, elem_id="app_title")
-        status_display = gr.Markdown(get_geoip(), elem_id="status_display")
-    with gr.Row(elem_id="float_display"):
-        user_info = gr.Markdown(value="getting user info...", elem_id="user_info")
-
-    with gr.Row().style(equal_height=True):
+        gr.HTML(CHUANHU_TITLE, elem_id="app-title")
+        status_display = gr.Markdown(get_geoip(), elem_id="status-display")
+    with gr.Row(elem_id="float-display"):
+        user_info = gr.Markdown(value="getting user info...", elem_id="user-info")
+        config_info = gr.HTML(get_html("config_info.html").format(bot_avatar=config.bot_avatar, user_avatar=config.user_avatar), visible=False, elem_id="config-info")
+        update_info = gr.HTML(get_html("update.html").format(
+            current_version=repo_tag_html(),
+            version_time=version_time(),
+            cancel_btn=i18n("取消"),
+            update_btn=i18n("更新"),
+            seenew_btn=i18n("详情"),
+            ok_btn=i18n("好"),
+            ), visible=check_update)
+
+    with gr.Row(equal_height=True):
         with gr.Column(scale=5):
             with gr.Row():
-                chatbot = gr.Chatbot(label="Chuanhu Chat", elem_id="chuanhu_chatbot").style(height="100%")
+                chatbot = gr.Chatbot(label="Chuanhu Chat", elem_id="chuanhu-chatbot", latex_delimiters=latex_delimiters_set, height=700)
             with gr.Row():
                 with gr.Column(min_width=225, scale=12):
                     user_input = gr.Textbox(
-                        elem_id="user_input_tb",
-                        show_label=False, placeholder=i18n("在这里输入")
-                    ).style(container=False)
+                        elem_id="user-input-tb",
+                        show_label=False, placeholder=i18n("在这里输入"),
+                        container=False
+                    )
                 with gr.Column(min_width=42, scale=1):
-                    submitBtn = gr.Button(value="", variant="primary", elem_id="submit_btn")
-                    cancelBtn = gr.Button(value="", variant="secondary", visible=False, elem_id="cancel_btn")
-            with gr.Row():
-                emptyBtn = gr.Button(
-                    i18n("🧹 新的对话"), elem_id="empty_btn"
-                )
-                retryBtn = gr.Button(i18n("🔄 重新生成"))
-                delFirstBtn = gr.Button(i18n("🗑️ 删除最旧对话"))
-                delLastBtn = gr.Button(i18n("🗑️ 删除最新对话"))
+                    submitBtn = gr.Button(value="", variant="primary", elem_id="submit-btn")
+                    cancelBtn = gr.Button(value="", variant="secondary", visible=False, elem_id="cancel-btn")
+            with gr.Row(elem_id="chatbot-buttons"):
+                with gr.Column(min_width=120, scale=1):
+                    emptyBtn = gr.Button(
+                        i18n("🧹 新的对话"), elem_id="empty-btn"
+                    )
+                with gr.Column(min_width=120, scale=1):
+                    retryBtn = gr.Button(i18n("🔄 重新生成"))
+                with gr.Column(min_width=120, scale=1):
+                    delFirstBtn = gr.Button(i18n("🗑️ 删除最旧对话"))
+                with gr.Column(min_width=120, scale=1):
+                    delLastBtn = gr.Button(i18n("🗑️ 删除最新对话"))
                 with gr.Row(visible=False) as like_dislike_area:
                     with gr.Column(min_width=20, scale=1):
                         likeBtn = gr.Button(i18n("👍"))
@@ -77,9 +97,9 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
                         label="API-Key",
                     )
                     if multi_api_key:
-                        usageTxt = gr.Markdown(i18n("多账号模式已开启，无需输入key，可直接开始对话"), elem_id="usage_display", elem_classes="insert_block")
+                        usageTxt = gr.Markdown(i18n("多账号模式已开启，无需输入key，可直接开始对话"), elem_id="usage-display", elem_classes="insert-block", visible=show_api_billing)
                     else:
-                        usageTxt = gr.Markdown(i18n("**发送消息** 或 **提交key** 以显示额度"), elem_id="usage_display", elem_classes="insert_block")
+                        usageTxt = gr.Markdown(i18n("**发送消息** 或 **提交key** 以显示额度"), elem_id="usage-display", elem_classes="insert-block", visible=show_api_billing)
                     model_select_dropdown = gr.Dropdown(
                         label=i18n("选择模型"), choices=MODELS, multiselect=False, value=MODELS[DEFAULT_MODEL], interactive=True
                     )
@@ -87,15 +107,15 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
                         label=i18n("选择LoRA模型"), choices=[], multiselect=False, interactive=True, visible=False
                     )
                     with gr.Row():
-                        single_turn_checkbox = gr.Checkbox(label=i18n("单轮对话"), value=False)
-                        use_websearch_checkbox = gr.Checkbox(label=i18n("使用在线搜索"), value=False)
+                        single_turn_checkbox = gr.Checkbox(label=i18n("单轮对话"), value=False, elem_classes="switch-checkbox")
+                        use_websearch_checkbox = gr.Checkbox(label=i18n("使用在线搜索"), value=False, elem_classes="switch-checkbox")
                     language_select_dropdown = gr.Dropdown(
                         label=i18n("选择回复语言（针对搜索&索引功能）"),
                         choices=REPLY_LANGUAGES,
                         multiselect=False,
                         value=REPLY_LANGUAGES[0],
                     )
-                    index_files = gr.Files(label=i18n("上传"), type="file")
+                    index_files = gr.Files(label=i18n("上传"), type="file", elem_id="upload-index-file")
                     two_column = gr.Checkbox(label=i18n("双栏pdf"), value=advance_docs["pdf"].get("two_column", False))
                     summarize_btn = gr.Button(i18n("总结"))
                     # TODO: 公式ocr
@@ -107,8 +127,8 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
                         placeholder=i18n("在这里输入System Prompt..."),
                         label="System prompt",
                         value=INITIAL_SYSTEM_PROMPT,
-                        lines=10,
-                    ).style(container=False)
+                        lines=10
+                    )
                     with gr.Accordion(label=i18n("加载Prompt模板"), open=True):
                         with gr.Column():
                             with gr.Row():
@@ -118,7 +138,8 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
                                         choices=get_template_names(plain=True),
                                         multiselect=False,
                                         value=get_template_names(plain=True)[0],
-                                    ).style(container=False)
+                                        container=False,
+                                    )
                                 with gr.Column(scale=1):
                                     templateRefreshBtn = gr.Button(i18n("🔄 刷新"))
                             with gr.Row():
@@ -129,7 +150,8 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
                                             get_template_names(plain=True)[0], mode=1
                                         ),
                                         multiselect=False,
-                                    ).style(container=False)
+                                        container=False,
+                                    )
 
                 with gr.Tab(label=i18n("保存/加载")):
                     with gr.Accordion(label=i18n("保存/加载对话历史记录"), open=True):
@@ -139,10 +161,14 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
                                     historyFileSelectDropdown = gr.Dropdown(
                                         label=i18n("从列表中加载对话"),
                                         choices=get_history_names(plain=True),
-                                        multiselect=False
+                                        multiselect=False,
+                                        container=False,
                                     )
-                                with gr.Column(scale=1):
-                                    historyRefreshBtn = gr.Button(i18n("🔄 刷新"))
+                                with gr.Row():
+                                    with gr.Column(min_width=42, scale=1):
+                                        historyRefreshBtn = gr.Button(i18n("🔄 刷新"))
+                                    with gr.Column(min_width=42, scale=1):
+                                        historyDeleteBtn = gr.Button(i18n("🗑️ 删除"))
                             with gr.Row():
                                 with gr.Column(scale=6):
                                     saveFileName = gr.Textbox(
@@ -150,7 +176,9 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
                                         placeholder=i18n("设置文件名: 默认为.json，可选为.md"),
                                         label=i18n("设置保存文件名"),
                                         value=i18n("对话历史记录"),
-                                    ).style(container=True)
+                                        elem_classes="no-container"
+                                        # container=False,
+                                    )
                                 with gr.Column(scale=1):
                                     saveHistoryBtn = gr.Button(i18n("💾 保存对话"))
                                     exportMarkdownBtn = gr.Button(i18n("📝 导出为Markdown"))
@@ -159,12 +187,32 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
                                 with gr.Column():
                                     downloadFile = gr.File(interactive=True)
 
+                with gr.Tab(label=i18n("微调")):
+                    openai_train_status = gr.Markdown(label=i18n("训练状态"), value=i18n("在这里[查看使用介绍](https://github.com/GaiZhenbiao/ChuanhuChatGPT/wiki/%E4%BD%BF%E7%94%A8%E6%95%99%E7%A8%8B#%E5%BE%AE%E8%B0%83-gpt-35)"))
+
+                    with gr.Tab(label=i18n("准备数据集")):
+                        dataset_preview_json = gr.JSON(label=i18n("数据集预览"), readonly=True)
+                        dataset_selection = gr.Files(label = i18n("选择数据集"), file_types=[".xlsx", ".jsonl"], file_count="single")
+                        upload_to_openai_btn = gr.Button(i18n("上传到OpenAI"), variant="primary", interactive=False)
+
+                    with gr.Tab(label=i18n("训练")):
+                        openai_ft_file_id = gr.Textbox(label=i18n("文件ID"), value="", lines=1, placeholder=i18n("上传到 OpenAI 后自动填充"))
+                        openai_ft_suffix = gr.Textbox(label=i18n("模型名称后缀"), value="", lines=1, placeholder=i18n("可选，用于区分不同的模型"))
+                        openai_train_epoch_slider = gr.Slider(label=i18n("训练轮数（Epochs）"), minimum=1, maximum=100, value=3, step=1, interactive=True)
+                        openai_start_train_btn = gr.Button(i18n("开始训练"), variant="primary", interactive=False)
+
+                    with gr.Tab(label=i18n("状态")):
+                        openai_status_refresh_btn = gr.Button(i18n("刷新状态"))
+                        openai_cancel_all_jobs_btn = gr.Button(i18n("取消所有任务"))
+                        add_to_models_btn = gr.Button(i18n("添加训练好的模型到模型列表"), interactive=False)
+
                 with gr.Tab(label=i18n("高级")):
-                    gr.Markdown(i18n("# ⚠️ 务必谨慎更改 ⚠️\n\n如果无法使用请恢复默认设置"))
-                    gr.HTML(get_html("appearance_switcher.html").format(label=i18n("切换亮暗色主题")), elem_classes="insert_block")
+                    gr.HTML(get_html("appearance_switcher.html").format(label=i18n("切换亮暗色主题")), elem_classes="insert-block")
                     use_streaming_checkbox = gr.Checkbox(
-                            label=i18n("实时传输回答"), value=True, visible=ENABLE_STREAMING_OPTION
+                            label=i18n("实时传输回答"), value=True, visible=ENABLE_STREAMING_OPTION, elem_classes="switch-checkbox"
                         )
+                    checkUpdateBtn = gr.Button(i18n("🔄 检查更新..."), visible=check_update)
+                    gr.Markdown(i18n("# ⚠️ 务必谨慎更改 ⚠️"), elem_id="advanced-warning")
                     with gr.Accordion(i18n("参数"), open=False):
                         temperature_slider = gr.Slider(
                             minimum=-0,
@@ -192,7 +240,7 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
                         )
                         stop_sequence_txt = gr.Textbox(
                             show_label=True,
-                            placeholder=i18n("在这里输入停止符，用英文逗号隔开..."),
+                            placeholder=i18n("停止符，用英文逗号隔开..."),
                             label="stop",
                             value="",
                             lines=1,
@@ -244,25 +292,36 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
                             lines=1,
                         )
 
-                    with gr.Accordion(i18n("网络设置"), open=False, visible=False):
+                    with gr.Accordion(i18n("网络参数"), open=False):
+                        gr.Markdown(i18n("---\n⚠️ 为保证API-Key安全，请在配置文件`config.json`中修改网络设置"), elem_id="netsetting-warning")
+                        default_btn = gr.Button(i18n("🔙 恢复默认网络设置"))
+                        # 网络代理
+                        proxyTxt = gr.Textbox(
+                            show_label=True,
+                            placeholder=i18n("未设置代理..."),
+                            label=i18n("代理地址"),
+                            value=config.http_proxy,
+                            lines=1,
+                            interactive=False,
+                            # container=False,
+                            elem_classes="view-only-textbox no-container",
+                        )
+                        # changeProxyBtn = gr.Button(i18n("🔄 设置代理地址"))
+
                         # 优先展示自定义的api_host
                         apihostTxt = gr.Textbox(
                             show_label=True,
-                            placeholder=i18n("在这里输入API-Host..."),
-                            label="API-Host",
+                            placeholder="api.openai.com",
+                            label="OpenAI API-Host",
                             value=config.api_host or shared.API_HOST,
                             lines=1,
+                            interactive=False,
+                            # container=False,
+                            elem_classes="view-only-textbox no-container",
                         )
-                        changeAPIURLBtn = gr.Button(i18n("🔄 切换API地址"))
-                        proxyTxt = gr.Textbox(
-                            show_label=True,
-                            placeholder=i18n("在这里输入代理地址..."),
-                            label=i18n("代理地址（示例：http://127.0.0.1:10809）"),
-                            value="",
-                            lines=2,
-                        )
-                        changeProxyBtn = gr.Button(i18n("🔄 设置代理地址"))
-                        default_btn = gr.Button(i18n("🔙 恢复默认设置"))
+                        # changeAPIURLBtn = gr.Button(i18n("🔄 切换API地址"))
+                        updateChuanhuBtn = gr.Button(visible=False, elem_classes="invisible-btn", elem_id="update-chuanhu-btn")
+
 
     gr.Markdown(CHUANHU_DESCRIPTION, elem_id="description")
     gr.HTML(get_html("footer.html").format(versions=versions_html()), elem_id="footer")
@@ -323,6 +382,10 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
         outputs=[saveFileName, systemPromptTxt, chatbot]
     )
 
+    refresh_history_args = dict(
+        fn=get_history_names, inputs=[gr.State(False), user_name], outputs=[historyFileSelectDropdown]
+    )
+
 
     # Chatbot
     cancelBtn.click(interrupt, [current_model], [])
@@ -341,6 +404,7 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
         inputs=[current_model],
         outputs=[chatbot, status_display],
         show_progress=True,
+        _js='clearChatbot',
     )
 
     retryBtn.click(**start_outputing_args).then(
@@ -391,7 +455,7 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
     keyTxt.change(set_key, [current_model, keyTxt], [user_api_key, status_display], api_name="set_key").then(**get_usage_args)
     keyTxt.submit(**get_usage_args)
     single_turn_checkbox.change(set_single_turn, [current_model, single_turn_checkbox], None)
-    model_select_dropdown.change(get_model, [model_select_dropdown, lora_select_dropdown, user_api_key, temperature_slider, top_p_slider, systemPromptTxt, user_name], [current_model, status_display, chatbot, lora_select_dropdown], show_progress=True, api_name="get_model")
+    model_select_dropdown.change(get_model, [model_select_dropdown, lora_select_dropdown, user_api_key, temperature_slider, top_p_slider, systemPromptTxt, user_name], [current_model, status_display, chatbot, lora_select_dropdown, user_api_key, keyTxt], show_progress=True, api_name="get_model")
     model_select_dropdown.change(toggle_like_btn_visibility, [model_select_dropdown], [like_dislike_area], show_progress=False)
     lora_select_dropdown.change(get_model, [model_select_dropdown, lora_select_dropdown, user_api_key, temperature_slider, top_p_slider, systemPromptTxt, user_name], [current_model, status_display, chatbot], show_progress=True)
 
@@ -425,10 +489,23 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
         downloadFile,
         show_progress=True,
     )
-    historyRefreshBtn.click(get_history_names, [gr.State(False), user_name], [historyFileSelectDropdown])
+    historyRefreshBtn.click(**refresh_history_args)
+    historyDeleteBtn.click(delete_chat_history, [current_model, historyFileSelectDropdown, user_name], [status_display, historyFileSelectDropdown, chatbot], _js='(a,b,c)=>{return showConfirmationDialog(a, b, c);}')
     historyFileSelectDropdown.change(**load_history_from_file_args)
     downloadFile.change(upload_chat_history, [current_model, downloadFile, user_name], [saveFileName, systemPromptTxt, chatbot])
 
+    # Train
+    dataset_selection.upload(handle_dataset_selection, dataset_selection, [dataset_preview_json, upload_to_openai_btn, openai_train_status])
+    dataset_selection.clear(handle_dataset_clear, [], [dataset_preview_json, upload_to_openai_btn])
+    upload_to_openai_btn.click(upload_to_openai, [dataset_selection], [openai_ft_file_id, openai_train_status], show_progress=True)
+
+    openai_ft_file_id.change(lambda x: gr.update(interactive=True) if len(x) > 0 else gr.update(interactive=False), [openai_ft_file_id], [openai_start_train_btn])
+    openai_start_train_btn.click(start_training, [openai_ft_file_id, openai_ft_suffix, openai_train_epoch_slider], [openai_train_status])
+
+    openai_status_refresh_btn.click(get_training_status, [], [openai_train_status, add_to_models_btn])
+    add_to_models_btn.click(add_to_models, [], [model_select_dropdown, openai_train_status], show_progress=True)
+    openai_cancel_all_jobs_btn.click(cancel_all_jobs, [], [openai_train_status], show_progress=True)
+
     # Advanced
     max_context_length_slider.change(set_token_upper_limit, [current_model, max_context_length_slider], None)
     temperature_slider.change(set_temperature, [current_model, temperature_slider], None)
@@ -444,15 +521,24 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
     default_btn.click(
         reset_default, [], [apihostTxt, proxyTxt, status_display], show_progress=True
     )
-    changeAPIURLBtn.click(
-        change_api_host,
-        [apihostTxt],
-        [status_display],
-        show_progress=True,
-    )
-    changeProxyBtn.click(
-        change_proxy,
-        [proxyTxt],
+    # changeAPIURLBtn.click(
+    #     change_api_host,
+    #     [apihostTxt],
+    #     [status_display],
+    #     show_progress=True,
+    # )
+    # changeProxyBtn.click(
+    #     change_proxy,
+    #     [proxyTxt],
+    #     [status_display],
+    #     show_progress=True,
+    # )
+    checkUpdateBtn.click(fn=None, _js='manualCheckUpdate')
+
+    # Invisible elements
+    updateChuanhuBtn.click(
+        update_chuanhu,
+        [],
         [status_display],
         show_progress=True,
     )
@@ -469,5 +555,5 @@ if __name__ == "__main__":
     reload_javascript()
     demo.queue(concurrency_count=CONCURRENT_COUNT).launch(
         blocked_paths=["config.json"],
-        favicon_path="./assets/favicon.ico"
+        favicon_path="./web_assets/favicon.ico",
     )

Dockerfile

@@ -1,15 +1,18 @@
-FROM python:3.9 as builder
-RUN apt-get update && apt-get install -y build-essential
+FROM python:3.9-slim-buster as builder
+RUN apt-get update \
+    && apt-get install -y build-essential \
+    && apt-get clean \
+    && rm -rf /var/lib/apt/lists/*
 COPY requirements.txt .
 COPY requirements_advanced.txt .
-RUN pip install --user -r requirements.txt
-# RUN pip install --user -r requirements_advanced.txt
+RUN pip install --user --no-cache-dir -r requirements.txt
+# RUN pip install --user --no-cache-dir -r requirements_advanced.txt
 
-FROM python:3.9
-MAINTAINER iskoldt
+FROM python:3.9-slim-buster
+LABEL maintainer="iskoldt"
 COPY --from=builder /root/.local /root/.local
 ENV PATH=/root/.local/bin:$PATH
 COPY . /app
 WORKDIR /app
-ENV dockerrun yes
-CMD ["python3", "-u", "ChuanhuChatbot.py", "2>&1", "|", "tee", "/var/log/application.log"]
+ENV dockerrun=yes
+CMD ["python3", "-u", "ChuanhuChatbot.py","2>&1", "|", "tee", "/var/log/application.log"]

config_example.json

@@ -1,24 +1,60 @@
 {
-    // 你的OpenAI API Key，一般必填，
-    // 若缺省填为 "openai_api_key": "" 则必须再在图形界面中填入API Key
-    "openai_api_key": "",
-    // 你的xmchat API Key，与OpenAI API Key不同
-    "xmchat_api_key": "",
-    "language": "auto",
-    // 如果使用代理，请取消注释下面的两行，并替换代理URL
-    // "https_proxy": "http://127.0.0.1:1079",
-    // "http_proxy": "http://127.0.0.1:1079",
+    // 各配置具体说明，见 [https://github.com/GaiZhenbiao/ChuanhuChatGPT/wiki/使用教程#配置-configjson]
+
+    //== API 配置 ==
+    "openai_api_key": "", // 你的 OpenAI API Key，一般必填，若空缺则需在图形界面中填入API Key
+    "google_palm_api_key": "", // 你的 Google PaLM API Key，用于 Google PaLM 对话模型
+    "xmchat_api_key": "", // 你的 xmchat API Key，用于 XMChat 对话模型
+    "minimax_api_key": "", // 你的 MiniMax API Key，用于 MiniMax 对话模型
+    "minimax_group_id": "", // 你的 MiniMax Group ID，用于 MiniMax 对话模型
+    "midjourney_proxy_api_base": "https://xxx/mj", // 你的 https://github.com/novicezk/midjourney-proxy 代理地址
+    "midjourney_proxy_api_secret": "", // 你的 MidJourney Proxy API Secret，用于鉴权访问 api，可选
+    "midjourney_discord_proxy_url": "", // 你的 MidJourney Discord Proxy URL，用于对生成对图进行反代，可选
+    "midjourney_temp_folder": "./tmp", // 你的 MidJourney 临时文件夹，用于存放生成的图片，填空则关闭自动下载切图（直接显示MJ的四宫格图）
+
+
+    //== Azure ==
+    "openai_api_type": "openai", // 可选项：azure, openai
+    "azure_openai_api_key": "", // 你的 Azure OpenAI API Key，用于 Azure OpenAI 对话模型
+    "azure_openai_api_base_url": "", // 你的 Azure Base URL
+    "azure_openai_api_version": "2023-05-15", // 你的 Azure OpenAI API 版本
+    "azure_deployment_name": "", // 你的 Azure OpenAI Chat 模型 Deployment 名称
+    "azure_embedding_deployment_name": "", // 你的 Azure OpenAI Embedding 模型 Deployment 名称
+    "azure_embedding_model_name": "text-embedding-ada-002", // 你的 Azure OpenAI Embedding 模型名称
+
+    //== 基础配置 ==
+    "language": "auto", // 界面语言，可选"auto", "zh-CN", "en-US", "ja-JP", "ko-KR", "sv-SE"
     "users": [], // 用户列表，[[用户名1, 密码1], [用户名2, 密码2], ...]
     "local_embedding": false, //是否在本地编制索引
+    "hide_history_when_not_logged_in": false, //未登录情况下是否不展示对话历史
+    "check_update": true, //是否启用检查更新
     "default_model": "gpt-3.5-turbo", // 默认模型
+    "bot_avatar": "default", // 机器人头像，可填写图片链接、Data URL (base64)，或者"none"（不显示头像）
+    "user_avatar": "default", // 用户头像，可填写图片链接、Data URL (base64)，或者"none"（不显示头像）
+
+    //== API 用量 ==
+    "show_api_billing": false, //是否显示OpenAI API用量（启用需要填写sensitive_id）
+    "sensitive_id": "", // 你 OpenAI 账户的 Sensitive ID，用于查询 API 用量
+    "usage_limit": 120, // 该 OpenAI API Key 的当月限额，单位：美元，用于计算百分比和显示上限
+    "legacy_api_usage": false, // 是否使用旧版 API 用量查询接口（OpenAI现已关闭该接口，但是如果你在使用第三方 API，第三方可能仍然支持此接口）
+
+    //== 川虎助理设置 ==
+    "default_chuanhu_assistant_model": "gpt-4", //川虎助理使用的模型，可选gpt-3.5-turbo或者gpt-4等
+    "GOOGLE_CSE_ID": "", //谷歌搜索引擎ID，用于川虎助理Pro模式，获取方式请看 https://stackoverflow.com/questions/37083058/programmatically-searching-google-in-python-using-custom-search
+    "GOOGLE_API_KEY": "", //谷歌API Key，用于川虎助理Pro模式
+    "WOLFRAM_ALPHA_APPID": "", //Wolfram Alpha API Key，用于川虎助理Pro模式，获取方式请看 https://products.wolframalpha.com/api/
+    "SERPAPI_API_KEY": "", //SerpAPI API Key，用于川虎助理Pro模式，获取方式请看 https://serpapi.com/
+
+    //== 文档处理与显示 ==
+    "latex_option": "default", // LaTeX 公式渲染策略，可选"default", "strict", "all"或者"disabled"
     "advance_docs": {
         "pdf": {
-            // 是否认为PDF是双栏的
-            "two_column": false,
-            // 是否使用OCR识别PDF中的公式
-            "formula_ocr": true
+            "two_column": false, // 是否认为PDF是双栏的
+            "formula_ocr": true // 是否使用OCR识别PDF中的公式
         }
     },
+
+    //== 高级配置 ==
     // 是否多个API Key轮换使用
     "multi_api_key": false,
     "api_key_list": [
@@ -26,7 +62,12 @@
         "sk-xxxxxxxxxxxxxxxxxxxxxxxx2",
         "sk-xxxxxxxxxxxxxxxxxxxxxxxx3"
     ],
-    // 如果使用自定义端口、自定义ip，请取消注释并替换对应内容
+    // 自定义OpenAI API Base
+    // "openai_api_base": "https://api.openai.com",
+    // 自定��使用代理（请替换代理URL）
+    // "https_proxy": "http://127.0.0.1:1079",
+    // "http_proxy": "http://127.0.0.1:1079",
+    // 自定义端口、自定义ip（请替换对应内容）
     // "server_name": "0.0.0.0",
     // "server_port": 7860,
     // 如果要share到gradio，设置为true

locale/en_US.json

@@ -32,24 +32,33 @@
     "📝 导出为Markdown": "📝 Export as Markdown",
     "默认保存于history文件夹": "Default save in history folder",
     "高级": "Advanced",
-    "# ⚠️ 务必谨慎更改 ⚠️\n\n如果无法使用请恢复默认设置": "# ⚠️ Caution: Changes require care. ⚠️\n\nIf unable to use, restore default settings.",
+    "# ⚠️ 务必谨慎更改 ⚠️": "# ⚠️ Caution: Changes require care. ⚠️",
     "参数": "Parameters",
-    "在这里输入停止符，用英文逗号隔开...": "Type in stop token here, separated by comma...",
+    "停止符，用英文逗号隔开...": "Type in stop token here, separated by comma...",
     "用于定位滥用行为": "Used to locate abuse",
     "用户名": "Username",
-    "网络设置": "Network Settings",
     "在这里输入API-Host...": "Type in API-Host here...",
     "🔄 切换API地址": "🔄 Switch API Address",
-    "在这里输入代理地址...": "Type in proxy address here...",
-    "代理地址（示例：http://127.0.0.1:10809）": "Proxy address (example: http://127.0.0.1:10809）",
+    "未设置代理...": "No proxy...",
+    "代理地址": "Proxy address",
     "🔄 设置代理地址": "🔄 Set Proxy Address",
-    "🔙 恢复默认设置": "🔙 Restore Default Settings",
+    "🔙 恢复默认网络设置": "🔙 Reset Network Settings",
+    "🔄 检查更新...": "🔄 Check for Update...",
+    "取消": "Cancel",
+    "更新": "Update",
+    "详情": "Details",
+    "好": "OK",
+    "更新成功，请重启本程序": "Updated successfully, please restart this program",
+    "更新失败，请尝试[手动更新](https://github.com/GaiZhenbiao/ChuanhuChatGPT/wiki/使用教程#手动更新)": "Update failed, please try [manually updating](https://github.com/GaiZhenbiao/ChuanhuChatGPT/wiki/使用教程#手动更新)",
     "川虎Chat 🚀": "Chuanhu Chat 🚀",
     "开始实时传输回答……": "Start streaming output...",
     "Token 计数: ": "Token Count: ",
-    "，本次对话累计消耗了 ": "，Total cost for this dialogue is ",
+    "，本次对话累计消耗了 ": ", Total cost for this dialogue is ",
     "**获取API使用情况失败**": "**Failed to get API usage**",
+    "**获取API使用情况失败**，需在填写`config.json`中正确填写sensitive_id": "**Failed to get API usage**, correct sensitive_id needed in `config.json`",
+    "**获取API使用情况失败**，sensitive_id错误或已过期": "**Failed to get API usage**, wrong or expired sensitive_id",
     "**本月使用金额** ": "**Monthly usage** ",
+    "本月使用金额": "Monthly usage",
     "获取API使用情况失败:": "Failed to get API usage:",
     "API密钥更改为了": "The API key is changed to",
     "JSON解析错误,收到的内容: ": "JSON parsing error, received content: ",
@@ -64,10 +73,15 @@
     "API key为空，请检查是否输入正确。": "API key is empty, check whether it is entered correctly.",
     "请输入对话内容。": "Enter the content of the conversation.",
     "账单信息不适用": "Billing information is not applicable",
-    "由Bilibili [土川虎虎虎](https://space.bilibili.com/29125536) 和 [明昭MZhao](https://space.bilibili.com/24807452)开发<br />访问川虎Chat的 [GitHub项目](https://github.com/GaiZhenbiao/ChuanhuChatGPT) 下载最新版脚本": "developor: Bilibili [土川虎虎虎](https://space.bilibili.com/29125536) and [明昭MZhao](https://space.bilibili.com/24807452)\n\nDownload latest code from [GitHub](https://github.com/GaiZhenbiao/ChuanhuChatGPT)",
+    "由Bilibili [土川虎虎虎](https://space.bilibili.com/29125536)、[明昭MZhao](https://space.bilibili.com/24807452) 和 [Keldos](https://github.com/Keldos-Li) 开发<br />访问川虎Chat的 [GitHub项目](https://github.com/GaiZhenbiao/ChuanhuChatGPT) 下载最新版脚本": "Developed by Bilibili [土川虎虎虎](https://space.bilibili.com/29125536), [明昭MZhao](https://space.bilibili.com/24807452) and [Keldos](https://github.com/Keldos-Li)\n\nDownload latest code from [GitHub](https://github.com/GaiZhenbiao/ChuanhuChatGPT)",
     "切换亮暗色主题": "Switch light/dark theme",
     "您的IP区域：未知。": "Your IP region: Unknown.",
     "获取IP地理位置失败。原因：": "Failed to get IP location. Reason: ",
     "。你仍然可以使用聊天功能。": ". You can still use the chat function.",
-    "您的IP区域：": "Your IP region: "
+    "您的IP区域：": "Your IP region: ",
+    "总结": "Summarize",
+    "生成内容总结中……": "Generating content summary...",
+    "由于下面的原因，Google 拒绝返回 PaLM 的回答：\n\n": "Due to the following reasons, Google refuses to provide an answer to PaLM: \n\n",
+    "---\n⚠️ 为保证API-Key安全，请在配置文件`config.json`中修改网络设置": "---\n⚠️ To ensure the security of API-Key, please modify the network settings in the configuration file `config.json`.",
+    "网络参数": "Network parameter"
 }

locale/ja_JP.json

@@ -32,24 +32,33 @@
     "📝 导出为Markdown": "📝 Markdownでエクスポート",
     "默认保存于history文件夹": "デフォルトでhistoryフォルダに保存されます",
     "高级": "Advanced",
-    "# ⚠️ 务必谨慎更改 ⚠️\n\n如果无法使用请恢复默认设置": "# ⚠️ 変更には慎重に ⚠️\n\nもし動作しない場合は、デフォルト設定に戻してください。",
+    "# ⚠️ 务必谨慎更改 ⚠️": "# ⚠️ 変更には慎重に ⚠️",
     "参数": "パラメータ",
-    "在这里输入停止符，用英文逗号隔开...": "ここにストップ文字を英語のカンマで区切って入力してください...",
+    "停止符，用英文逗号隔开...": "ここにストップ文字を英語のカンマで区切って入力してください...",
     "用于定位滥用行为": "不正行為を特定するために使用されます",
     "用户名": "ユーザー名",
-    "网络设置": "ネットワーク設定",
     "在这里输入API-Host...": "API-Hostを入力してください...",
     "🔄 切换API地址": "🔄 APIアドレスを切り替え",
-    "在这里输入代理地址...": "プロキシアドレスを入力してください...",
-    "代理地址（示例：http://127.0.0.1:10809）": "プロキシアドレス（例：http://127.0.0.1:10809）",
+    "未设置代理...": "代理が設定されていません...",
+    "代理地址": "プロキシアドレス",
     "🔄 设置代理地址": "🔄 プロキシアドレスを設定",
-    "🔙 恢复默认设置": "🔙 デフォルト設定に戻す",
+    "🔙 恢复默认网络设置": "🔙 ネットワーク設定のリセット",
+    "🔄 检查更新...": "🔄 アップデートをチェック...",
+    "取消": "キャンセル",
+    "更新": "アップデート",
+    "详情": "詳細",
+    "好": "はい",
+    "更新成功，请重启本程序": "更新が成功しました、このプログラムを再起動してください",
+    "更新失败，请尝试[手动更新](https://github.com/GaiZhenbiao/ChuanhuChatGPT/wiki/使用教程#手动更新)": "更新に失敗しました、[手動での更新](https://github.com/GaiZhenbiao/ChuanhuChatGPT/wiki/使用教程#手动更新)をお試しください。",
     "川虎Chat 🚀": "川虎Chat 🚀",
     "开始实时传输回答……": "ストリーム出力開始……",
     "Token 计数: ": "Token数: ",
     "，本次对话累计消耗了 ": ", 今の会話で消費合計 ",
     "**获取API使用情况失败**": "**API使用状況の取得に失敗しました**",
+    "**获取API使用情况失败**，需在填写`config.json`中正确填写sensitive_id": "**API使用状況の取得に失敗しました**、`config.json`に正しい`sensitive_id`を入力する必要があります",
+    "**获取API使用情况失败**，sensitive_id错误或已过期": "**API使用状況の取得に失敗しました**、sensitive_idが間違っているか、期限切れです",
     "**本月使用金额** ": "**今月の使用料金** ",
+    "本月使用金额": "今月の使用料金",
     "获取API使用情况失败:": "API使用状況の取得に失敗しました:",
     "API密钥更改为了": "APIキーが変更されました",
     "JSON解析错误,收到的内容: ": "JSON解析エラー、受信内容: ",
@@ -64,10 +73,15 @@
     "API key为空，请检查是否输入正确。": "APIキーが入力されていません。正しく入力されているか確認してください。",
     "请输入对话内容。": "会話内容を入力してください。",
     "账单信息不适用": "課金情報は対象外です",
-    "由Bilibili [土川虎虎虎](https://space.bilibili.com/29125536) 和 [明昭MZhao](https://space.bilibili.com/24807452)开发<br />访问川虎Chat的 [GitHub项目](https://github.com/GaiZhenbiao/ChuanhuChatGPT) 下载最新版脚本": "開発：Bilibili [土川虎虎虎](https://space.bilibili.com/29125536) と [明昭MZhao](https://space.bilibili.com/24807452)\n\n最新コードは川虎Chatのサイトへ [GitHubプロジェクト](https://github.com/GaiZhenbiao/ChuanhuChatGPT)",
+    "由Bilibili [土川虎虎虎](https://space.bilibili.com/29125536)、[明昭MZhao](https://space.bilibili.com/24807452) 和 [Keldos](https://github.com/Keldos-Li) 开发<br />访问川虎Chat的 [GitHub项目](https://github.com/GaiZhenbiao/ChuanhuChatGPT) 下载最新版脚本": "開発：Bilibili [土川虎虎虎](https://space.bilibili.com/29125536) と [明昭MZhao](https://space.bilibili.com/24807452) と [Keldos](https://github.com/Keldos-Li)\n\n最新コードは川虎Chatのサイトへ [GitHubプロジェクト](https://github.com/GaiZhenbiao/ChuanhuChatGPT)",
     "切换亮暗色主题": "テーマの明暗切替",
     "您的IP区域：未知。": "あなたのIPアドレス地域：不明",
     "获取IP地理位置失败。原因：": "IPアドレス地域の取得に失敗しました。理由：",
     "。你仍然可以使用聊天功能。": "。あなたはまだチャット機能を使用できます。",
-    "您的IP区域：": "あなたのIPアドレス地域："
-}
+    "您的IP区域：": "あなたのIPアドレス地域：",
+    "总结": "要約する",
+    "生成内容总结中……": "コンテンツ概要を生成しています...",
+    "由于下面的原因，Google 拒绝返回 PaLM 的回答：\n\n": "Googleは以下の理由から、PaLMの回答を返すことを拒否しています:\n\n",
+    "---\n⚠️ 为保证API-Key安全，请在配置文件`config.json`中修改网络设置": "---\n⚠️ APIキーの安全性を確保するために、`config.json`ファイルでネットワーク設定を変更してください。",
+    "网络参数": "ネットワークパラメータ"
+}

locale/ko_KR.json

@@ -0,0 +1,89 @@
+{
+    "未命名对话历史记录": "이름없는 대화 기록",
+    "在这里输入": "여기에 입력하세요",
+    "🧹 新的对话": "🧹 새로운 대화",
+    "🔄 重新生成": "🔄 재생성",
+    "🗑️ 删除最旧对话": "🗑️ 가장 오래된 대화 삭제",
+    "🗑️ 删除最新对话": "🗑️ 최신 대화 삭제",
+    "🗑️ 删除": "🗑️ 삭제",
+    "模型": "LLM 모델",
+    "多账号模式已开启，无需输入key，可直接开始对话": "다중 계정 모드가 활성화되어 있으므로 키를 입력할 필요가 없이 바로 대화를 시작할 수 있습니다",
+    "**发送消息** 或 **提交key** 以显示额度": "**메세지를 전송** 하거나 **Key를 입력**하여 크레딧 표시",
+    "选择模型": "모델 선택",
+    "选择LoRA模型": "LoRA 모델 선택",
+    "实时传输回答": "실시간 전송",
+    "单轮对话": "단일 대화",
+    "使用在线搜索": "온라인 검색 사용",
+    "选择回复语言（针对搜索&索引功能）": "답장 언어 선택 (검색 & 인덱스용)",
+    "上传索引文件": "업로드",
+    "双栏pdf": "2-column pdf",
+    "识别公式": "formula OCR",
+    "在这里输入System Prompt...": "여기에 시스템 프롬프트를 입력하세요...",
+    "加载Prompt模板": "프롬프트 템플릿 불러오기",
+    "选择Prompt模板集合文件": "프롬프트 콜렉션 파일 선택",
+    "🔄 刷新": "🔄 새로고침",
+    "从Prompt模板中加载": "프롬프트 템플릿에서 불러오기",
+    "保存/加载": "저장/불러오기",
+    "保存/加载对话历史记录": "대화 기록 저장/불러오기",
+    "从列表中加载对话": "리스트에서 대화 불러오기",
+    "设置文件名: 默认为.json，可选为.md": "파일 이름 설정: 기본값: .json, 선택: .md",
+    "设置保存文件名": "저장 파일명 설정",
+    "对话历史记录": "대화 기록",
+    "💾 保存对话": "💾 대화 저장",
+    "📝 导出为Markdown": "📝 마크다운으로 내보내기",
+    "默认保存于history文件夹": "히스토리 폴더에 기본 저장",
+    "高级": "고급",
+    "# ⚠️ 务必谨慎更改 ⚠️": "# ⚠️ 주의: 변경시 주의하세요. ⚠️",
+    "参数": "파라미터들",
+    "停止符，用英文逗号隔开...": "여기에 정지 토큰 입력, ','로 구분됨...",
+    "用于定位滥用行为": "악용 사례 파악에 활용됨",
+    "用户名": "사용자 이름",
+    "在这里输入API-Host...": "여기에 API host를 입력하세요...",
+    "🔄 切换API地址": "🔄 API 주소 변경",
+    "未设置代理...": "대리인이 설정되지 않았습니다...",
+    "代理地址": "프록시 주소",
+    "🔄 设置代理地址": "🔄 프록시 주소 설정",
+    "🔙 恢复默认网络设置": "🔙 네트워크 설정 초기화",
+    "🔄 检查更新...": "🔄 업데이트 확인...",
+    "取消": "취소",
+    "更新": "업데이트",
+    "详情": "상세",
+    "好": "예",
+    "更新成功，请重启本程序": "업데이트 성공, 이 프로그램을 재시작 해주세요",
+    "更新失败，请尝试[手动更新](https://github.com/GaiZhenbiao/ChuanhuChatGPT/wiki/使用教程#手动更新)": "업데이트 실패, [수동 업데이트](https://github.com/GaiZhenbiao/ChuanhuChatGPT/wiki/使用教程#手动更新)를 시도하십시오",
+    "川虎Chat 🚀": "Chuanhu Chat 🚀",
+    "开始实时传输回答……": "실시간 응답 출력 시작...",
+    "Token 计数: ": "토큰 수: ",
+    "，本次对话累计消耗了 ": "，이 대화의 전체 비용은 ",
+    "**获取API使用情况失败**": "**API 사용량 가져오기 실패**",
+    "**获取API使用情况失败**，需在填写`config.json`中正确填写sensitive_id": "**API 사용량 가져오기 실패**. `config.json`에 올바른 `sensitive_id`를 입력해야 합니다",
+    "**获取API使用情况失败**，sensitive_id错误或已过期": "**API 사용량 가져오기 실패**. sensitive_id가 잘못되었거나 만료되었습니다",
+    "**本月使用金额** ": "**이번 달 사용금액** ",
+    "本月使用金额": "이번 달 사용금액",
+    "获取API使用情况失败:": "API 사용량 가져오기 실패:",
+    "API密钥更改为了": "API 키가 변경되었습니다.",
+    "JSON解析错误,收到的内容: ": "JSON 파싱 에러, 응답: ",
+    "模型设置为了：": "설정된 모델: ",
+    "☹️发生了错误：": "☹️에러: ",
+    "获取对话时发生错误，请查看后台日志": "대화를 가져오는 중 에러가 발생했습니다. 백그라운드 로그를 확인하세요",
+    "请检查网络连接，或者API-Key是否有效。": "네트워크 연결 또는 API키가 유효한지 확인하세요",
+    "连接超时，无法获取对话。": "연결 시간 초과, 대화를 가져올 수 없습니다.",
+    "读取超时，无法获取对话。": "읽기 시간 초과, 대화를 가져올 수 없습니다.",
+    "代理错误，无法获取对话。": "프록시 에러, 대화를 가져올 수 없습니다.",
+    "SSL错误，无法获取对话。": "SSL 에러, 대화를 가져올 수 없습니다.",
+    "API key为空，请检查是否输入正确。": "API 키가 비어 있습니다. 올바르게 입력되었는지 확인하십세요.",
+    "请输入对话内容。": "대화 내용을 입력하세요.",
+    "账单信息不适用": "청구 정보를 가져올 수 없습니다",
+    "由Bilibili [土川虎虎虎](https://space.bilibili.com/29125536)、[明昭MZhao](https://space.bilibili.com/24807452) 和 [Keldos](https://github.com/Keldos-Li) 开发<br />访问川虎Chat的 [GitHub项目](https://github.com/GaiZhenbiao/ChuanhuChatGPT) 下载最新版脚本": "제작: Bilibili [土川虎虎虎](https://space.bilibili.com/29125536), [明昭MZhao](https://space.bilibili.com/24807452), [Keldos](https://github.com/Keldos-Li)\n\n최신 코드 다운로드: [GitHub](https://github.com/GaiZhenbiao/ChuanhuChatGPT)",
+    "切换亮暗色主题": "라이트/다크 테마 전환",
+    "您的IP区域：未知。": "IP 지역: 알 수 없음.",
+    "获取IP地理位置失败。原因：": "다음과 같은 이유로 IP 위치를 가져올 수 없습니다. 이유: ",
+    "。你仍然可以使用聊天功能。": ". 채팅 기능을 계속 사용할 수 있습니다.",
+    "您的IP区域：": "당신의 IP 지역: ",
+    "总结": "요약",
+    "生成内容总结中……": "콘텐츠 요약 생성중...",
+    "上传": "업로드",
+    "由于下面的原因，Google 拒绝返回 PaLM 的回答：\n\n": "구글은 다음과 같은 이유로 인해 PaLM의 응답을 거부합니다: \n\n",
+    "---\n⚠️ 为保证API-Key安全，请在配置文件`config.json`中修改网络设置": "---\n⚠️ API-Key의 안전을 보장하기 위해 네트워크 설정을 `config.json` 구성 파일에서 수정해주세요.",
+    "网络参数": "네트워크 매개변수"
+}

locale/sv-SE.json

@@ -0,0 +1,87 @@
+{  
+    "未命名对话历史记录": "Onämnd Dialoghistorik",  
+    "在这里输入": "Skriv in här",  
+    "🧹 新的对话": "🧹 Ny Dialog",  
+    "🔄 重新生成": "🔄 Regenerera",  
+    "🗑️ 删除最旧对话": "🗑️ Ta bort äldsta dialogen",  
+    "🗑️ 删除最新对话": "🗑️ Ta bort senaste dialogen",  
+    "模型": "Modell",  
+    "多账号模式已开启，无需输入key，可直接开始对话": "Flerkontoläge är aktiverat, ingen nyckel behövs, du kan starta dialogen direkt",  
+    "**发送消息** 或 **提交key** 以显示额度": "**Skicka meddelande** eller **Skicka in nyckel** för att visa kredit",  
+    "选择模型": "Välj Modell",  
+    "选择LoRA模型": "Välj LoRA Modell",  
+    "实时传输回答": "Strömmande utdata",  
+    "单轮对话": "Enkel dialog",  
+    "使用在线搜索": "Använd online-sökning",  
+    "选择回复语言（针对搜索&索引功能）": "Välj svarspråk (för sök- och indexfunktion)",  
+    "上传索引文件": "Ladda upp",  
+    "双栏pdf": "Två-kolumns pdf",  
+    "识别公式": "Formel OCR",  
+    "在这里输入System Prompt...": "Skriv in System Prompt här...",  
+    "加载Prompt模板": "Ladda Prompt-mall",  
+    "选择Prompt模板集合文件": "Välj Prompt-mall Samlingsfil",  
+    "🔄 刷新": "🔄 Uppdatera",  
+    "从Prompt模板中加载": "Ladda från Prompt-mall",  
+    "保存/加载": "Spara/Ladda",  
+    "保存/加载对话历史记录": "Spara/Ladda Dialoghistorik",  
+    "从列表中加载对话": "Ladda dialog från lista",  
+    "设置文件名: 默认为.json，可选为.md": "Ställ in filnamn: standard är .json, valfritt är .md",  
+    "设置保存文件名": "Ställ in sparfilnamn",  
+    "对话历史记录": "Dialoghistorik",  
+    "💾 保存对话": "💾 Spara Dialog",  
+    "📝 导出为Markdown": "📝 Exportera som Markdown",  
+    "默认保存于history文件夹": "Sparas som standard i mappen history",  
+    "高级": "Avancerat",  
+    "# ⚠️ 务必谨慎更改 ⚠️": "# ⚠️ Var försiktig med ändringar. ⚠️",  
+    "参数": "Parametrar",  
+    "停止符，用英文逗号隔开...": "Skriv in stopptecken här, separerade med kommatecken...",  
+    "用于定位滥用行为": "Används för att lokalisera missbruk",  
+    "用户名": "Användarnamn",  
+    "在这里输入API-Host...": "Skriv in API-Host här...",  
+    "🔄 切换API地址": "🔄 Byt API-adress",  
+    "未设置代理...": "Inte inställd proxy...",  
+    "代理地址": "Proxyadress",  
+    "🔄 设置代理地址": "🔄 Ställ in Proxyadress",  
+    "🔙 恢复网络默认设置": "🔙 Återställ Nätverksinställningar",  
+    "🔄 检查更新...": "🔄 Sök efter uppdateringar...",  
+    "取消": "Avbryt",  
+    "更新": "Uppdatera",  
+    "详情": "Detaljer",  
+    "好": "OK",  
+    "更新成功，请重启本程序": "Uppdaterat framgångsrikt, starta om programmet",
+    "更新失败，请尝试[手动更新](https://github.com/GaiZhenbiao/ChuanhuChatGPT/wiki/使用教程#手动更新)": "Uppdateringen misslyckades, prova att [uppdatera manuellt](https://github.com/GaiZhenbiao/ChuanhuChatGPT/wiki/使用教程#手动更新)",
+    "川虎Chat 🚀": "Chuanhu Chat 🚀",  
+    "开始实时传输回答……": "Börjar strömma utdata...",  
+    "Token 计数: ": "Tokenräkning: ",  
+    "，本次对话累计消耗了 ": ", Total kostnad för denna dialog är ",  
+    "**获取API使用情况失败**": "**Misslyckades med att hämta API-användning**",  
+    "**获取API使用情况失败**，需在填写`config.json`中正确填写sensitive_id": "**Misslyckades med att hämta API-användning**, korrekt sensitive_id behövs i `config.json`",  
+    "**获取API使用情况失败**，sensitive_id错误或已过期": "**Misslyckades med att hämta API-användning**, felaktig eller utgången sensitive_id",  
+    "**本月使用金额** ": "**Månadens användning** ",  
+    "本月使用金额": "Månadens användning",  
+    "获取API使用情况失败:": "Misslyckades med att hämta API-användning:",  
+    "API密钥更改为了": "API-nyckeln har ändrats till",  
+    "JSON解析错误,收到的内容: ": "JSON-tolkningsfel, mottaget innehåll: ",  
+    "模型设置为了：": "Modellen är inställd på: ",  
+    "☹️发生了错误：": "☹️Fel: ",  
+    "获取对话时发生错误，请查看后台日志": "Ett fel uppstod när dialogen hämtades, kontrollera bakgrundsloggen",  
+    "请检查网络连接，或者API-Key是否有效。": "Kontrollera nätverksanslutningen eller om API-nyckeln är giltig.",  
+    "连接超时，无法获取对话。": "Anslutningen tog för lång tid, kunde inte hämta dialogen.",  
+    "读取超时，无法获取对话。": "Läsningen tog för lång tid, kunde inte hämta dialogen.",  
+    "代理错误，无法获取对话。": "Proxyfel, kunde inte hämta dialogen.",  
+    "SSL错误，无法获取对话。": "SSL-fel, kunde inte hämta dialogen.",  
+    "API key为空，请检查是否输入正确。": "API-nyckeln är tom, kontrollera om den är korrekt inmatad.",  
+    "请输入对话内容。": "Ange dialoginnehåll.",  
+    "账单信息不适用": "Faktureringsinformation är inte tillämplig",  
+    "由Bilibili [土川虎虎虎](https://space.bilibili.com/29125536)、[明昭MZhao](https://space.bilibili.com/24807452) 和 [Keldos](https://github.com/Keldos-Li) 开发<br />访问川虎Chat的 [GitHub项目](https://github.com/GaiZhenbiao/ChuanhuChatGPT) 下载最新版脚本": "Utvecklad av Bilibili [土川虎虎虎](https://space.bilibili.com/29125536), [明昭MZhao](https://space.bilibili.com/24807452) och [Keldos](https://github.com/Keldos-Li)\n\nLadda ner senaste koden från [GitHub](https://github.com/GaiZhenbiao/ChuanhuChatGPT)",  
+    "切换亮暗色主题": "Byt ljus/mörk tema",  
+    "您的IP区域：未知。": "Din IP-region: Okänd.",  
+    "获取IP地理位置失败。原因：": "Misslyckades med att hämta IP-plats. Orsak: ",  
+    "。你仍然可以使用聊天功能。": ". Du kan fortfarande använda chattfunktionen.",  
+    "您的IP区域：": "Din IP-region: ",  
+    "总结": "Sammanfatta",  
+    "生成内容总结中……": "Genererar innehållssammanfattning...",  
+    "由于下面的原因，Google 拒绝返回 PaLM 的回答：\n\n": "På grund av följande skäl vägrar Google att ge ett svar till PaLM: \n\n",
+    "---\n⚠️ 为保证API-Key安全，请在配置文件`config.json`中修改网络设置": "---\n⚠️ För att säkerställa säkerheten för API-nyckeln, vänligen ändra nätverksinställningarna i konfigurationsfilen `config.json`.",
+    "网络参数": "nätverksparametrar"
+}  

modules/config.py

@@ -11,11 +11,11 @@ from . import presets
 
 __all__ = [
     "my_api_key",
+    "sensitive_id",
     "authflag",
     "auth_list",
     "dockerflag",
     "retrieve_proxy",
-    "log_level",
     "advance_docs",
     "update_doc_config",
     "usage_limit",
@@ -23,8 +23,11 @@ __all__ = [
     "server_name",
     "server_port",
     "share",
+    "check_update",
+    "latex_delimiters_set",
     "hide_history_when_not_logged_in",
-    "default_chuanhu_assistant_model"
+    "default_chuanhu_assistant_model",
+    "show_api_billing"
 ]
 
 # 添加一个统一的config文件，避免文件过多造成的疑惑（优先级最低）
@@ -35,10 +38,22 @@ if os.path.exists("config.json"):
 else:
     config = {}
 
+
+def load_config_to_environ(key_list):
+    global config
+    for key in key_list:
+        if key in config:
+            os.environ[key.upper()] = os.environ.get(key.upper(), config[key])
+
+
 lang_config = config.get("language", "auto")
 language = os.environ.get("LANGUAGE", lang_config)
 
-hide_history_when_not_logged_in = config.get("hide_history_when_not_logged_in", False)
+hide_history_when_not_logged_in = config.get(
+    "hide_history_when_not_logged_in", False)
+check_update = config.get("check_update", True)
+show_api_billing = config.get("show_api_billing", False)
+show_api_billing = bool(os.environ.get("SHOW_API_BILLING", show_api_billing))
 
 if os.path.exists("api_key.txt"):
     logging.info("检测到api_key.txt文件，正在进行迁移...")
@@ -52,26 +67,44 @@ if os.path.exists("auth.json"):
     logging.info("检测到auth.json文件，正在进行迁移...")
     auth_list = []
     with open("auth.json", "r", encoding='utf-8') as f:
-            auth = json.load(f)
-            for _ in auth:
-                if auth[_]["username"] and auth[_]["password"]:
-                    auth_list.append((auth[_]["username"], auth[_]["password"]))
-                else:
-                    logging.error("请检查auth.json文件中的用户名和密码！")
-                    sys.exit(1)
+        auth = json.load(f)
+        for _ in auth:
+            if auth[_]["username"] and auth[_]["password"]:
+                auth_list.append((auth[_]["username"], auth[_]["password"]))
+            else:
+                logging.error("请检查auth.json文件中的用户名和密码！")
+                sys.exit(1)
     config["users"] = auth_list
     os.rename("auth.json", "auth(deprecated).json")
     with open("config.json", "w", encoding='utf-8') as f:
         json.dump(config, f, indent=4, ensure_ascii=False)
 
-## 处理docker if we are running in Docker
+# 处理docker if we are running in Docker
 dockerflag = config.get("dockerflag", False)
 if os.environ.get("dockerrun") == "yes":
     dockerflag = True
 
-## 处理 api-key 以及 允许的用户列表
+# 处理 api-key 以及 允许的用户列表
 my_api_key = config.get("openai_api_key", "")
 my_api_key = os.environ.get("OPENAI_API_KEY", my_api_key)
+os.environ["OPENAI_API_KEY"] = my_api_key
+os.environ["OPENAI_EMBEDDING_API_KEY"] = my_api_key
+
+if config.get("legacy_api_usage", False):
+    sensitive_id = my_api_key
+else:
+    sensitive_id = config.get("sensitive_id", "")
+    sensitive_id = os.environ.get("SENSITIVE_ID", sensitive_id)
+
+# 模型配置
+if "extra_models" in  config:
+    presets.MODELS.extend(config["extra_models"])
+    logging.info(f"已添加额外的模型：{config['extra_models']}")
+
+google_palm_api_key = config.get("google_palm_api_key", "")
+google_palm_api_key = os.environ.get(
+    "GOOGLE_PALM_API_KEY", google_palm_api_key)
+os.environ["GOOGLE_PALM_API_KEY"] = google_palm_api_key
 
 xmchat_api_key = config.get("xmchat_api_key", "")
 os.environ["XMCHAT_API_KEY"] = xmchat_api_key
@@ -81,11 +114,23 @@ os.environ["MINIMAX_API_KEY"] = minimax_api_key
 minimax_group_id = config.get("minimax_group_id", "")
 os.environ["MINIMAX_GROUP_ID"] = minimax_group_id
 
+midjourney_proxy_api_base = config.get("midjourney_proxy_api_base", "")
+os.environ["MIDJOURNEY_PROXY_API_BASE"] = midjourney_proxy_api_base
+midjourney_proxy_api_secret = config.get("midjourney_proxy_api_secret", "")
+os.environ["MIDJOURNEY_PROXY_API_SECRET"] = midjourney_proxy_api_secret
+midjourney_discord_proxy_url = config.get("midjourney_discord_proxy_url", "")
+os.environ["MIDJOURNEY_DISCORD_PROXY_URL"] = midjourney_discord_proxy_url
+midjourney_temp_folder = config.get("midjourney_temp_folder", "")
+os.environ["MIDJOURNEY_TEMP_FOLDER"] = midjourney_temp_folder
+
+load_config_to_environ(["openai_api_type", "azure_openai_api_key", "azure_openai_api_base_url",
+                       "azure_openai_api_version", "azure_deployment_name", "azure_embedding_deployment_name", "azure_embedding_model_name"])
+
 
 usage_limit = os.environ.get("USAGE_LIMIT", config.get("usage_limit", 120))
 
-## 多账户机制
-multi_api_key = config.get("multi_api_key", False) # 是否开启多账户机制
+# 多账户机制
+multi_api_key = config.get("multi_api_key", False)  # 是否开启多账户机制
 if multi_api_key:
     api_key_list = config.get("api_key_list", [])
     if len(api_key_list) == 0:
@@ -93,21 +138,26 @@ if multi_api_key:
         sys.exit(1)
     shared.state.set_api_key_queue(api_key_list)
 
-auth_list = config.get("users", []) # 实际上是使用者的列表
+auth_list = config.get("users", [])  # 实际上是使用者的列表
 authflag = len(auth_list) > 0  # 是否开启认证的状态值，改为判断auth_list长度
 
 # 处理自定义的api_host，优先读环境变量的配置，如果存在则自动装配
-api_host = os.environ.get("OPENAI_API_BASE", config.get("openai_api_base", None))
+api_host = os.environ.get(
+    "OPENAI_API_BASE", config.get("openai_api_base", None))
 if api_host is not None:
     shared.state.set_api_host(api_host)
+    os.environ["OPENAI_API_BASE"] = f"{api_host}/v1"
+    logging.info(f"OpenAI API Base set to: {os.environ['OPENAI_API_BASE']}")
 
-default_chuanhu_assistant_model = config.get("default_chuanhu_assistant_model", "gpt-3.5-turbo")
+default_chuanhu_assistant_model = config.get(
+    "default_chuanhu_assistant_model", "gpt-3.5-turbo")
 for x in ["GOOGLE_CSE_ID", "GOOGLE_API_KEY", "WOLFRAM_ALPHA_APPID", "SERPAPI_API_KEY"]:
     if config.get(x, None) is not None:
         os.environ[x] = config[x]
 
+
 @contextmanager
-def retrieve_openai_api(api_key = None):
+def retrieve_openai_api(api_key=None):
     old_api_key = os.environ.get("OPENAI_API_KEY", "")
     if api_key is None:
         os.environ["OPENAI_API_KEY"] = my_api_key
@@ -117,24 +167,20 @@ def retrieve_openai_api(api_key = None):
         yield api_key
     os.environ["OPENAI_API_KEY"] = old_api_key
 
-## 处理log
-log_level = config.get("log_level", "INFO")
-logging.basicConfig(
-    level=log_level,
-    format="%(asctime)s [%(levelname)s] [%(filename)s:%(lineno)d] %(message)s",
-)
 
-## 处理代理：
-http_proxy = config.get("http_proxy", "")
-https_proxy = config.get("https_proxy", "")
-http_proxy = os.environ.get("HTTP_PROXY", http_proxy)
-https_proxy = os.environ.get("HTTPS_PROXY", https_proxy)
+
+# 处理代理：
+http_proxy = os.environ.get("HTTP_PROXY", "")
+https_proxy = os.environ.get("HTTPS_PROXY", "")
+http_proxy = config.get("http_proxy", http_proxy)
+https_proxy = config.get("https_proxy", https_proxy)
 
 # 重置系统变量，在不需要设置的时候不设置环境变量，以免引起全局代理报错
 os.environ["HTTP_PROXY"] = ""
 os.environ["HTTPS_PROXY"] = ""
 
-local_embedding = config.get("local_embedding", False) # 是否使用本地embedding
+local_embedding = config.get("local_embedding", False)  # 是否使用本地embedding
+
 
 @contextmanager
 def retrieve_proxy(proxy=None):
@@ -151,22 +197,62 @@ def retrieve_proxy(proxy=None):
         old_var = os.environ["HTTP_PROXY"], os.environ["HTTPS_PROXY"]
         os.environ["HTTP_PROXY"] = http_proxy
         os.environ["HTTPS_PROXY"] = https_proxy
-        yield http_proxy, https_proxy # return new proxy
+        yield http_proxy, https_proxy  # return new proxy
 
         # return old proxy
         os.environ["HTTP_PROXY"], os.environ["HTTPS_PROXY"] = old_var
 
 
-## 处理advance docs
+# 处理latex options
+user_latex_option = config.get("latex_option", "default")
+if user_latex_option == "default":
+    latex_delimiters_set = [
+        {"left": "$$", "right": "$$", "display": True},
+        {"left": "$", "right": "$", "display": False},
+        {"left": "\\(", "right": "\\)", "display": False},
+        {"left": "\\[", "right": "\\]", "display": True},
+    ]
+elif user_latex_option == "strict":
+    latex_delimiters_set = [
+        {"left": "$$", "right": "$$", "display": True},
+        {"left": "\\(", "right": "\\)", "display": False},
+        {"left": "\\[", "right": "\\]", "display": True},
+    ]
+elif user_latex_option == "all":
+    latex_delimiters_set = [
+        {"left": "$$", "right": "$$", "display": True},
+        {"left": "$", "right": "$", "display": False},
+        {"left": "\\(", "right": "\\)", "display": False},
+        {"left": "\\[", "right": "\\]", "display": True},
+        {"left": "\\begin{equation}", "right": "\\end{equation}", "display": True},
+        {"left": "\\begin{align}", "right": "\\end{align}", "display": True},
+        {"left": "\\begin{alignat}", "right": "\\end{alignat}", "display": True},
+        {"left": "\\begin{gather}", "right": "\\end{gather}", "display": True},
+        {"left": "\\begin{CD}", "right": "\\end{CD}", "display": True},
+    ]
+elif user_latex_option == "disabled":
+    latex_delimiters_set = []
+else:
+    latex_delimiters_set = [
+        {"left": "$$", "right": "$$", "display": True},
+        {"left": "$", "right": "$", "display": False},
+        {"left": "\\(", "right": "\\)", "display": False},
+        {"left": "\\[", "right": "\\]", "display": True},
+    ]
+
+# 处理advance docs
 advance_docs = defaultdict(lambda: defaultdict(dict))
 advance_docs.update(config.get("advance_docs", {}))
+
+
 def update_doc_config(two_column_pdf):
     global advance_docs
     advance_docs["pdf"]["two_column"] = two_column_pdf
 
     logging.info(f"更新后的文件参数为：{advance_docs}")
 
-## 处理gradio.launch参数
+
+# 处理gradio.launch参数
 server_name = config.get("server_name", None)
 server_port = config.get("server_port", None)
 if server_name is None:
@@ -188,3 +274,7 @@ except ValueError:
     pass
 
 share = config.get("share", False)
+
+# avatar
+bot_avatar = config.get("bot_avatar", "default")
+user_avatar = config.get("user_avatar", "default")

modules/index_func.py

@@ -1,7 +1,7 @@
 import os
 import logging
 
-import colorama
+import hashlib
 import PyPDF2
 from tqdm import tqdm
 
@@ -10,19 +10,6 @@ from modules.utils import *
 from modules.config import local_embedding
 
 
-def get_index_name(file_src):
-    file_paths = [x.name for x in file_src]
-    file_paths.sort(key=lambda x: os.path.basename(x))
-
-    md5_hash = hashlib.md5()
-    for file_path in file_paths:
-        with open(file_path, "rb") as f:
-            while chunk := f.read(8192):
-                md5_hash.update(chunk)
-
-    return md5_hash.hexdigest()
-
-
 def get_documents(file_src):
     from langchain.schema import Document
     from langchain.text_splitter import TokenTextSplitter
@@ -47,11 +34,12 @@ def get_documents(file_src):
                     pdftext = parse_pdf(filepath, two_column).text
                 except:
                     pdftext = ""
-                    with open(filepath, "rb", encoding="utf-8") as pdfFileObj:
+                    with open(filepath, "rb") as pdfFileObj:
                         pdfReader = PyPDF2.PdfReader(pdfFileObj)
                         for page in tqdm(pdfReader.pages):
                             pdftext += page.extract_text()
-                texts = [Document(page_content=pdftext, metadata={"source": filepath})]
+                texts = [Document(page_content=pdftext,
+                                  metadata={"source": filepath})]
             elif file_type == ".docx":
                 logging.debug("Loading Word...")
                 from langchain.document_loaders import UnstructuredWordDocumentLoader
@@ -72,7 +60,8 @@ def get_documents(file_src):
                 text_list = excel_to_string(filepath)
                 texts = []
                 for elem in text_list:
-                    texts.append(Document(page_content=elem, metadata={"source": filepath}))
+                    texts.append(Document(page_content=elem,
+                                 metadata={"source": filepath}))
             else:
                 logging.debug("Loading text file...")
                 from langchain.document_loaders import TextLoader
@@ -111,14 +100,20 @@ def construct_index(
     embedding_limit = None if embedding_limit == 0 else embedding_limit
     separator = " " if separator == "" else separator
 
-    index_name = get_index_name(file_src)
+    index_name = get_file_hash(file_src)
     index_path = f"./index/{index_name}"
     if local_embedding:
         from langchain.embeddings.huggingface import HuggingFaceEmbeddings
-        embeddings = HuggingFaceEmbeddings(model_name = "sentence-transformers/distiluse-base-multilingual-cased-v2")
+        embeddings = HuggingFaceEmbeddings(
+            model_name="sentence-transformers/distiluse-base-multilingual-cased-v2")
     else:
         from langchain.embeddings import OpenAIEmbeddings
-        embeddings = OpenAIEmbeddings(openai_api_base=os.environ.get("OPENAI_API_BASE", None), openai_api_key=os.environ.get("OPENAI_EMBEDDING_API_KEY", api_key))
+        if os.environ.get("OPENAI_API_TYPE", "openai") == "openai":
+            embeddings = OpenAIEmbeddings(openai_api_base=os.environ.get(
+                "OPENAI_API_BASE", None), openai_api_key=os.environ.get("OPENAI_EMBEDDING_API_KEY", api_key))
+        else:
+            embeddings = OpenAIEmbeddings(deployment=os.environ["AZURE_EMBEDDING_DEPLOYMENT_NAME"], openai_api_key=os.environ["AZURE_OPENAI_API_KEY"],
+                                          model=os.environ["AZURE_EMBEDDING_MODEL_NAME"], openai_api_base=os.environ["AZURE_OPENAI_API_BASE_URL"], openai_api_type="azure")
     if os.path.exists(index_path):
         logging.info("找到了缓存的索引文件，加载中……")
         return FAISS.load_local(index_path, embeddings)

modules/models/Google_PaLM.py

@@ -0,0 +1,26 @@
+from .base_model import BaseLLMModel
+import google.generativeai as palm
+
+class Google_PaLM_Client(BaseLLMModel):
+    def __init__(self, model_name, api_key, user_name="") -> None:
+        super().__init__(model_name=model_name, user=user_name)
+        self.api_key = api_key
+
+    def _get_palm_style_input(self):
+        new_history = []
+        for item in self.history:
+            if item["role"] == "user":
+                new_history.append({'author': '1', 'content': item["content"]})
+            else:
+                new_history.append({'author': '0', 'content': item["content"]})
+        return new_history
+
+    def get_answer_at_once(self):
+        palm.configure(api_key=self.api_key)
+        messages = self._get_palm_style_input()
+        response = palm.chat(context=self.system_prompt, messages=messages, temperature=self.temperature, top_p=self.top_p)
+        if response.last is not None:
+            return response.last, len(response.last)
+        else:
+            reasons = '\n\n'.join(reason['reason'].name for reason in response.filters)
+            return "由于下面的原因，Google 拒绝返回 PaLM 的回答：\n\n" + reasons, 0

modules/models/azure.py

@@ -0,0 +1,17 @@
+from langchain.chat_models import AzureChatOpenAI
+import os
+
+from .base_model import Base_Chat_Langchain_Client
+
+# load_config_to_environ(["azure_openai_api_key", "azure_api_base_url", "azure_openai_api_version", "azure_deployment_name"])
+
+class Azure_OpenAI_Client(Base_Chat_Langchain_Client):
+    def setup_model(self):
+        # inplement this to setup the model then return it
+        return AzureChatOpenAI(
+            openai_api_base=os.environ["AZURE_OPENAI_API_BASE_URL"],
+            openai_api_version=os.environ["AZURE_OPENAI_API_VERSION"],
+            deployment_name=os.environ["AZURE_DEPLOYMENT_NAME"],
+            openai_api_key=os.environ["AZURE_OPENAI_API_KEY"],
+            openai_api_type="azure",
+        )

modules/models/base_model.py

@@ -29,6 +29,8 @@ from langchain.input import print_text
 from langchain.schema import AgentAction, AgentFinish, LLMResult
 from threading import Thread, Condition
 from collections import deque
+from langchain.chat_models.base import BaseChatModel
+from langchain.schema import HumanMessage, AIMessage, SystemMessage, BaseMessage
 
 from ..presets import *
 from ..index_func import *
@@ -36,6 +38,7 @@ from ..utils import *
 from .. import shared
 from ..config import retrieve_proxy
 
+
 class CallbackToIterator:
     def __init__(self):
         self.queue = deque()
@@ -52,7 +55,8 @@ class CallbackToIterator:
 
     def __next__(self):
         with self.cond:
-            while not self.queue and not self.finished:  # Wait for a value to be added to the queue.
+            # Wait for a value to be added to the queue.
+            while not self.queue and not self.finished:
                 self.cond.wait()
             if not self.queue:
                 raise StopIteration()
@@ -63,6 +67,7 @@ class CallbackToIterator:
             self.finished = True
             self.cond.notify()  # Wake up the generator if it's waiting.
 
+
 def get_action_description(text):
     match = re.search('```(.*?)```', text, re.S)
     json_text = match.group(1)
@@ -72,10 +77,11 @@ def get_action_description(text):
     action_name = json_dict['action']
     action_input = json_dict['action_input']
     if action_name != "Final Answer":
-        return f'<p style="font-size: smaller; color: gray;">{action_name}: {action_input}</p>'
+        return f'<!-- S O PREFIX --><p class="agent-prefix">{action_name}: {action_input}\n\n</p><!-- E O PREFIX -->'
     else:
         return ""
 
+
 class ChuanhuCallbackHandler(BaseCallbackHandler):
 
     def __init__(self, callback) -> None:
@@ -117,6 +123,10 @@ class ChuanhuCallbackHandler(BaseCallbackHandler):
         """Run on new LLM token. Only available when streaming is enabled."""
         self.callback(token)
 
+    def on_chat_model_start(self, serialized: Dict[str, Any], messages: List[List[BaseMessage]],  **kwargs: Any) -> Any:
+        """Run when a chat model starts running."""
+        pass
+
 
 class ModelType(Enum):
     Unknown = -1
@@ -129,6 +139,9 @@ class ModelType(Enum):
     YuanAI = 6
     Minimax = 7
     ChuanhuAgent = 8
+    GooglePaLM = 9
+    LangchainChat = 10
+    Midjourney = 11
 
     @classmethod
     def get_type(cls, model_name: str):
@@ -152,6 +165,12 @@ class ModelType(Enum):
             model_type = ModelType.Minimax
         elif "川虎助理" in model_name_lower:
             model_type = ModelType.ChuanhuAgent
+        elif "palm" in model_name_lower:
+            model_type = ModelType.GooglePaLM
+        elif "midjourney" in model_name_lower:
+            model_type = ModelType.Midjourney
+        elif "azure" in model_name_lower or "api" in model_name_lower:
+            model_type = ModelType.LangchainChat
         else:
             model_type = ModelType.Unknown
         return model_type
@@ -161,7 +180,7 @@ class BaseLLMModel:
     def __init__(
         self,
         model_name,
-        system_prompt="",
+        system_prompt=INITIAL_SYSTEM_PROMPT,
         temperature=1.0,
         top_p=1.0,
         n_choices=1,
@@ -201,7 +220,8 @@ class BaseLLMModel:
         conversations are stored in self.history, with the most recent question, in OpenAI format
         should return a generator, each time give the next word (str) in the answer
         """
-        logging.warning("stream predict not implemented, using at once predict instead")
+        logging.warning(
+            "stream predict not implemented, using at once predict instead")
         response, _ = self.get_answer_at_once()
         yield response
 
@@ -212,7 +232,8 @@ class BaseLLMModel:
         the answer (str)
         total token count (int)
         """
-        logging.warning("at once predict not implemented, using stream predict instead")
+        logging.warning(
+            "at once predict not implemented, using stream predict instead")
         response_iter = self.get_answer_stream_iter()
         count = 0
         for response in response_iter:
@@ -246,7 +267,8 @@ class BaseLLMModel:
         stream_iter = self.get_answer_stream_iter()
 
         if display_append:
-            display_append = "<hr>" +display_append
+            display_append = '\n\n<hr class="append-display no-in-raw" />' + display_append
+        partial_text = ""
         for partial_text in stream_iter:
             chatbot[-1] = (chatbot[-1][0], partial_text + display_append)
             self.all_token_counts[-1] += 1
@@ -273,9 +295,11 @@ class BaseLLMModel:
             self.history[-2] = construct_user(fake_input)
         chatbot[-1] = (chatbot[-1][0], ai_reply + display_append)
         if fake_input is not None:
-            self.all_token_counts[-1] += count_token(construct_assistant(ai_reply))
+            self.all_token_counts[-1] += count_token(
+                construct_assistant(ai_reply))
         else:
-            self.all_token_counts[-1] = total_token_count - sum(self.all_token_counts)
+            self.all_token_counts[-1] = total_token_count - \
+                sum(self.all_token_counts)
         status_text = self.token_message()
         return chatbot, status_text
 
@@ -299,10 +323,13 @@ class BaseLLMModel:
             from langchain.chat_models import ChatOpenAI
             from langchain.callbacks import StdOutCallbackHandler
             prompt_template = "Write a concise summary of the following:\n\n{text}\n\nCONCISE SUMMARY IN " + language + ":"
-            PROMPT = PromptTemplate(template=prompt_template, input_variables=["text"])
+            PROMPT = PromptTemplate(
+                template=prompt_template, input_variables=["text"])
             llm = ChatOpenAI()
-            chain = load_summarize_chain(llm, chain_type="map_reduce", return_intermediate_steps=True, map_prompt=PROMPT, combine_prompt=PROMPT)
-            summary = chain({"input_documents": list(index.docstore.__dict__["_dict"].values())}, return_only_outputs=True)["output_text"]
+            chain = load_summarize_chain(
+                llm, chain_type="map_reduce", return_intermediate_steps=True, map_prompt=PROMPT, combine_prompt=PROMPT)
+            summary = chain({"input_documents": list(index.docstore.__dict__[
+                            "_dict"].values())}, return_only_outputs=True)["output_text"]
             print(i18n("总结") + f": {summary}")
             chatbot.append([i18n("上传了")+str(len(files))+"个文件", summary])
         return chatbot, status
@@ -323,9 +350,12 @@ class BaseLLMModel:
             msg = "索引获取成功，生成回答中……"
             logging.info(msg)
             with retrieve_proxy():
-                retriever = VectorStoreRetriever(vectorstore=index, search_type="similarity_score_threshold",search_kwargs={"k":6, "score_threshold": 0.5})
-                relevant_documents = retriever.get_relevant_documents(real_inputs)
-            reference_results = [[d.page_content.strip("�"), os.path.basename(d.metadata["source"])] for d in relevant_documents]
+                retriever = VectorStoreRetriever(vectorstore=index, search_type="similarity_score_threshold", search_kwargs={
+                                                 "k": 6, "score_threshold": 0.5})
+                relevant_documents = retriever.get_relevant_documents(
+                    real_inputs)
+            reference_results = [[d.page_content.strip("�"), os.path.basename(
+                d.metadata["source"])] for d in relevant_documents]
             reference_results = add_source_numbers(reference_results)
             display_append = add_details(reference_results)
             display_append = "\n\n" + "".join(display_append)
@@ -348,10 +378,12 @@ class BaseLLMModel:
                 reference_results.append([result['body'], result['href']])
                 display_append.append(
                     # f"{idx+1}. [{domain_name}]({result['href']})\n"
-                    f"<li><a href=\"{result['href']}\" target=\"_blank\">{result['title']}</a></li>\n"
+                    f"<a href=\"{result['href']}\" target=\"_blank\">{idx+1}.&nbsp;{result['title']}</a>"
                 )
             reference_results = add_source_numbers(reference_results)
-            display_append = "<ol>\n\n" + "".join(display_append) + "</ol>"
+            # display_append = "<ol>\n\n" + "".join(display_append) + "</ol>"
+            display_append = '<div class = "source-a">' + \
+                "".join(display_append) + '</div>'
             real_inputs = (
                 replace_today(WEBSEARCH_PTOMPT_TEMPLATE)
                 .replace("{query}", real_inputs)
@@ -375,14 +407,16 @@ class BaseLLMModel:
 
         status_text = "开始生成回答……"
         logging.info(
-             "用户" + f"{self.user_identifier}" + "的输入为：" + colorama.Fore.BLUE + f"{inputs}" + colorama.Style.RESET_ALL
+            "用户" + f"{self.user_identifier}" + "的输入为：" +
+            colorama.Fore.BLUE + f"{inputs}" + colorama.Style.RESET_ALL
         )
         if should_check_token_count:
             yield chatbot + [(inputs, "")], status_text
         if reply_language == "跟随问题语言（不稳定）":
             reply_language = "the same language as the question, such as English, 中文, 日本語, Español, Français, or Deutsch."
 
-        limited_context, fake_inputs, display_append, inputs, chatbot = self.prepare_inputs(real_inputs=inputs, use_websearch=use_websearch, files=files, reply_language=reply_language, chatbot=chatbot)
+        limited_context, fake_inputs, display_append, inputs, chatbot = self.prepare_inputs(
+            real_inputs=inputs, use_websearch=use_websearch, files=files, reply_language=reply_language, chatbot=chatbot)
         yield chatbot + [(fake_inputs, "")], status_text
 
         if (
@@ -434,7 +468,7 @@ class BaseLLMModel:
                 yield chatbot, status_text
         except Exception as e:
             traceback.print_exc()
-            status_text = STANDARD_ERROR_MSG + str(e)
+            status_text = STANDARD_ERROR_MSG + beautify_err_msg(str(e))
             yield chatbot, status_text
 
         if len(self.history) > 1 and self.history[-1]["content"] != inputs:
@@ -568,10 +602,13 @@ class BaseLLMModel:
         self.system_prompt = new_system_prompt
 
     def set_key(self, new_access_key):
-        self.api_key = new_access_key.strip()
-        msg = i18n("API密钥更改为了") + hide_middle_chars(self.api_key)
-        logging.info(msg)
-        return self.api_key, msg
+        if "*" not in new_access_key:
+            self.api_key = new_access_key.strip()
+            msg = i18n("API密钥更改为了") + hide_middle_chars(self.api_key)
+            logging.info(msg)
+            return self.api_key, msg
+        else:
+            return gr.update(), gr.update()
 
     def set_single_turn(self, new_single_turn):
         self.single_turn = new_single_turn
@@ -580,7 +617,8 @@ class BaseLLMModel:
         self.history = []
         self.all_token_counts = []
         self.interrupted = False
-        pathlib.Path(os.path.join(HISTORY_DIR, self.user_identifier, new_auto_history_filename(os.path.join(HISTORY_DIR, self.user_identifier)))).touch()
+        pathlib.Path(os.path.join(HISTORY_DIR, self.user_identifier, new_auto_history_filename(
+            os.path.join(HISTORY_DIR, self.user_identifier)))).touch()
         return [], self.token_message([0])
 
     def delete_first_conversation(self):
@@ -623,7 +661,8 @@ class BaseLLMModel:
 
     def auto_save(self, chatbot):
         history_file_path = get_history_filepath(self.user_identifier)
-        save_file(history_file_path, self.system_prompt, self.history, chatbot, self.user_identifier)
+        save_file(history_file_path, self.system_prompt,
+                  self.history, chatbot, self.user_identifier)
 
     def export_markdown(self, filename, chatbot, user_name):
         if filename == "":
@@ -639,7 +678,8 @@ class BaseLLMModel:
             filename = filename.name
         try:
             if "/" not in filename:
-                history_file_path = os.path.join(HISTORY_DIR, user_name, filename)
+                history_file_path = os.path.join(
+                    HISTORY_DIR, user_name, filename)
             else:
                 history_file_path = filename
             with open(history_file_path, "r", encoding="utf-8") as f:
@@ -665,15 +705,33 @@ class BaseLLMModel:
             logging.info(f"没有找到对话历史记录 {filename}")
             return gr.update(), self.system_prompt, gr.update()
 
+    def delete_chat_history(self, filename, user_name):
+        if filename == "CANCELED":
+            return gr.update(), gr.update(), gr.update()
+        if filename == "":
+            return i18n("你没有选择任何对话历史"), gr.update(), gr.update()
+        if not filename.endswith(".json"):
+            filename += ".json"
+        if "/" not in filename:
+            history_file_path = os.path.join(HISTORY_DIR, user_name, filename)
+        else:
+            history_file_path = filename
+        try:
+            os.remove(history_file_path)
+            return i18n("删除对话历史成功"), get_history_names(False, user_name), []
+        except:
+            logging.info(f"删除对话历史失败 {history_file_path}")
+            return i18n("对话历史")+filename+i18n("已经被删除啦"), gr.update(), gr.update()
+
     def auto_load(self):
         if self.user_identifier == "":
             self.reset()
             return self.system_prompt, gr.update()
         history_file_path = get_history_filepath(self.user_identifier)
-        filename, system_prompt, chatbot = self.load_chat_history(history_file_path, self.user_identifier)
+        filename, system_prompt, chatbot = self.load_chat_history(
+            history_file_path, self.user_identifier)
         return system_prompt, chatbot
 
-
     def like(self):
         """like the last response, implement if needed
         """
@@ -683,3 +741,47 @@ class BaseLLMModel:
         """dislike the last response, implement if needed
         """
         return gr.update()
+
+
+class Base_Chat_Langchain_Client(BaseLLMModel):
+    def __init__(self, model_name, user_name=""):
+        super().__init__(model_name, user=user_name)
+        self.need_api_key = False
+        self.model = self.setup_model()
+
+    def setup_model(self):
+        # inplement this to setup the model then return it
+        pass
+
+    def _get_langchain_style_history(self):
+        history = [SystemMessage(content=self.system_prompt)]
+        for i in self.history:
+            if i["role"] == "user":
+                history.append(HumanMessage(content=i["content"]))
+            elif i["role"] == "assistant":
+                history.append(AIMessage(content=i["content"]))
+        return history
+
+    def get_answer_at_once(self):
+        assert isinstance(
+            self.model, BaseChatModel), "model is not instance of LangChain BaseChatModel"
+        history = self._get_langchain_style_history()
+        response = self.model.generate(history)
+        return response.content, sum(response.content)
+
+    def get_answer_stream_iter(self):
+        it = CallbackToIterator()
+        assert isinstance(
+            self.model, BaseChatModel), "model is not instance of LangChain BaseChatModel"
+        history = self._get_langchain_style_history()
+
+        def thread_func():
+            self.model(messages=history, callbacks=[
+                ChuanhuCallbackHandler(it.callback)])
+            it.finish()
+        t = Thread(target=thread_func)
+        t.start()
+        partial_text = ""
+        for value in it:
+            partial_text += value
+            yield partial_text

modules/models/midjourney.py

@@ -0,0 +1,385 @@
+import base64
+import io
+import json
+import logging
+import pathlib
+import time
+import tempfile
+import os
+
+from datetime import datetime
+
+import requests
+import tiktoken
+from PIL import Image
+
+from modules.config import retrieve_proxy
+from modules.models.models import XMChat
+
+mj_proxy_api_base = os.getenv("MIDJOURNEY_PROXY_API_BASE")
+mj_discord_proxy_url = os.getenv("MIDJOURNEY_DISCORD_PROXY_URL")
+mj_temp_folder = os.getenv("MIDJOURNEY_TEMP_FOLDER")
+
+
+class Midjourney_Client(XMChat):
+
+    class FetchDataPack:
+        """
+        A class to store data for current fetching data from Midjourney API
+        """
+
+        action: str  # current action, e.g. "IMAGINE", "UPSCALE", "VARIATION"
+        prefix_content: str  # prefix content, task description and process hint
+        task_id: str  # task id
+        start_time: float  # task start timestamp
+        timeout: int  # task timeout in seconds
+        finished: bool  # whether the task is finished
+        prompt: str  # prompt for the task
+
+        def __init__(self, action, prefix_content, task_id, timeout=900):
+            self.action = action
+            self.prefix_content = prefix_content
+            self.task_id = task_id
+            self.start_time = time.time()
+            self.timeout = timeout
+            self.finished = False
+
+    def __init__(self, model_name, api_key, user_name=""):
+        super().__init__(api_key, user_name)
+        self.model_name = model_name
+        self.history = []
+        self.api_key = api_key
+        self.headers = {
+            "Content-Type": "application/json",
+            "mj-api-secret": f"{api_key}"
+        }
+        self.proxy_url = mj_proxy_api_base
+        self.command_splitter = "::"
+
+        if mj_temp_folder:
+            temp = "./tmp"
+            if user_name:
+                temp = os.path.join(temp, user_name)
+            if not os.path.exists(temp):
+                os.makedirs(temp)
+            self.temp_path = tempfile.mkdtemp(dir=temp)
+            logging.info("mj temp folder: " + self.temp_path)
+        else:
+            self.temp_path = None
+
+    def use_mj_self_proxy_url(self, img_url):
+        """
+        replace discord cdn url with mj self proxy url
+        """
+        return img_url.replace(
+            "https://cdn.discordapp.com/",
+            mj_discord_proxy_url and mj_discord_proxy_url or "https://cdn.discordapp.com/"
+        )
+
+    def split_image(self, image_url):
+        """
+        when enabling temp dir, split image into 4 parts
+        """
+        with retrieve_proxy():
+            image_bytes = requests.get(image_url).content
+        img = Image.open(io.BytesIO(image_bytes))
+        width, height = img.size
+        # calculate half width and height
+        half_width = width // 2
+        half_height = height // 2
+        # create coordinates (top-left x, top-left y, bottom-right x, bottom-right y)
+        coordinates = [(0, 0, half_width, half_height),
+                       (half_width, 0, width, half_height),
+                       (0, half_height, half_width, height),
+                       (half_width, half_height, width, height)]
+
+        images = [img.crop(c) for c in coordinates]
+        return images
+
+    def auth_mj(self):
+        """
+        auth midjourney api
+        """
+        # TODO: check if secret is valid
+        return {'status': 'ok'}
+
+    def request_mj(self, path: str, action: str, data: str, retries=3):
+        """
+        request midjourney api
+        """
+        mj_proxy_url = self.proxy_url
+        if mj_proxy_url is None or not (mj_proxy_url.startswith("http://") or mj_proxy_url.startswith("https://")):
+            raise Exception('please set MIDJOURNEY_PROXY_API_BASE in ENV or in config.json')
+
+        auth_ = self.auth_mj()
+        if auth_.get('error'):
+            raise Exception('auth not set')
+
+        fetch_url = f"{mj_proxy_url}/{path}"
+        # logging.info(f"[MJ Proxy] {action} {fetch_url} params: {data}")
+
+        for _ in range(retries):
+            try:
+                with retrieve_proxy():
+                    res = requests.request(method=action, url=fetch_url, headers=self.headers, data=data)
+                break
+            except Exception as e:
+                print(e)
+
+        if res.status_code != 200:
+            raise Exception(f'{res.status_code} - {res.content}')
+
+        return res
+
+    def fetch_status(self, fetch_data: FetchDataPack):
+        """
+        fetch status of current task
+        """
+        if fetch_data.start_time + fetch_data.timeout < time.time():
+            fetch_data.finished = True
+            return "任务超时，请检查 dc 输出。描述：" + fetch_data.prompt
+
+        time.sleep(3)
+        status_res = self.request_mj(f"task/{fetch_data.task_id}/fetch", "GET", '')
+        status_res_json = status_res.json()
+        if not (200 <= status_res.status_code < 300):
+            raise Exception("任务状态获取失败：" + status_res_json.get(
+                'error') or status_res_json.get('description') or '未知错误')
+        else:
+            fetch_data.finished = False
+            if status_res_json['status'] == "SUCCESS":
+                content = status_res_json['imageUrl']
+                fetch_data.finished = True
+            elif status_res_json['status'] == "FAILED":
+                content = status_res_json['failReason'] or '未知原因'
+                fetch_data.finished = True
+            elif status_res_json['status'] == "NOT_START":
+                content = f'任务未开始，已等待 {time.time() - fetch_data.start_time:.2f} 秒'
+            elif status_res_json['status'] == "IN_PROGRESS":
+                content = '任务正在运行'
+                if status_res_json.get('progress'):
+                    content += f"，进度：{status_res_json['progress']}"
+            elif status_res_json['status'] == "SUBMITTED":
+                content = '任务已提交处理'
+            elif status_res_json['status'] == "FAILURE":
+                fetch_data.finished = True
+                return "任务处理失败，原因：" + status_res_json['failReason'] or '未知原因'
+            else:
+                content = status_res_json['status']
+            if fetch_data.finished:
+                img_url = self.use_mj_self_proxy_url(status_res_json['imageUrl'])
+                if fetch_data.action == "DESCRIBE":
+                    return f"\n{status_res_json['prompt']}"
+                time_cost_str = f"\n\n{fetch_data.action} 花费时间：{time.time() - fetch_data.start_time:.2f} 秒"
+                upscale_str = ""
+                variation_str = ""
+                if fetch_data.action in ["IMAGINE", "UPSCALE", "VARIATION"]:
+                    upscale = [f'/mj UPSCALE{self.command_splitter}{i+1}{self.command_splitter}{fetch_data.task_id}'
+                               for i in range(4)]
+                    upscale_str = '\n放大图片：\n\n' + '\n\n'.join(upscale)
+                    variation = [f'/mj VARIATION{self.command_splitter}{i+1}{self.command_splitter}{fetch_data.task_id}'
+                                 for i in range(4)]
+                    variation_str = '\n图片变体：\n\n' + '\n\n'.join(variation)
+                if self.temp_path and fetch_data.action in ["IMAGINE", "VARIATION"]:
+                    try:
+                        images = self.split_image(img_url)
+                        # save images to temp path
+                        for i in range(4):
+                            images[i].save(pathlib.Path(self.temp_path) / f"{fetch_data.task_id}_{i}.png")
+                        img_str = '\n'.join(
+                            [f"![{fetch_data.task_id}](/file={self.temp_path}/{fetch_data.task_id}_{i}.png)"
+                             for i in range(4)])
+                        return fetch_data.prefix_content + f"{time_cost_str}\n\n{img_str}{upscale_str}{variation_str}"
+                    except Exception as e:
+                        logging.error(e)
+                return fetch_data.prefix_content + \
+                    f"{time_cost_str}[![{fetch_data.task_id}]({img_url})]({img_url}){upscale_str}{variation_str}"
+            else:
+                content = f"**任务状态:** [{(datetime.now()).strftime('%Y-%m-%d %H:%M:%S')}] - {content}"
+                content += f"\n\n花费时间：{time.time() - fetch_data.start_time:.2f} 秒"
+                if status_res_json['status'] == 'IN_PROGRESS' and status_res_json.get('imageUrl'):
+                    img_url = status_res_json.get('imageUrl')
+                    return f"{content}\n[![{fetch_data.task_id}]({img_url})]({img_url})"
+                return content
+        return None
+
+    def handle_file_upload(self, files, chatbot, language):
+        """
+        handle file upload
+        """
+        if files:
+            for file in files:
+                if file.name:
+                    logging.info(f"尝试读取图像: {file.name}")
+                    self.try_read_image(file.name)
+            if self.image_path is not None:
+                chatbot = chatbot + [((self.image_path,), None)]
+            if self.image_bytes is not None:
+                logging.info("使用图片作为输入")
+        return None, chatbot, None
+
+    def reset(self):
+        self.image_bytes = None
+        self.image_path = None
+        return [], "已重置"
+
+    def get_answer_at_once(self):
+        content = self.history[-1]['content']
+        answer = self.get_help()
+
+        if not content.lower().startswith("/mj"):
+            return answer, len(content)
+
+        prompt = content[3:].strip()
+        action = "IMAGINE"
+        first_split_index = prompt.find(self.command_splitter)
+        if first_split_index > 0:
+            action = prompt[:first_split_index]
+        if action not in ["IMAGINE", "DESCRIBE", "UPSCALE",
+                          # "VARIATION", "BLEND", "REROLL"
+                          ]:
+            raise Exception("任务提交失败：未知的任务类型")
+        else:
+            action_index = None
+            action_use_task_id = None
+            if action in ["VARIATION", "UPSCALE", "REROLL"]:
+                action_index = int(prompt[first_split_index + 2:first_split_index + 3])
+                action_use_task_id = prompt[first_split_index + 5:]
+
+            try:
+                res = None
+                if action == "IMAGINE":
+                    data = {
+                        "prompt": prompt
+                    }
+                    if self.image_bytes is not None:
+                        data["base64"] = 'data:image/png;base64,' + self.image_bytes
+                    res = self.request_mj("submit/imagine", "POST",
+                                          json.dumps(data))
+                elif action == "DESCRIBE":
+                    res = self.request_mj("submit/describe", "POST",
+                                          json.dumps({"base64": 'data:image/png;base64,' + self.image_bytes}))
+                elif action == "BLEND":
+                    res = self.request_mj("submit/blend", "POST", json.dumps(
+                        {"base64Array": [self.image_bytes, self.image_bytes]}))
+                elif action in ["UPSCALE", "VARIATION", "REROLL"]:
+                    res = self.request_mj(
+                        "submit/change", "POST",
+                        json.dumps({"action": action, "index": action_index, "taskId": action_use_task_id}))
+                res_json = res.json()
+                if not (200 <= res.status_code < 300) or (res_json['code'] not in [1, 22]):
+                    answer = "任务提交失败：" + res_json.get('error', res_json.get('description', '未知错误'))
+                else:
+                    task_id = res_json['result']
+                    prefix_content = f"**画面描述:** {prompt}\n**任务ID:** {task_id}\n"
+
+                    fetch_data = Midjourney_Client.FetchDataPack(
+                        action=action,
+                        prefix_content=prefix_content,
+                        task_id=task_id,
+                    )
+                    fetch_data.prompt = prompt
+                    while not fetch_data.finished:
+                        answer = self.fetch_status(fetch_data)
+            except Exception as e:
+                logging.error("submit failed", e)
+                answer = "任务提交错误：" + str(e.args[0]) if e.args else '未知错误'
+
+        return answer, tiktoken.get_encoding("cl100k_base").encode(content)
+
+    def get_answer_stream_iter(self):
+        content = self.history[-1]['content']
+        answer = self.get_help()
+
+        if not content.lower().startswith("/mj"):
+            yield answer
+            return
+
+        prompt = content[3:].strip()
+        action = "IMAGINE"
+        first_split_index = prompt.find(self.command_splitter)
+        if first_split_index > 0:
+            action = prompt[:first_split_index]
+        if action not in ["IMAGINE", "DESCRIBE", "UPSCALE",
+                          "VARIATION", "BLEND", "REROLL"
+                          ]:
+            yield "任务提交失败：未知的任务类型"
+            return
+
+        action_index = None
+        action_use_task_id = None
+        if action in ["VARIATION", "UPSCALE", "REROLL"]:
+            action_index = int(prompt[first_split_index + 2:first_split_index + 3])
+            action_use_task_id = prompt[first_split_index + 5:]
+
+        try:
+            res = None
+            if action == "IMAGINE":
+                data = {
+                    "prompt": prompt
+                }
+                if self.image_bytes is not None:
+                    data["base64"] = 'data:image/png;base64,' + self.image_bytes
+                res = self.request_mj("submit/imagine", "POST",
+                                      json.dumps(data))
+            elif action == "DESCRIBE":
+                res = self.request_mj("submit/describe", "POST", json.dumps(
+                    {"base64": 'data:image/png;base64,' + self.image_bytes}))
+            elif action == "BLEND":
+                res = self.request_mj("submit/blend", "POST", json.dumps(
+                    {"base64Array": [self.image_bytes, self.image_bytes]}))
+            elif action in ["UPSCALE", "VARIATION", "REROLL"]:
+                res = self.request_mj(
+                    "submit/change", "POST",
+                    json.dumps({"action": action, "index": action_index, "taskId": action_use_task_id}))
+            res_json = res.json()
+            if not (200 <= res.status_code < 300) or (res_json['code'] not in [1, 22]):
+                yield "任务提交失败：" + res_json.get('error', res_json.get('description', '未知错误'))
+            else:
+                task_id = res_json['result']
+                prefix_content = f"**画面描述:** {prompt}\n**任务ID:** {task_id}\n"
+                content = f"[{(datetime.now()).strftime('%Y-%m-%d %H:%M:%S')}] - 任务提交成功：" + \
+                    res_json.get('description') or '请稍等片刻'
+                yield content
+
+                fetch_data = Midjourney_Client.FetchDataPack(
+                    action=action,
+                    prefix_content=prefix_content,
+                    task_id=task_id,
+                )
+                while not fetch_data.finished:
+                    yield self.fetch_status(fetch_data)
+        except Exception as e:
+            logging.error('submit failed', e)
+            yield "任务提交错误：" + str(e.args[0]) if e.args else '未知错误'
+
+    def get_help(self):
+        return """```
+【绘图帮助】
+所有命令都需要以 /mj 开头，如：/mj a dog
+IMAGINE - 绘图，可以省略该命令，后面跟上绘图内容
+    /mj a dog
+    /mj IMAGINE::a cat
+DESCRIBE - 描述图片，需要在右下角上传需要描述的图片内容
+    /mj DESCRIBE::
+UPSCALE - 确认后放大图片，第一个数值为需要放大的图片（1~4），第二参数为任务ID
+    /mj UPSCALE::1::123456789
+    请使用SD进行UPSCALE
+VARIATION - 图片变体，第一个数值为需要放大的图片（1~4），第二参数为任务ID
+    /mj VARIATION::1::123456789
+    
+【绘图参数】
+所有命令默认会带上参数--v 5.2
+其他参数参照 https://docs.midjourney.com/docs/parameter-list
+长宽比 --aspect/--ar
+    --ar 1:2
+    --ar 16:9
+负面tag --no
+    --no plants
+    --no hands
+随机种子 --seed
+    --seed 1
+生成动漫风格（NijiJourney） --niji
+    --niji
+```
+"""

modules/models/models.py

@@ -24,7 +24,7 @@ from ..presets import *
 from ..index_func import *
 from ..utils import *
 from .. import shared
-from ..config import retrieve_proxy, usage_limit
+from ..config import retrieve_proxy, usage_limit, sensitive_id
 from modules import config
 from .base_model import BaseLLMModel, ModelType
 
@@ -87,21 +87,23 @@ class OpenAIClient(BaseLLMModel):
             try:
                 usage_data = self._get_billing_data(usage_url)
             except Exception as e:
-                logging.error(f"获取API使用情况失败:" + str(e))
+                # logging.error(f"获取API使用情况失败: " + str(e))
+                if "Invalid authorization header" in str(e):
+                    return i18n("**获取API使用情况失败**，需在填写`config.json`中正确填写sensitive_id")
+                elif "Incorrect API key provided: sess" in str(e):
+                    return i18n("**获取API使用情况失败**，sensitive_id错误或已过期")
                 return i18n("**获取API使用情况失败**")
             # rounded_usage = "{:.5f}".format(usage_data["total_usage"] / 100)
             rounded_usage = round(usage_data["total_usage"] / 100, 5)
             usage_percent = round(usage_data["total_usage"] / usage_limit, 2)
+            from ..webui import get_html
             # return i18n("**本月使用金额** ") + f"\u3000 ${rounded_usage}"
-            return """\
-                <b>""" + i18n("本月使用金额") + f"""</b>
-                <div class="progress-bar">
-                    <div class="progress" style="width: {usage_percent}%;">
-                        <span class="progress-text">{usage_percent}%</span>
-                    </div>
-                </div>
-                <div style="display: flex; justify-content: space-between;"><span>${rounded_usage}</span><span>${usage_limit}</span></div>
-                """
+            return get_html("billing_info.html").format(
+                    label = i18n("本月使用金额"),
+                    usage_percent = usage_percent,
+                    rounded_usage = rounded_usage,
+                    usage_limit = usage_limit
+                )
         except requests.exceptions.ConnectTimeout:
             status_text = (
                 STANDARD_ERROR_MSG + CONNECTION_TIMEOUT_MSG + ERROR_RETRIEVE_MSG
@@ -161,7 +163,7 @@ class OpenAIClient(BaseLLMModel):
 
         # 如果有自定义的api-host，使用自定义host发送请求，否则使用默认设置发送请求
         if shared.state.completion_url != COMPLETION_URL:
-            logging.info(f"使用自定义API URL: {shared.state.completion_url}")
+            logging.debug(f"使用自定义API URL: {shared.state.completion_url}")
 
         with retrieve_proxy():
             try:
@@ -179,9 +181,10 @@ class OpenAIClient(BaseLLMModel):
     def _refresh_header(self):
         self.headers = {
             "Content-Type": "application/json",
-            "Authorization": f"Bearer {self.api_key}",
+            "Authorization": f"Bearer {sensitive_id}",
         }
 
+
     def _get_billing_data(self, billing_url):
         with retrieve_proxy():
             response = requests.get(
@@ -206,7 +209,7 @@ class OpenAIClient(BaseLLMModel):
                 chunk_length = len(chunk)
                 try:
                     chunk = json.loads(chunk[6:])
-                except json.JSONDecodeError:
+                except:
                     print(i18n("JSON解析错误,收到的内容: ") + f"{chunk}")
                     error_msg += chunk
                     continue
@@ -560,6 +563,7 @@ def get_model(
     try:
         if model_type == ModelType.OpenAI:
             logging.info(f"正在加载OpenAI模型: {model_name}")
+            access_key = os.environ.get("OPENAI_API_KEY", access_key)
             model = OpenAIClient(
                 model_name=model_name,
                 api_key=access_key,
@@ -610,16 +614,29 @@ def get_model(
         elif model_type == ModelType.ChuanhuAgent:
             from .ChuanhuAgent import ChuanhuAgent_Client
             model = ChuanhuAgent_Client(model_name, access_key, user_name=user_name)
+        elif model_type == ModelType.GooglePaLM:
+            from .Google_PaLM import Google_PaLM_Client
+            access_key = os.environ.get("GOOGLE_PALM_API_KEY", access_key)
+            model = Google_PaLM_Client(model_name, access_key, user_name=user_name)
+        elif model_type == ModelType.LangchainChat:
+            from .azure import Azure_OpenAI_Client
+            model = Azure_OpenAI_Client(model_name, user_name=user_name)
+        elif model_type == ModelType.Midjourney:
+            from .midjourney import Midjourney_Client
+            mj_proxy_api_secret = os.getenv("MIDJOURNEY_PROXY_API_SECRET")
+            model = Midjourney_Client(model_name, mj_proxy_api_secret, user_name=user_name)
         elif model_type == ModelType.Unknown:
             raise ValueError(f"未知模型: {model_name}")
         logging.info(msg)
     except Exception as e:
-        logging.error(e)
+        import traceback
+        traceback.print_exc()
         msg = f"{STANDARD_ERROR_MSG}: {e}"
+    presudo_key = hide_middle_chars(access_key)
     if dont_change_lora_selector:
-        return model, msg, chatbot
+        return model, msg, chatbot, gr.update(), access_key, presudo_key
     else:
-        return model, msg, chatbot, gr.Dropdown.update(choices=lora_choices, visible=lora_selector_visibility)
+        return model, msg, chatbot, gr.Dropdown.update(choices=lora_choices, visible=lora_selector_visibility), access_key, presudo_key
 
 
 if __name__ == "__main__":

modules/overwrites.py

@@ -71,23 +71,36 @@ def postprocess_chat_messages(
         else:
             raise ValueError(f"Invalid message for Chatbot component: {chat_message}")
 
-with open("./assets/custom.js", "r", encoding="utf-8") as f, \
-    open("./assets/external-scripts.js", "r", encoding="utf-8") as f1:
-    customJS = f.read()
-    externalScripts = f1.read()
-
-
-def reload_javascript():
-    print("Reloading javascript...")
-    js = f'<script>{customJS}</script><script async>{externalScripts}</script>'
-    # if render_latex:
-    #     js += """\"""
-    def template_response(*args, **kwargs):
-        res = GradioTemplateResponseOriginal(*args, **kwargs)
-        res.body = res.body.replace(b'</html>', f'{js}</html>'.encode("utf8"))
-        res.init_headers()
-        return res
-
-    gr.routes.templates.TemplateResponse = template_response
-
-GradioTemplateResponseOriginal = gr.routes.templates.TemplateResponse
+
+
+def add_classes_to_gradio_component(comp):
+    """
+    this adds gradio-* to the component for css styling (ie gradio-button to gr.Button), as well as some others
+    code from stable-diffusion-webui <AUTOMATIC1111/stable-diffusion-webui>
+    """
+
+    comp.elem_classes = [f"gradio-{comp.get_block_name()}", *(comp.elem_classes or [])]
+
+    if getattr(comp, 'multiselect', False):
+        comp.elem_classes.append('multiselect')
+
+
+def IOComponent_init(self, *args, **kwargs):
+    res = original_IOComponent_init(self, *args, **kwargs)
+    add_classes_to_gradio_component(self)
+
+    return res
+
+original_IOComponent_init = gr.components.IOComponent.__init__
+gr.components.IOComponent.__init__ = IOComponent_init
+
+
+def BlockContext_init(self, *args, **kwargs):
+    res = original_BlockContext_init(self, *args, **kwargs)
+    add_classes_to_gradio_component(self)
+
+    return res
+
+original_BlockContext_init = gr.blocks.BlockContext.__init__
+gr.blocks.BlockContext.__init__ = BlockContext_init
+

modules/presets.py

@@ -60,19 +60,24 @@ ONLINE_MODELS = [
     "gpt-4-32k-0613",
     "川虎助理",
     "川虎助理 Pro",
+    "GooglePaLM",
     "xmchat",
+    "Azure OpenAI",
     "yuanai-1.0-base_10B",
     "yuanai-1.0-translate",
     "yuanai-1.0-dialog",
     "yuanai-1.0-rhythm_poems",
     "minimax-abab4-chat",
     "minimax-abab5-chat",
+    "midjourney"
 ]
 
 LOCAL_MODELS = [
     "chatglm-6b",
     "chatglm-6b-int4",
-    "chatglm-6b-int4-qe",
+    "chatglm-6b-int4-ge",
+    "chatglm2-6b",
+    "chatglm2-6b-int4",
     "StableLM",
     "MOSS",
     "llama-7b-hf",
@@ -121,6 +126,7 @@ REPLY_LANGUAGES = [
     "Español",
     "Français",
     "Deutsch",
+    "한국어",
     "跟随问题语言（不稳定）"
 ]
 
@@ -170,6 +176,8 @@ SUMMARIZE_PROMPT = """Write a concise summary of the following:
 CONCISE SUMMARY IN 中文:"""
 
 ALREADY_CONVERTED_MARK = "<!-- ALREADY CONVERTED BY PARSER. -->"
+START_OF_OUTPUT_MARK = "<!-- SOO IN MESSAGE -->"
+END_OF_OUTPUT_MARK = "<!-- EOO IN MESSAGE -->"
 
 small_and_beautiful_theme = gr.themes.Soft(
         primary_hue=gr.themes.Color(
@@ -222,7 +230,7 @@ small_and_beautiful_theme = gr.themes.Soft(
         # button_primary_background_fill_hover="*primary_400",
         # button_primary_border_color="*primary_500",
         button_primary_border_color_dark="*primary_600",
-        button_primary_text_color="wihte",
+        button_primary_text_color="white",
         button_primary_text_color_dark="white",
         button_secondary_background_fill="*neutral_100",
         button_secondary_background_fill_hover="*neutral_50",

modules/repo.py

@@ -0,0 +1,239 @@
+# -*- coding:utf-8 -*-
+import os
+import sys
+import subprocess
+from functools import lru_cache
+import logging
+import gradio as gr
+import datetime
+
+# This file is mainly used to describe repo version info, execute the git command, python pip command, shell command, etc.
+# Part of the code in this file is referenced from stable-diffusion-webui/modules/launch_utils.py
+
+python = sys.executable
+pip = os.environ.get('PIP', "pip")
+git = os.environ.get('GIT', "git")
+
+# Pypi index url
+index_url = os.environ.get('INDEX_URL', "")
+
+# Whether to default to printing command output
+default_command_live = True
+
+
+def run(command, desc=None, errdesc=None, custom_env=None, live: bool = default_command_live) -> str:
+    if desc is not None:
+        print(desc)
+    run_kwargs = {
+        "args": command,
+        "shell": True,
+        "env": os.environ if custom_env is None else custom_env,
+        "encoding": 'utf8',
+        "errors": 'ignore',
+    }
+
+    if not live:
+        run_kwargs["stdout"] = run_kwargs["stderr"] = subprocess.PIPE
+
+    result = subprocess.run(**run_kwargs)
+    if result.returncode != 0:
+        error_bits = [
+            f"{errdesc or 'Error running command'}.",
+            f"Command: {command}",
+            f"Error code: {result.returncode}",
+        ]
+        if result.stdout:
+            error_bits.append(f"stdout: {result.stdout}")
+        if result.stderr:
+            error_bits.append(f"stderr: {result.stderr}")
+        raise RuntimeError("\n".join(error_bits))
+
+    return (result.stdout or "")
+
+
+def run_pip(command, desc=None, pref=None, live=default_command_live):
+    # if args.skip_install:
+    #     return
+
+    index_url_line = f' --index-url {index_url}' if index_url != '' else ''
+    return run(
+        f'"{python}" -m pip {command} --prefer-binary{index_url_line}', 
+        desc=f"{pref} Installing {desc}...", 
+        errdesc=f"Couldn't install {desc}", 
+        live=live
+    )
+
+
+@lru_cache()
+def commit_hash():
+    try:
+        return subprocess.check_output([git, "rev-parse", "HEAD"], shell=False, encoding='utf8').strip()
+    except Exception:
+        return "<none>"
+
+def commit_html():
+    commit = commit_hash()
+    if commit != "<none>":
+        short_commit = commit[0:7]
+        commit_info = f'<a style="text-decoration:none;color:inherit" href="https://github.com/GaiZhenbiao/ChuanhuChatGPT/commit/{short_commit}">{short_commit}</a>'
+    else:
+        commit_info = "unknown \U0001F615"
+    return commit_info
+
+@lru_cache()
+def tag_html():
+    try:
+        latest_tag = run(f"{git} describe --tags --abbrev=0", live=False).strip()
+        try:
+            # tag = subprocess.check_output([git, "describe", "--tags", "--exact-match"], shell=False, encoding='utf8').strip()
+            tag = run(f"{git} describe --tags --exact-match", live=False).strip()
+        except Exception: 
+            tag = "<edited>"
+    except Exception:
+        tag = "<none>"
+
+    if tag == "<none>":
+        tag_info = "unknown \U0001F615"
+    elif tag == "<edited>":
+        tag_info = f'<a style="text-decoration:none;color:inherit" href="https://github.com/GaiZhenbiao/ChuanhuChatGPT/releases/tag/{latest_tag}">{latest_tag}</a><span style="font-size:smaller">*</span>'
+    else:
+        tag_info = f'<a style="text-decoration:none;color:inherit" href="https://github.com/GaiZhenbiao/ChuanhuChatGPT/releases/tag/{tag}">{tag}</a>'
+ 
+    return tag_info
+
+def repo_tag_html():
+    commit_version = commit_html()
+    tag_version = tag_html()
+    return tag_version if tag_version != "unknown \U0001F615" else commit_version
+
+def versions_html():
+    python_version = ".".join([str(x) for x in sys.version_info[0:3]])
+    repo_version = repo_tag_html()
+    return f"""
+        Python: <span title="{sys.version}">{python_version}</span>
+         • 
+        Gradio: {gr.__version__}
+         • 
+        <a style="text-decoration:none;color:inherit" href="https://github.com/GaiZhenbiao/ChuanhuChatGPT">ChuanhuChat</a>: {repo_version}
+        """
+
+def version_time():
+    try:
+        commit_time = subprocess.check_output(f"TZ=UTC {git} log -1 --format=%cd --date='format-local:%Y-%m-%dT%H:%M:%SZ'", shell=True, encoding='utf8').strip()
+        # commit_time = run(f"TZ=UTC {git} log -1 --format=%cd --date='format-local:%Y-%m-%dT%H:%M:%SZ'").strip()
+    except Exception:
+        commit_time = "unknown"
+    return commit_time
+
+
+
+def get_current_branch():
+    try:
+        # branch = run(f"{git} rev-parse --abbrev-ref HEAD").strip()
+        branch = subprocess.check_output([git, "rev-parse", "--abbrev-ref", "HEAD"], shell=False, encoding='utf8').strip()
+    except Exception:
+        branch = "<none>"
+    return branch
+
+
+def get_latest_release():
+    try:
+        import requests
+        release = requests.get("https://api.github.com/repos/GaiZhenbiao/ChuanhuChatGPT/releases/latest").json()
+        tag = release["tag_name"]
+        release_note = release["body"]
+        need_pip = release_note.find("requirements reinstall needed") != -1
+    except Exception:
+        tag = "<none>"
+        release_note = ""
+        need_pip = False
+    return {"tag": tag, "release_note": release_note, "need_pip": need_pip}
+
+def get_tag_commit_hash(tag):
+    try:
+        import requests
+        tags = requests.get("https://api.github.com/repos/GaiZhenbiao/ChuanhuChatGPT/tags").json()
+        commit_hash = [x["commit"]["sha"] for x in tags if x["name"] == tag][0]
+    except Exception:
+        commit_hash = "<none>"
+    return commit_hash
+
+def repo_need_stash():
+    try:
+        return subprocess.check_output([git, "diff-index", "--quiet", "HEAD", "--"], shell=False, encoding='utf8').strip() != ""
+    except Exception:
+        return True
+
+def background_update():
+    # {git} fetch --all && ({git} pull https://github.com/GaiZhenbiao/ChuanhuChatGPT.git main -f || ({git} stash && {git} pull https://github.com/GaiZhenbiao/ChuanhuChatGPT.git main -f && {git} stash pop)) && {pip} install -r requirements.txt")
+    try:
+        latest_release = get_latest_release()
+        latest_release_tag = latest_release["tag"]
+        latest_release_hash = get_tag_commit_hash(latest_release_tag)
+        need_pip = latest_release["need_pip"]
+        need_stash = repo_need_stash()
+
+        timestamp = datetime.datetime.now().strftime("%Y%m%d%H%M%S")
+        current_branch = get_current_branch()
+        updater_branch = f'tmp_{timestamp}'
+        backup_branch = f'backup_{timestamp}'
+        track_repo = "https://github.com/GaiZhenbiao/ChuanhuChatGPT.git"
+        try:
+            try:
+                run(f"{git} fetch {track_repo}", desc="[Updater] Fetching from github...", live=False)
+            except Exception:
+                logging.error(f"Update failed in fetching, check your network connection")
+                return "failed"
+            
+            run(f'{git} stash push --include-untracked -m "updater-{timestamp}"', 
+                desc=f'[Updater] Restoring you local changes on stash updater-{timestamp}', live=False) if need_stash else None
+            
+            run(f"{git} checkout -b {backup_branch}", live=False)
+            run(f"{git} checkout -b {updater_branch}", live=False)
+            run(f"{git} reset --hard FETCH_HEAD", live=False)
+            run(f"{git} reset --hard {latest_release_hash}", desc=f'[Updater] Checking out {latest_release_tag}...', live=False)
+            run(f"{git} checkout {current_branch}", live=False)
+            
+            try:
+                run(f"{git} merge --no-edit {updater_branch} -q", desc=f"[Updater] Trying to apply latest update on version {latest_release_tag}...")
+            except Exception:
+                logging.error(f"Update failed in merging")
+                try:
+                    run(f"{git} merge --abort", desc="[Updater] Conflict detected, canceling update...")
+                    run(f"{git} reset --hard {backup_branch}", live=False)
+                    run(f"{git} branch -D -f {updater_branch}", live=False)
+                    run(f"{git} branch -D -f {backup_branch}", live=False)
+                    run(f"{git} stash pop", live=False) if need_stash else None
+                    logging.error(f"Update failed, but your file was safely reset to the state before the update.")
+                    return "failed"
+                except Exception as e:
+                    logging.error(f"!!!Update failed in resetting, try to reset your files manually. {e}")
+                    return "failed"
+
+            if need_stash:
+                try:
+                    run(f"{git} stash apply", desc="[Updater] Trying to restore your local modifications...", live=False)
+                except Exception:
+                    run(f"{git} reset --hard {backup_branch}", desc="[Updater] Conflict detected, canceling update...", live=False)
+                    run(f"{git} branch -D -f {updater_branch}", live=False)
+                    run(f"{git} branch -D -f {backup_branch}", live=False)
+                    run(f"{git} stash pop", live=False)
+                    logging.error(f"Update failed in applying your local changes, but your file was safely reset to the state before the update.")
+                    return "failed"
+                run(f"{git} stash drop", live=False)
+
+            run(f"{git} branch -D -f {updater_branch}", live=False)
+            run(f"{git} branch -D -f {backup_branch}", live=False)
+        except Exception as e:
+            logging.error(f"Update failed: {e}")
+            return "failed"
+        if need_pip:
+            try: 
+                run_pip(f"install -r requirements.txt", pref="[Updater]", desc="requirements", live=False) 
+            except Exception:
+                logging.error(f"Update failed in pip install")
+                return "failed"
+        return "success"
+    except Exception as e:
+        logging.error(f"Update failed: {e}")
+        return "failed" 

modules/shared.py

@@ -62,3 +62,4 @@ state = State()
 
 modules_path = os.path.dirname(os.path.realpath(__file__))
 chuanhu_path = os.path.dirname(modules_path)
+assets_path = os.path.join(chuanhu_path, "web_assets")

modules/train_func.py

@@ -0,0 +1,161 @@
+import os
+import logging
+import traceback
+
+import openai
+import gradio as gr
+import ujson as json
+import commentjson
+import openpyxl
+
+import modules.presets as presets
+from modules.utils import get_file_hash, count_token
+from modules.presets import i18n
+
+def excel_to_jsonl(filepath, preview=False):
+    # 打开Excel文件
+    workbook = openpyxl.load_workbook(filepath)
+
+    # 获取第一个工作表
+    sheet = workbook.active
+
+    # 获取所有行数据
+    data = []
+    for row in sheet.iter_rows(values_only=True):
+        data.append(row)
+
+    # 构建字典列表
+    headers = data[0]
+    jsonl = []
+    for row in data[1:]:
+        row_data = dict(zip(headers, row))
+        if any(row_data.values()):
+            jsonl.append(row_data)
+    formatted_jsonl = []
+    for i in jsonl:
+            if "提问" in i and "答案" in i:
+                if "系统" in i :
+                    formatted_jsonl.append({
+                        "messages":[
+                            {"role": "system", "content": i["系统"]},
+                            {"role": "user", "content": i["提问"]},
+                            {"role": "assistant", "content": i["答案"]}
+                        ]
+                    })
+                else:
+                    formatted_jsonl.append({
+                        "messages":[
+                            {"role": "user", "content": i["提问"]},
+                            {"role": "assistant", "content": i["答案"]}
+                        ]
+                    })
+            else:
+                logging.warning(f"跳过一行数据，因为没有找到提问和答案: {i}")
+    return formatted_jsonl
+
+def jsonl_save_to_disk(jsonl, filepath):
+    file_hash = get_file_hash(file_paths = [filepath])
+    os.makedirs("files", exist_ok=True)
+    save_path = f"files/{file_hash}.jsonl"
+    with open(save_path, "w") as f:
+        f.write("\n".join([json.dumps(i, ensure_ascii=False) for i in jsonl]))
+    return save_path
+
+def estimate_cost(ds):
+    dialogues = []
+    for l in ds:
+        for m in l["messages"]:
+            dialogues.append(m["content"])
+    dialogues = "\n".join(dialogues)
+    tokens = count_token(dialogues)
+    return f"Token 数约为 {tokens}，预估每轮（epoch）费用约为 {tokens / 1000 * 0.008} 美元。"
+
+
+def handle_dataset_selection(file_src):
+    logging.info(f"Loading dataset {file_src.name}...")
+    preview = ""
+    if file_src.name.endswith(".jsonl"):
+        with open(file_src.name, "r") as f:
+            ds = [json.loads(l) for l in f.readlines()]
+    else:
+        ds = excel_to_jsonl(file_src.name)
+    preview = ds[0]
+
+    return preview, gr.update(interactive=True), estimate_cost(ds)
+
+def upload_to_openai(file_src):
+    openai.api_key = os.getenv("OPENAI_API_KEY")
+    dspath = file_src.name
+    msg = ""
+    logging.info(f"Uploading dataset {dspath}...")
+    if dspath.endswith(".xlsx"):
+        jsonl = excel_to_jsonl(dspath)
+        dspath = jsonl_save_to_disk(jsonl, dspath)
+    try:
+        uploaded = openai.File.create(
+            file=open(dspath, "rb"),
+            purpose='fine-tune'
+            )
+        return uploaded.id, f"上传成功"
+    except Exception as e:
+        traceback.print_exc()
+        return "", f"上传失败，原因：{ e }"
+
+def build_event_description(id, status, trained_tokens, name=i18n("暂时未知")):
+    # convert to markdown
+    return f"""
+    #### 训练任务 {id}
+
+    模型名称：{name}
+
+    状态：{status}
+
+    已经训练了 {trained_tokens} 个token
+    """
+
+def start_training(file_id, suffix, epochs):
+    openai.api_key = os.getenv("OPENAI_API_KEY")
+    try:
+        job = openai.FineTuningJob.create(training_file=file_id, model="gpt-3.5-turbo", suffix=suffix, hyperparameters={"n_epochs": epochs})
+        return build_event_description(job.id, job.status, job.trained_tokens)
+    except Exception as e:
+        traceback.print_exc()
+        if "is not ready" in str(e):
+            return "训练出错，因为文件还没准备好。OpenAI 需要一点时间准备文件，过几分钟再来试试。"
+        return f"训练失败，原因：{ e }"
+
+def get_training_status():
+    openai.api_key = os.getenv("OPENAI_API_KEY")
+    active_jobs = [build_event_description(job["id"], job["status"], job["trained_tokens"], job["fine_tuned_model"]) for job in openai.FineTuningJob.list(limit=10)["data"] if job["status"] != "cancelled"]
+    return "\n\n".join(active_jobs), gr.update(interactive=True) if len(active_jobs) > 0 else gr.update(interactive=False)
+
+def handle_dataset_clear():
+    return gr.update(value=None), gr.update(interactive=False)
+
+def add_to_models():
+    openai.api_key = os.getenv("OPENAI_API_KEY")
+    succeeded_jobs = [job for job in openai.FineTuningJob.list()["data"] if job["status"] == "succeeded"]
+    extra_models = [job["fine_tuned_model"] for job in succeeded_jobs]
+    for i in extra_models:
+        if i not in presets.MODELS:
+            presets.MODELS.append(i)
+
+    with open('config.json', 'r') as f:
+        data = commentjson.load(f)
+    if 'extra_models' in data:
+        for i in extra_models:
+            if i not in data['extra_models']:
+                data['extra_models'].append(i)
+    else:
+        data['extra_models'] = extra_models
+    with open('config.json', 'w') as f:
+        commentjson.dump(data, f, indent=4)
+
+    return gr.update(choices=presets.MODELS), f"成功添加了 {len(succeeded_jobs)} 个模型。"
+
+def cancel_all_jobs():
+    openai.api_key = os.getenv("OPENAI_API_KEY")
+    jobs = [job for job in openai.FineTuningJob.list()["data"] if job["status"] not in ["cancelled", "succeeded"]]
+    for job in jobs:
+        openai.FineTuningJob.cancel(job["id"])
+    return f"成功取消了 {len(jobs)} 个训练任务。"

modules/utils.py

@@ -2,16 +2,14 @@
 from __future__ import annotations
 from typing import TYPE_CHECKING, Any, Callable, Dict, List, Tuple, Type
 import logging
-import json
+import commentjson as json
 import os
 import datetime
-import hashlib
 import csv
 import requests
 import re
 import html
-import sys
-import subprocess
+import hashlib
 
 import gradio as gr
 from pypinyin import lazy_pinyin
@@ -47,6 +45,9 @@ def set_key(current_model, *args):
 def load_chat_history(current_model, *args):
     return current_model.load_chat_history(*args)
 
+def delete_chat_history(current_model, *args):
+    return current_model.delete_chat_history(*args)
+
 def interrupt(current_model, *args):
     return current_model.interrupt(*args)
 
@@ -125,9 +126,10 @@ def dislike(current_model, *args):
     return current_model.dislike(*args)
 
 
-def count_token(message):
+def count_token(input_str):
     encoding = tiktoken.get_encoding("cl100k_base")
-    input_str = f"role: {message['role']}, content: {message['content']}"
+    if type(input_str) == dict:
+        input_str = f"role: {input_str['role']}, content: {input_str['content']}"
     length = len(encoding.encode(input_str))
     return length
 
@@ -202,6 +204,28 @@ def convert_mdtext(md_text): # deprecated
     output += ALREADY_CONVERTED_MARK
     return output
 
+
+def clip_rawtext(chat_message, need_escape=True):
+    # first, clip hr line
+    hr_pattern = r'\n\n<hr class="append-display no-in-raw" />(.*?)'
+    hr_match = re.search(hr_pattern, chat_message, re.DOTALL)
+    message_clipped = chat_message[:hr_match.start()] if hr_match else chat_message
+    # second, avoid agent-prefix being escaped
+    agent_prefix_pattern = r'<!-- S O PREFIX --><p class="agent-prefix">(.*?)<\/p><!-- E O PREFIX -->'
+    agent_matches = re.findall(agent_prefix_pattern, message_clipped)
+    final_message = ""
+    if agent_matches:
+        agent_parts = re.split(agent_prefix_pattern, message_clipped)
+        for i, part in enumerate(agent_parts):
+            if i % 2 == 0:
+                final_message += escape_markdown(part) if need_escape else part
+            else:
+                final_message += f'<!-- S O PREFIX --><p class="agent-prefix">{part}</p><!-- E O PREFIX -->'
+    else:
+        final_message = escape_markdown(message_clipped) if need_escape else message_clipped
+    return final_message
+
+
 def convert_bot_before_marked(chat_message):
     """
     注意不能给输出加缩进, 否则会被marked解析成代码块
@@ -209,12 +233,13 @@ def convert_bot_before_marked(chat_message):
     if '<div class="md-message">' in chat_message:
         return chat_message
     else:
+        raw = f'<div class="raw-message hideM"><pre>{clip_rawtext(chat_message)}</pre></div>'
+        # really_raw = f'{START_OF_OUTPUT_MARK}<div class="really-raw hideM">{clip_rawtext(chat_message, need_escape=False)}\n</div>{END_OF_OUTPUT_MARK}'
+
         code_block_pattern = re.compile(r"```(.*?)(?:```|$)", re.DOTALL)
         code_blocks = code_block_pattern.findall(chat_message)
         non_code_parts = code_block_pattern.split(chat_message)[::2]
         result = []
-
-        raw = f'<div class="raw-message hideM">{escape_markdown(chat_message)}</div>'
         for non_code, code in zip(non_code_parts, code_blocks + [""]):
             if non_code.strip():
                 result.append(non_code)
@@ -222,7 +247,7 @@ def convert_bot_before_marked(chat_message):
                 code = f"\n```{code}\n```"
                 result.append(code)
         result = "".join(result)
-        md = f'<div class="md-message">{result}\n</div>'
+        md = f'<div class="md-message">\n\n{result}\n</div>'
         return raw + md
 
 def convert_user_before_marked(chat_message):
@@ -236,7 +261,7 @@ def escape_markdown(text):
     Escape Markdown special characters to HTML-safe equivalents.
     """
     escape_chars = {
-        ' ': '&nbsp;',
+        # ' ': '&nbsp;',
         '_': '&#95;',
         '*': '&#42;',
         '[': '&#91;',
@@ -253,8 +278,12 @@ def escape_markdown(text):
         '`': '&#96;',
         '>': '&#62;',
         '<': '&#60;',
-        '|': '&#124;'
+        '|': '&#124;',
+        '$': '&#36;',
+        ':': '&#58;',
+        '\n': '<br>',
     }
+    text = text.replace('    ', '&nbsp;&nbsp;&nbsp;&nbsp;')
     return ''.join(escape_chars.get(c, c) for c in text)
 
 
@@ -508,55 +537,19 @@ def transfer_input(inputs):
     )
 
 
+def update_chuanhu():
+    from .repo import background_update
 
-def run(command, desc=None, errdesc=None, custom_env=None, live=False):
-    if desc is not None:
-        print(desc)
-    if live:
-        result = subprocess.run(command, shell=True, env=os.environ if custom_env is None else custom_env)
-        if result.returncode != 0:
-            raise RuntimeError(f"""{errdesc or 'Error running command'}.
-                Command: {command}
-                Error code: {result.returncode}""")
-
-        return ""
-    result = subprocess.run(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True, env=os.environ if custom_env is None else custom_env)
-    if result.returncode != 0:
-        message = f"""{errdesc or 'Error running command'}.
-            Command: {command}
-            Error code: {result.returncode}
-            stdout: {result.stdout.decode(encoding="utf8", errors="ignore") if len(result.stdout)>0 else '<empty>'}
-            stderr: {result.stderr.decode(encoding="utf8", errors="ignore") if len(result.stderr)>0 else '<empty>'}
-            """
-        raise RuntimeError(message)
-    return result.stdout.decode(encoding="utf8", errors="ignore")
-
-def versions_html():
-    git = os.environ.get('GIT', "git")
-    python_version = ".".join([str(x) for x in sys.version_info[0:3]])
-    try:
-        commit_hash = run(f"{git} rev-parse HEAD").strip()
-    except Exception:
-        commit_hash = "<none>"
-    if commit_hash != "<none>":
-        short_commit = commit_hash[0:7]
-        commit_info = f"<a style=\"text-decoration:none;color:inherit\" href=\"https://github.com/GaiZhenbiao/ChuanhuChatGPT/commit/{short_commit}\">{short_commit}</a>"
+    print("[Updater] Trying to update...")
+    update_status = background_update()
+    if update_status == "success":
+        logging.info("Successfully updated, restart needed")
+        status = '<span id="update-status" class="hideK">success</span>'
+        return gr.Markdown.update(value=i18n("更新成功，请重启本程序")+status)
     else:
-        commit_info = "unknown \U0001F615"
-    return f"""
-        Python: <span title="{sys.version}">{python_version}</span>
-         • 
-        Gradio: {gr.__version__}
-         • 
-        <a style="text-decoration:none;color:inherit" href="https://github.com/GaiZhenbiao/ChuanhuChatGPT">ChuanhuChat</a>: {commit_info}
-        """
-
-def get_html(filename):
-    path = os.path.join(shared.chuanhu_path, "assets", "html", filename)
-    if os.path.exists(path):
-        with open(path, encoding="utf8") as file:
-            return file.read()
-    return ""
+        status = '<span id="update-status" class="hideK">failure</span>'
+        return gr.Markdown.update(value=i18n("更新失败，请尝试[手动更新](https://github.com/GaiZhenbiao/ChuanhuChatGPT/wiki/使用教程#手动更新)")+status)
+
 
 def add_source_numbers(lst, source_name = "Source", use_source = True):
     if use_source:
@@ -654,3 +647,37 @@ def get_history_filepath(username):
 
     latest_file = os.path.join(dirname, latest_file)
     return latest_file
+
+def beautify_err_msg(err_msg):
+    if "insufficient_quota" in  err_msg:
+        return i18n("剩余配额不足，[进一步了解](https://github.com/GaiZhenbiao/ChuanhuChatGPT/wiki/%E5%B8%B8%E8%A7%81%E9%97%AE%E9%A2%98#you-exceeded-your-current-quota-please-check-your-plan-and-billing-details)")
+    if "The model: gpt-4 does not exist" in err_msg:
+        return i18n("你没有权限访问 GPT4，[进一步了解](https://github.com/GaiZhenbiao/ChuanhuChatGPT/issues/843)")
+    if "Resource not found" in err_msg:
+        return i18n("请查看 config_example.json，配置 Azure OpenAI")
+    return err_msg
+
+def auth_from_conf(username, password):
+    try:
+        with open("config.json", encoding="utf-8") as f:
+            conf = json.load(f)
+        usernames, passwords = [i[0] for i in conf["users"]], [i[1] for i in conf["users"]]
+        if username in usernames:
+            if passwords[usernames.index(username)] == password:
+                return True
+        return False
+    except:
+        return False
+
+def get_file_hash(file_src=None, file_paths=None):
+    if file_src:
+        file_paths = [x.name for x in file_src]
+    file_paths.sort(key=lambda x: os.path.basename(x))
+
+    md5_hash = hashlib.md5()
+    for file_path in file_paths:
+        with open(file_path, "rb") as f:
+            while chunk := f.read(8192):
+                md5_hash.update(chunk)
+
+    return md5_hash.hexdigest()

modules/webui.py

@@ -0,0 +1,70 @@
+
+from collections import namedtuple
+import os
+import gradio as gr
+
+from . import shared
+
+# with open("./assets/ChuanhuChat.js", "r", encoding="utf-8") as f, \
+#     open("./assets/external-scripts.js", "r", encoding="utf-8") as f1:
+#     customJS = f.read()
+#     externalScripts = f1.read()
+
+
+def get_html(filename):
+    path = os.path.join(shared.chuanhu_path, "web_assets", "html", filename)
+    if os.path.exists(path):
+        with open(path, encoding="utf8") as file:
+            return file.read()
+    return ""
+
+def webpath(fn):
+    if fn.startswith(shared.assets_path):
+        web_path = os.path.relpath(fn, shared.chuanhu_path).replace('\\', '/')
+    else:
+        web_path = os.path.abspath(fn)
+    return f'file={web_path}?{os.path.getmtime(fn)}'
+
+ScriptFile = namedtuple("ScriptFile", ["basedir", "filename", "path"])
+
+def javascript_html():
+    head = ""
+    for script in list_scripts("javascript", ".js"):
+        head += f'<script type="text/javascript" src="{webpath(script.path)}"></script>\n'
+    for script in list_scripts("javascript", ".mjs"):
+        head += f'<script type="module" src="{webpath(script.path)}"></script>\n'
+    return head
+
+def css_html():
+    head = ""
+    for cssfile in list_scripts("stylesheet", ".css"):
+        head += f'<link rel="stylesheet" property="stylesheet" href="{webpath(cssfile.path)}">'
+    return head
+
+def list_scripts(scriptdirname, extension):
+    scripts_list = []
+    scripts_dir = os.path.join(shared.chuanhu_path, "web_assets", scriptdirname)
+    if os.path.exists(scripts_dir):
+        for filename in sorted(os.listdir(scripts_dir)):
+            scripts_list.append(ScriptFile(shared.assets_path, filename, os.path.join(scripts_dir, filename)))
+    scripts_list = [x for x in scripts_list if os.path.splitext(x.path)[1].lower() == extension and os.path.isfile(x.path)]
+    return scripts_list
+
+
+def reload_javascript():
+    js = javascript_html()
+    js += '<script async type="module" src="https://cdn.jsdelivr.net/npm/marked/marked.min.js"></script>'
+    js += '<script async type="module" src="http://spin.js.org/spin.umd.js"></script><link type="text/css" href="https://spin.js.org/spin.css" rel="stylesheet" />'
+    
+    css = css_html()
+
+    def template_response(*args, **kwargs):
+        res = GradioTemplateResponseOriginal(*args, **kwargs)
+        res.body = res.body.replace(b'</head>', f'{js}</head>'.encode("utf8"))
+        res.body = res.body.replace(b'</body>', f'{css}</body>'.encode("utf8"))
+        res.init_headers()
+        return res
+
+    gr.routes.templates.TemplateResponse = template_response
+
+GradioTemplateResponseOriginal = gr.routes.templates.TemplateResponse

readme/README_en.md

@@ -6,7 +6,7 @@
 <h1 align="center">川虎 Chat 🐯 Chuanhu Chat</h1>
 <div align="center">
   <a href="https://github.com/GaiZhenBiao/ChuanhuChatGPT">
-    <img src="https://user-images.githubusercontent.com/70903329/227087087-93b37d64-7dc3-4738-a518-c1cf05591c8a.png" alt="Logo" height="156">
+    <img src="https://github.com/GaiZhenbiao/ChuanhuChatGPT/assets/70903329/aca3a7ec-4f1d-4667-890c-a6f47bf08f63" alt="Logo" height="156">
   </a>
 
 <p align="center">
@@ -44,6 +44,23 @@
   </p>
 </div>
 
+## Supported LLM Models
+
+**LLM models via API**:
+
+- [ChatGPT](https://chat.openai.com) ([GPT-4](https://openai.com/product/gpt-4))
+- [Google PaLM](https://developers.generativeai.google/products/palm)
+- [Inspur Yuan 1.0](https://air.inspur.com/home)
+- [MiniMax](https://api.minimax.chat/)
+- [XMChat](https://github.com/MILVLG/xmchat)
+
+**LLM models via local deployment**:
+
+- [ChatGLM](https://github.com/THUDM/ChatGLM-6B) ([ChatGLM2](https://github.com/THUDM/ChatGLM2-6B))
+- [LLaMA](https://github.com/facebookresearch/llama)
+- [StableLM](https://github.com/Stability-AI/StableLM)
+- [MOSS](https://github.com/OpenLMLab/MOSS)
+
 ## Usage Tips
 
 - To better control the ChatGPT, use System Prompt.
@@ -51,11 +68,11 @@
 - To try again if the response is unsatisfactory, use `🔄 Regenerate` button.
 - To start a new line in the input box, press <kbd>Shift</kbd> + <kbd>Enter</kbd> keys.
 - To quickly switch between input history, press <kbd>↑</kbd> and <kbd>↓</kbd> key in the input box.
-- To deploy the program onto a server, change the last line of the program to `demo.launch(server_name="0.0.0.0", server_port=<your port number>)`.
-- To get a public shared link, change the last line of the program to `demo.launch(share=True)`. Please be noted that the program must be running in order to be accessed via a public link.
+- To deploy the program onto a server, set `"server_name": "0.0.0.0", "server_port" <your port number>,` in `config.json`.
+- To get a public shared link, set `"share": true,` in `config.json`. Please be noted that the program must be running in order to be accessed via a public link.
 - To use it in Hugging Face Spaces: It is recommended to **Duplicate Space** and run the program in your own Space for a faster and more secure experience.
 
-## Installation
+## Quickstart
 
 ```shell
 git clone https://github.com/GaiZhenbiao/ChuanhuChatGPT.git
@@ -87,10 +104,6 @@ When you encounter problems, you should try manually pulling the latest changes
    ```
    pip install -r requirements.txt
    ```
-3. Update Gradio
-   ```
-   pip install gradio --upgrade --force-reinstall
-   ```
 
 Generally, you can solve most problems by following these steps.
 

readme/README_ja.md

@@ -6,7 +6,7 @@
 <h1 align="center">川虎 Chat 🐯 Chuanhu Chat</h1>
 <div align="center">
   <a href="https://github.com/GaiZhenBiao/ChuanhuChatGPT">
-    <img src="https://user-images.githubusercontent.com/70903329/227087087-93b37d64-7dc3-4738-a518-c1cf05591c8a.png" alt="Logo" height="156">
+    <img src="https://github.com/GaiZhenbiao/ChuanhuChatGPT/assets/70903329/aca3a7ec-4f1d-4667-890c-a6f47bf08f63" alt="Logo" height="156">
   </a>
 
 <p align="center">
@@ -44,17 +44,34 @@
   </p>
 </div>
 
+## サポートされている大規模言語モデル
+
+**APIを通じてアクセス可能な大規模言語モデル**：
+
+- [ChatGPT](https://chat.openai.com) ([GPT-4](https://openai.com/product/gpt-4))
+- [Google PaLM](https://developers.generativeai.google/products/palm)
+- [Inspur Yuan 1.0](https://air.inspur.com/home)
+- [MiniMax](https://api.minimax.chat/)
+- [XMChat](https://github.com/MILVLG/xmchat)
+
+**ローカルに展開された大規模言語モデル**：
+
+- [ChatGLM](https://github.com/THUDM/ChatGLM-6B) ([ChatGLM2](https://github.com/THUDM/ChatGLM2-6B))
+- [LLaMA](https://github.com/facebookresearch/llama)
+- [StableLM](https://github.com/Stability-AI/StableLM)
+- [MOSS](https://github.com/OpenLMLab/MOSS)
+
 ## 使う上でのTips
 
 - ChatGPTをより適切に制御するために、システムプロンプトを使用できます。
 - プロンプトテンプレートを使用するには、プロンプトテンプレートコレクションを選択し、ドロップダウンメニューから特定のプロンプトを選択。回答が不十分な場合は、`🔄再生成`ボタンを使って再試行します。
 - 入力ボックスで改行するには、<kbd>Shift</kbd> + <kbd>Enter</kbd>キーを押してください。
 - 入力履歴を素早く切り替えるには、入力ボックスで <kbd>↑</kbd>と<kbd>↓</kbd>キーを押す。
-- プログラムをサーバにデプロイするには、プログラムの最終行を `demo.launch(server_name="0.0.0.0", server_port=<your port number>)`に変更します。
-- 共有リンクを取得するには、プログラムの最後の行を `demo.launch(share=True)` に変更してください。なお、公開リンクでアクセスするためには、プログラムが実行されている必要があることに注意してください。
+- プログラムをサーバーに展開するには、`config.json` 内の `"server_name": "0.0.0.0", "server_port": <ポート番号>`を設定してください。
+- 共有リンクを取得するには、 `config.json` 内の `"share": true` を設定してください。なお、公開リンクでアクセスするためには、プログラムが実行されている必要があることに注意してください。
 - Hugging Face Spacesで使用する場合： より速く、より安全に利用するために、**Duplicate Space**を使用し、自分のスペースでプログラムを実行することをお勧めします。
 
-## インストール
+## クイックスタート
 
 ```shell
 git clone https://github.com/GaiZhenbiao/ChuanhuChatGPT.git
@@ -86,10 +103,6 @@ python ChuanhuChatbot.py
    ```
    pip install -r requirements.txt
    ```
-3. Gradioを更新
-   ```
-   pip install gradio --upgrade --force-reinstall
-   ```
 
 一般的に、以下の手順でほとんどの問題を解決することができます。
 

requirements.txt

@@ -1,5 +1,5 @@
-gradio==3.33.1
-gradio_client==0.2.5
+gradio==3.40.0
+gradio_client==0.4.0
 pypinyin
 tiktoken
 socksio
@@ -7,7 +7,7 @@ tqdm
 colorama
 googlesearch-python
 Pygments
-langchain==0.0.173
+langchain==0.0.276
 markdown
 PyPDF2
 pdfplumber
@@ -16,10 +16,13 @@ commentjson
 openpyxl
 pandoc
 wolframalpha
-faiss-cpu
+faiss-cpu==1.7.4
 duckduckgo-search
 arxiv
 wikipedia
 google.generativeai
-openai
+openai>=0.27.9
 unstructured
+google-api-python-client
+tabulate
+ujson

requirements_advanced.txt

@@ -0,0 +1,11 @@
+transformers
+huggingface_hub
+torch
+icetk
+protobuf==3.19.0
+git+https://github.com/OptimalScale/LMFlow.git
+cpm-kernels
+sentence_transformers
+accelerate
+sentencepiece
+datasets