utils/retrieval.py

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

class ChatQABot:
    def __init__(self, model_name='Qwen/Qwen1.5-1.8B-Chat'):
        self.tokenizer = AutoTokenizer.from_pretrained(
            model_name, 
            trust_remote_code=True
        )
        self.model = AutoModelForCausalLM.from_pretrained(
            model_name,
            trust_remote_code=True,
            torch_dtype=torch.float16,
            device_map="auto" if torch.cuda.is_available() else "cpu"
        )
    
    def generate_answer(self, query: str, context: str) -> str:
        """基于检索到的上下文生成答案"""
        prompt = f"""基于以下聊天记录回答问题：

聊天记录：
{context}

问题：{query}

请根据聊天记录准确回答，如果聊天记录中没有相关信息，请说"根据现有聊天记录无法回答这个问题"。回答要简洁准确。"""
        
        messages = [
            {"role": "system", "content": "你是一个专业的聊天记录分析助手。"},
            {"role": "user", "content": prompt}
        ]
        
        text = self.tokenizer.apply_chat_template(
            messages,
            tokenize=False,
            add_generation_prompt=True
        )
        
        inputs = self.tokenizer(text, return_tensors="pt").to(self.model.device)
        
        with torch.no_grad():
            outputs = self.model.generate(
                **inputs,
                max_new_tokens=200,
                temperature=0.7,
                do_sample=True,
                pad_token_id=self.tokenizer.eos_token_id
            )
        
        response = self.tokenizer.decode(outputs[0][inputs.input_ids.shape[1]:], skip_special_tokens=True)
        return response