rag-gateway/app/service/ragflow.py

import httpx
from typing import Union, Dict, List

from fastapi import HTTPException
from starlette import status

from app.config.config import settings
from app.utils.rsa_crypto import RagflowCrypto


class RagflowService:
    def __init__(self, base_url: str):
        self.base_url = base_url

    def _handle_response(self, response: httpx.Response) -> Union[Dict, List]:
        if response.status_code != 200:
            return {}

        data = response.json()
        ret_code = data.get("retcode")
        if ret_code == 401:
            raise HTTPException(
                status_code=status.HTTP_401_UNAUTHORIZED,
                detail="登录过期",
            )
        if ret_code != 0:
            return {}

        # 检查返回的数据类型
        if isinstance(data.get("data"), dict):
            return data.get("data", {})
        elif isinstance(data.get("data"), list):
            return data.get("data", [])
        else:
            return {}

    async def register(self, username: str, password: str):
        password = RagflowCrypto(settings.PUBLIC_KEY, settings.PRIVATE_KEY).encrypt(password)
        async with httpx.AsyncClient() as client:
            response = await client.post(
                f"{self.base_url}/v1/user/register",
                headers={'Content-Type': 'application/json'},
                json={"nickname": username, "email": f"{username}@example.com", "password": password}
            )
            if response.status_code != 200:
                raise Exception(f"Ragflow registration failed: {response.text}")

    async def login(self, username: str, password: str) -> str:
        password = RagflowCrypto(settings.PUBLIC_KEY, settings.PRIVATE_KEY).encrypt(password)
        async with httpx.AsyncClient() as client:
            response = await client.post(
                f"{self.base_url}/v1/user/login",
                headers={'Content-Type': 'application/json'},
                json={"email": f"{username}@example.com", "password": password}
            )
            if response.status_code != 200:
                raise Exception(f"Ragflow login failed: {response.text}")
            authorization = response.headers.get('Authorization')
            if not authorization:
                raise Exception("Authorization header not found in response")
            return authorization

    async def chat(self, token: str, chat_id: str, chat_history: list):
        data = {
            "conversation_id": chat_id,
            "messages": chat_history
        }

        print(f"send to ragflow chat: {data}")

        target_url = f"{self.base_url}/v1/conversation/completion"
        async with httpx.AsyncClient(timeout=300.0) as client:
            headers = {
                'Content-Type': 'application/json',
                'Authorization': token
            }
            async with client.stream("POST", target_url, json=data, headers=headers) as response:
                if response.status_code == 200:
                    try:
                        async for answer in response.aiter_text():
                            print(f"response of ragflow chat: {answer}")
                            yield answer
                    except GeneratorExit as e:
                        print(e)
                        return
                else:
                    yield f"Error: {response.status_code}"

    async def get_chat_sessions(self, token: str, dialog_id: str) -> list:
        url = f"{self.base_url}/v1/conversation/list?dialog_id={dialog_id}"
        headers = {"Authorization": token}
        async with httpx.AsyncClient() as client:
            response = await client.get(url, headers=headers)
            data = self._handle_response(response)
            result = [
                {
                    "id": item["id"],
                    "name": item["name"],
                    "updated_time": item["update_time"]
                }
                for item in data
            ]
            return result

    async def set_session(self, token: str, dialog_id: str, message: dict, chat_id: str, is_new: bool) -> list:
        url = f"{self.base_url}/v1/conversation/set?dialog_id={dialog_id}"
        headers = {"Authorization": token}
        data = {
            "dialog_id": dialog_id,
            "name": message["message"],
            "is_new": is_new,
            "conversation_id": chat_id,
        }
        async with httpx.AsyncClient() as client:
            response = await client.post(url, headers=headers, json=data)
            data = self._handle_response(response)
            return [
                {
                    "content": "你好！ 我是你的助理，有什么可以帮到你的吗？",
                    "role": "assistant"
                },
                {
                    "content": message["message"],
                    "doc_ids":message.get("doc_ids", []),
                    "role": "user"
                }
            ] if data else []

    async def get_session_history(self, token: str, chat_id: str) -> list:
        url = f"{self.base_url}/v1/conversation/get?conversation_id={chat_id}"
        headers = {"Authorization": token}
        async with httpx.AsyncClient() as client:
            response = await client.get(url, headers=headers)
            data = self._handle_response(response)
            return data.get("message", [])

    async def upload_and_parse(self, token: str, chat_id: str, filename: str, file: bytes) -> str:
        url = f"{self.base_url}/v1/document/upload_and_parse"
        headers = {"Authorization": token}
        data = {"conversation_id": chat_id}

        # 创建表单数据，包含文件
        files = {"file": (filename, file)}
        async with httpx.AsyncClient(timeout=60) as client:
            response = await client.post(url, headers=headers, files=files, data=data)
            data = self._handle_response(response)
            return data