Langchain-Chatchat/server/model_workers/zhipu.py

from contextlib import contextmanager

import httpx
import requests
from fastchat.conversation import Conversation
from httpx_sse import EventSource

from server.model_workers.base import *
from fastchat import conversation as conv
import sys
from typing import List, Dict, Iterator, Literal, Any
import jwt
import time
import json

@contextmanager
def connect_sse(client: httpx.Client, method: str, url: str, **kwargs: Any):
    with client.stream(method, url, **kwargs) as response:
        yield EventSource(response)


def generate_token(apikey: str, exp_seconds: int):
    try:
        id, secret = apikey.split(".")
    except Exception as e:
        raise Exception("invalid apikey", e)

    payload = {
        "api_key": id,
        "exp": int(round(time.time() * 1000)) + exp_seconds * 1000,
        "timestamp": int(round(time.time() * 1000)),
    }

    return jwt.encode(
        payload,
        secret,
        algorithm="HS256",
        headers={"alg": "HS256", "sign_type": "SIGN"},
    )


class ChatGLMWorker(ApiModelWorker):
    DEFAULT_EMBED_MODEL = "embedding-2"
    
    def __init__(
            self,
            *,
            model_names: List[str] = ("zhipu-api",),
            controller_addr: str = None,
            worker_addr: str = None,
            version: Literal["glm-4"] = "glm-4",
            **kwargs,
    ):
        kwargs.update(model_names=model_names, controller_addr=controller_addr, worker_addr=worker_addr)
        kwargs.setdefault("context_len", 4096)
        super().__init__(**kwargs)
        self.version = version

    def do_chat(self, params: ApiChatParams) -> Iterator[Dict]:
        params.load_config(self.model_names[0])
        token = generate_token(params.api_key, 60)
        headers = {
            "Content-Type": "application/json",
            "Authorization": f"Bearer {token}"
        }
        data = {
            "model": params.version,
            "messages": params.messages,
            "max_tokens": params.max_tokens,
            "temperature": params.temperature,
            "stream": True
        }

        url = "https://open.bigmodel.cn/api/paas/v4/chat/completions"
        text = ""
        response = requests.post(url, headers=headers, json=data, stream=True)
        for chunk in response.iter_lines():
            if  chunk:
                if chunk.startswith(b'data:'):
                    json_str = chunk.decode('utf-8')[6:]
                    try:
                        data = json.loads(json_str)
                        if 'finish_reason' in data and data.get('finish_reason') =="stop":
                            break
                        else:
                            msg = data['choices'][0]['delta']['content']
                            text += msg
                            yield {"error_code": 0, "text": text}
                    except json.JSONDecodeError as e:
                        pass


    def do_embeddings(self, params: ApiEmbeddingsParams) -> Dict:
        embed_model = params.embed_model or self.DEFAULT_EMBED_MODEL

        params.load_config(self.model_names[0])
        i = 0
        batch_size = 1
        result = []
        while i < len(params.texts):
            token = generate_token(params.api_key, 60)
            headers = {
                "Content-Type": "application/json",
                "Authorization": f"Bearer {token}"
            }
            data = {
                "model": embed_model,
                "input": "".join(params.texts[i: i + batch_size])
            }
            embedding_data = self.request_embedding_api(headers, data, 1)
            if embedding_data:
                result.append(embedding_data)
            i += batch_size
            print(f"请求{embed_model}接口处理第{i}块文本，返回embeddings: \n{embedding_data}")

        return {"code": 200, "data": result}

    # 请求接口，支持重试
    def request_embedding_api(self, headers, data, retry=0):
        response = ''
        try:
            url = "https://open.bigmodel.cn/api/paas/v4/embeddings"
            response = requests.post(url, headers=headers, json=data)
            ans = response.json()
            return ans["data"][0]["embedding"]
        except Exception as e:
            print(f"request_embedding_api error={e} \nresponse={response}")
            if retry > 0:
                return self.request_embedding_api(headers, data, retry - 1)
            else:
                return None

    def get_embeddings(self, params):
        print("embedding")
        print(params)

    def make_conv_template(self, conv_template: str = None, model_path: str = None) -> Conversation:
        return conv.Conversation(
            name=self.model_names[0],
            system_message="你是智谱AI小助手，请根据用户的提示来完成任务",
            messages=[],
            roles=["user", "assistant", "system"],
            sep="\n###",
            stop_str="###",
        )


if __name__ == "__main__":
    import uvicorn
    from server.utils import MakeFastAPIOffline
    from fastchat.serve.model_worker import app

    worker = ChatGLMWorker(
        controller_addr="http://127.0.0.1:20001",
        worker_addr="http://127.0.0.1:21001",
    )
    sys.modules["fastchat.serve.model_worker"].worker = worker
    MakeFastAPIOffline(app)
    uvicorn.run(app, port=21001)