app.py

import json
import sseclient
import requests
from flask import Flask, request, Response, stream_with_context
import random

app = Flask(__name__)

def generate_random_ip():
    return f"{random.randint(1,255)}.{random.randint(0,255)}.{random.randint(0,255)}.{random.randint(0,255)}"

def format_openai_response(content, finish_reason=None):
    return {
        "id": "chatcmpl-123",
        "object": "chat.completion.chunk",
        "created": 1677652288,
        "model": "gpt-4o",
        "choices": [{
            "delta": {"content": content} if content else {"finish_reason": finish_reason},
            "index": 0,
            "finish_reason": finish_reason
        }]
    }

@app.route('/hf/v1/chat/completions', methods=['POST'])
def chat_completions():
    data = request.json
    messages = data.get('messages', [])
    stream = data.get('stream', False)
    
    if not messages:
        return {"error": "No messages provided"}, 400
    
    model = data.get('model', 'gpt-4o')

    if model.startswith('gpt'):
        endpoint = "openAI"
        original_api_url = 'https://chatpro.ai-pro.org/api/ask/openAI'
    elif model.startswith('claude'):
        endpoint = "claude"
        original_api_url = 'https://chatpro.ai-pro.org/api/ask/claude'
    else:
        return {"error": "Unsupported model"}, 400

    headers = {
        'content-type': 'application/json',
        'X-Forwarded-For': generate_random_ip(),
        'origin': 'https://chatpro.ai-pro.org',
        'user-agent': generate_user_agent()
    }

    def generate():
        nonlocal messages
        full_response = ""
        while True:
            conversation = "\n".join([f"{msg['role']}: {msg['content']}" for msg in messages])
            conversation += "\n请关注并回复user最近的消息并避免总结对话历史的回答"
            
            payload = {
                "text": conversation,
                "endpoint": endpoint,
                "model": model
            }
            
            response = requests.post(original_api_url, headers=headers, json=payload, stream=True)
            client = sseclient.SSEClient(response)
            
            for event in client.events():
                if event.data.startswith('{"text":'):
                    data = json.loads(event.data)
                    new_content = data['text'][len(full_response):]
                    full_response = data['text']
                    
                    if new_content:
                        yield f"data: {json.dumps(format_openai_response(new_content))}\n\n"
                
                elif '"final":true' in event.data:
                    final_data = json.loads(event.data)
                    finish_reason = final_data.get('responseMessage', {}).get('finish_reason', 'stop')
                    if finish_reason == 'length':
                        # 如果因为长度被截断，添加已生成的回复到消息列表，并继续生成
                        messages.append({"role": "assistant", "content": full_response})
                        messages.append({"role": "user", "content": "请继续你的输出"})
                        break
                    else:
                        yield f"data: {json.dumps(format_openai_response('', finish_reason))}\n\n"
                        yield "data: [DONE]\n\n"
                        return

    if stream:
        return Response(stream_with_context(generate()), content_type='text/event-stream')
    else:
        full_response = ""
        finish_reason = "stop"
        for chunk in generate():
            if chunk.startswith("data: ") and not chunk.strip() == "data: [DONE]":
                response_data = json.loads(chunk[6:])
                if 'choices' in response_data and response_data['choices']:
                    delta = response_data['choices'][0].get('delta', {})
                    if 'content' in delta:
                        full_response += delta['content']
                    if 'finish_reason' in delta:
                        finish_reason = delta['finish_reason']

        return {
            "id": "chatcmpl-123",
            "object": "chat.completion",
            "created": 1677652288,
            "model": model,
            "choices": [{
                "index": 0,
                "message": {
                    "role": "assistant",
                    "content": full_response
                },
                "finish_reason": finish_reason
            }],
            "usage": {
                "prompt_tokens": 0,
                "completion_tokens": 0,
                "total_tokens": 0
            }
        }

if __name__ == '__main__':
    app.run(debug=True, port=5000)