# pip install rank-bm25
from rank_bm25 import BM25Okapi
import jieba    # pip install jieba  中文分词

class HybridRetriever:
    def __init__(self, docs: list[str]):
        self.docs = docs
        # BM25 用分词后的词列表
        tokenized = [list(jieba.cut(d)) for d in docs]
        self.bm25 = BM25Okapi(tokenized)

    def bm25_search(self, query: str, top_k: int = 5) -> list[str]:
        tokens = list(jieba.cut(query))
        scores = self.bm25.get_scores(tokens)
        ranked = sorted(enumerate(scores), key=lambda x: x[1], reverse=True)
        return [self.docs[i] for i, _ in ranked[:top_k]]

    def hybrid_search(self, query: str, top_k: int = 3) -> list[str]:
        # 并发跑两路检索
        vec_results = retrieve(query, top_k=5)    # pgvector 向量检索
        bm25_results = self.bm25_search(query, top_k=5)

        # RRF 融合
        scores: dict[str, float] = {}
        for rank, doc in enumerate(vec_results):
            scores[doc] = scores.get(doc, 0) + 1 / (60 + rank)
        for rank, doc in enumerate(bm25_results):
            scores[doc] = scores.get(doc, 0) + 1 / (60 + rank)

        merged = sorted(scores, key=lambda d: scores[d], reverse=True)
        return merged[:top_k]

# 替换 rag_query 里的 retrieve
retriever = HybridRetriever(all_docs)
contexts = retriever.hybrid_search("VDOT 88 配速")