Chainlit接入FastGpt接口完美对接，实现全新的用户聊天界面

前言

由于fastgpt只提供了一个分享用的网页应用，网页访问地址没法自定义，虽然可以接入NextWeb/ChatGPT web等开源应用。但是如果我们想直接给客户应用，还需要客户去设置配置，里面还有很多我们不想展示给客户的东西怎么办？于是，我使用Chainlit实现了一个无缝快速接入fastgpt实现自定义用户使用界面的应用，代码清晰简单。还可以自定义logo、欢迎语、网站图标等。
之前写一个一篇文章《Chainlit接入FastGpt接口快速实现自定义用户聊天界面》文章中实现了Fastgpt的对话接口，可以实现聊天，但是有些人想。接入Fastgpt的欢迎语，获取聊天历史记录等。本此的代码更新可以实现这些，下面我主要讲一些这次代码更新实现的亮点吧！

1. 接入Fastgpt后，可以自动更新成你的网站图标和网站名称

在这里插入图片描述

2. 接入欢迎语

在这里插入图片描述

3. 接入聊天记录

在这里插入图片描述

不仅可以实现聊天记录的自动添加还可实现删除，搜索聊天记录的功能，原fastgpt的分享聊天网页可没有聊天搜索功能呦！

4 可以开启黑白主题的切换

在这里插入图片描述

5.可以自定义修改访问路径

当然这个功能需要你自己修改启动地址

快速开始

获取Fastgpt的信息

获取fastgpt的base_url和share_id

复制base_url和share_id，后面需要配置到Chainlit的环境变量中

在这里插入图片描述

获取fastgpt的API_KEY

只需要复制API_KEY即可，后面需要配置到Chainlit的环境变量中

Chainlit网页搭建

创建一个文件夹，例如“chainlit_fastgpt”

mkdir chainlit_fastgpt

进入 chainlit_chat文件夹下，执行命令创建python 虚拟环境空间(需要提前安装好python sdk。 Chainlit 需要python>=3.8。,具体操作，由于文章长度问题就不在叙述，自行百度)，命令如下：

python -m venv .venv

这一步是避免python第三方库冲突，省事版可以跳过
.venv是创建的虚拟空间文件夹可以自定义

接下来激活你创建虚拟空间，命令如下：

#linux or mac
source .venv/bin/activate
#windows
.venv\Scripts\activate

在项目根目录下创建`requirements.txt`，内容如下：

chainlit~=1.1.402
aiohttp~=3.10.5
requests~=2.32.3
literalai~=0.0.607

在项目根目录下创建app.py文件，代码如下：

import hashlib
import json
import os
from typing import Optional, Dictimport aiohttp
import chainlit as cl
import chainlit.data as cl_data
import requestsfrom fastgpt_data import FastgptDataLayer, now, share_id, app_name, welcome_textfastgpt_base_url = os.getenv("FASTGPT_BASE_URL")
fastgpt_api_key = os.getenv("FASTGPT_API_KEY")
fastgpt_api_detail = os.getenv("FASTGPT_API_DETAIL", False)cl_data._data_layer = FastgptDataLayer()
cl.config.ui.name = app_namedef download_logo():local_filename = "./public/favicon.svg"directory = os.path.dirname(local_filename)os.makedirs(directory, exist_ok=True)# Streaming, so we can iterate over the response.with requests.get(f"{fastgpt_base_url}/icon/logo.svg", stream=True) as r:r.raise_for_status()  # Check if the request was successfulwith open(local_filename, 'wb') as f:for chunk in r.iter_content(chunk_size=8192):# If you have chunk encoded response uncomment if# and set chunk_size parameter to None.f.write(chunk)download_logo()@cl.on_chat_start
async def chat_start():if welcome_text:# elements = [cl.Text(content=welcomeText, display="inline")]await cl.Message(content=welcome_text).send()@cl.on_message
async def handle_message(message: cl.Message):msg = cl.Message(content="")url = f"{fastgpt_base_url}/api/v1/chat/completions"print('message.thread_id',message.thread_id)headers = {"Authorization": f"Bearer {fastgpt_api_key}","Content-Type": "application/json"}data = {"messages": [{"role": "user","content": message.content}],"variables": {"cTime": now},"responseChatItemId": message.id,"shareId": share_id,"chatId": message.thread_id,"appType": "advanced","outLinkUid": cl.context.session.user.identifier,"detail": fastgpt_api_detail,"stream": True}async for data in fetch_sse(url, headers=headers, data=json.dumps(data), detail=fastgpt_api_detail):delta = data['choices'][0]['delta']if delta:await msg.stream_token(delta['content'])await msg.send()@cl.header_auth_callback
def header_auth_callback(headers: Dict) -> Optional[cl.User]:print(headers)# 创建一个md5 hash对象md5_hash = hashlib.md5()user_agent_bytes = headers.get('user-agent').encode('utf-8')# 更新这个hash对象的内容md5_hash.update(user_agent_bytes)# 获取md5哈希值的十六进制表示形式md5_hex_digest = md5_hash.hexdigest()out_link_uid = md5_hex_digestprint("MD5加密后的结果:", md5_hex_digest)return cl.User(identifier=out_link_uid, display_name="visitor")@cl.on_chat_resume
async def on_chat_resume():passasync def fetch_sse(url, headers, data, detail):async with aiohttp.ClientSession() as session:async with session.post(url, headers=headers, data=data) as response:async for line in response.content:if line:  # 过滤掉空行data = line.decode('utf-8').rstrip('\n\r')# print(f"Received: {data}")# 检查是否为数据行，并且是我们感兴趣的事件类型if detail:if data.startswith('event:'):event_type = data.split(':', 1)[1].strip()  # 提取事件类型elif data.startswith('data:') and event_type == 'flowNodeStatus':data = data.split(':', 1)[1].strip()flowNodeStatus = json.loads(data)current_step = cl.context.current_stepcurrent_step.name = flowNodeStatus['name']elif data.startswith('data:') and event_type == 'answer':data = data.split(':', 1)[1].strip()  # 提取数据内容# 如果数据包含换行符，可能需要进一步处理（这取决于你的具体需求）# 这里我们简单地打印出来if data != '[DONE]':yield json.loads(data)else:if data.startswith('data:'):data = data.split(':', 1)[1].strip()  # 提取数据内容# 如果数据包含换行符，可能需要进一步处理（这取决于你的具体需求）# 这里我们简单地打印出来if data != '[DONE]':yield json.loads(data)

传入的model，temperature等参数字段均无效，这些字段由编排决定，不会根据 API 参数改变。
不会返回实际消耗Token值，如果需要，可以设置detail=true，并手动计算 responseData 里的tokens值。

在项目根目录下创建fastgpt_data.py文件，代码如下：

import json
import os
import uuid
from typing import Optional, List, Dictimport requests
from chainlit import PersistedUser
from chainlit.data import BaseDataLayer
from chainlit.types import PageInfo, ThreadFilter, ThreadDict, Pagination, PaginatedResponse
from literalai.helper import utc_nowfastgpt_base_url = os.getenv("FASTGPT_BASE_URL")
share_id = os.getenv("FASTGPT_SHARE_ID")
now = utc_now()
user_cur_threads = []
thread_user_dict = {}def change_type(user_type: str):if user_type == 'AI':return 'assistant_message'if user_type == 'Human':return 'user_message'def get_app_info():with requests.get(f"{fastgpt_base_url}/api/core/chat/outLink/init?chatId=&shareId={share_id}&outLinkUid=123456") as resp:app = {}if resp.status_code == 200:res = json.loads(resp.content)app = res.get('data').get('app')appId = res.get('data').get('appId')app['id'] = appIdreturn appapp_info = get_app_info()app_id = app_info.get('id')
app_name = app_info.get('name')
welcome_text = app_info.get('chatConfig').get('welcomeText')def getHistories(user_id):histories = []if user_id:with requests.post(f"{fastgpt_base_url}/api/core/chat/getHistories",data={"shareId": share_id, "outLinkUid": user_id}) as resp:if resp.status_code == 200:res = json.loads(resp.content)data = res["data"]print(data)histories = [{"id": item["chatId"],"name": item["title"],"createdAt": item["updateTime"],"userId": user_id,"userIdentifier": user_id}for item in data]if user_cur_threads:thread = next((t for t in user_cur_threads if t["userId"] == user_id), None)if thread:  # 确保 thread 不为 Nonethread_id = thread.get("id")if histories:# 检查 thread 的 ID 是否已存在于 threads 中if not any(t.get("id") == thread_id for t in histories):histories.insert(0, thread)else:# 如果 threads 是空列表，则直接插入 threadhistories.insert(0, thread)for item in histories:thread_user_dict[item.get('id')] = item.get('userId')return historiesclass FastgptDataLayer(BaseDataLayer):async def get_user(self, identifier: str):print('get_user', identifier)return PersistedUser(id=identifier, createdAt=now, identifier=identifier)async def update_thread(self,thread_id: str,name: Optional[str] = None,user_id: Optional[str] = None,metadata: Optional[Dict] = None,tags: Optional[List[str]] = None,):print('---------update_thread----------',thread_id)thread = next((t for t in user_cur_threads if t["userId"] == user_id), None)if thread:if thread_id:thread["id"] = thread_idif name:thread["name"] = nameif user_id:thread["userId"] = user_idthread["userIdentifier"] = user_idif metadata:thread["metadata"] = metadataif tags:thread["tags"] = tagsthread["createdAt"] = utc_now()else:print('---------update_thread----------thread_id ', thread_id, name)user_cur_threads.append({"id": thread_id,"name": name,"metadata": metadata,"tags": tags,"createdAt": utc_now(),"userId": user_id,"userIdentifier": user_id,})async def get_thread_author(self, thread_id: str):print('get_thread_author')return thread_user_dict.get(thread_id, None)async def list_threads(self, pagination: Pagination, filters: ThreadFilter) -> PaginatedResponse[ThreadDict]:threads = []if filters:threads = getHistories(filters.userId)search = filters.search if filters.search else ""filtered_threads = [thread for thread in threads if search in thread.get('name', '')]start = 0if pagination.cursor:for i, thread in enumerate(filtered_threads):if thread["id"] == pagination.cursor:  # Find the start index using pagination.cursorstart = i + 1breakend = start + pagination.firstpaginated_threads = filtered_threads[start:end] or []has_next_page = len(paginated_threads) > endstart_cursor = paginated_threads[0]["id"] if paginated_threads else Noneend_cursor = paginated_threads[-1]["id"] if paginated_threads else Nonereturn PaginatedResponse(pageInfo=PageInfo(hasNextPage=has_next_page,startCursor=start_cursor,endCursor=end_cursor,),data=paginated_threads,)async def get_thread(self, thread_id: str):print('get_thread', thread_id)user_id = thread_user_dict.get(thread_id, None)thread = Noneif user_id:params = {'chatId': thread_id,'shareId': share_id,'outLinkUid': user_id,}with requests.get(f"{fastgpt_base_url}/api/core/chat/outLink/init",params=params,) as resp:if resp.status_code == 200:res = json.loads(resp.content)data = res["data"]if data:history = data['history']files = []texts = []for item in history:for entry in item['value']:if entry.get('type') == 'text':text = {"id": item["_id"],"threadId": thread_id,"name": item["obj"],"type": change_type(item["obj"]),"input": None,"createdAt": utc_now(),"output": entry.get('text').get('content'),}texts.append(text)if entry.get('type') == 'file':file = {"id": str(uuid.UUID),"threadId": thread_id,"forId": item["_id"],"name": entry.get('file').get('name'),"type": entry.get('file').get('type'),"url": entry.get('file').get('url'),"display": "inline","size": "medium"}files.append(file)thread = {"id": thread_id,"name": data.get("title", ''),"createdAt": utc_now(),"userId": "admin","userIdentifier": "admin","metadata": {"appId": data["appId"]},"steps": texts,"elements": files,}return threadreturn threadasync def delete_thread(self, thread_id: str):print('delete_thread')thread = next((t for t in user_cur_threads if t["id"] == thread_id), None)user_id = thread_user_dict.get(thread_id, None)if thread:user_cur_threads.remove(thread)if user_id:params = {'appId': app_id,'chatId': thread_id,'shareId': share_id,'outLinkUid': user_id,}requests.get(f"{fastgpt_base_url}/api/core/chat/delHistory",params=params)

在项目根目录下创建`.env`环境变量，配置如下：

CHAINLIT_AUTH_SECRET="xOIPIMBGfI7N*VK6O~KOVIRC/cGRNSmk%bmO4Q@el647hR?^mdW6=8KlQBuWWTbk"
FASTGPT_BASE_URL="https://share.fastgpt.in"
FASTGPT_API_KEY="fastgpt-key"
FASTGPT_SHARE_ID=""
FASTGPT_API_DETAIL=False