from llama_index.core import VectorStoreIndex, Settings
from llama_index.core.query_engine import RetrieverQueryEngine
from llama_index.core.retrievers import VectorIndexRetriever
from llama_index.core.response_synthesizers import get_response_synthesizer, ResponseMode
from llama_index.core.prompts import PromptTemplate
from llama_index.retrievers.bm25 import BM25Retriever
from llama_index.core.retrievers import QueryFusionRetriever
from my_logging import log_message
from config import CUSTOM_PROMPT, PROMPT_SIMPLE_POISK

def create_vector_index(documents):
    log_message("Строю векторный индекс")
    return VectorStoreIndex.from_documents(documents)

def rerank_nodes(query, nodes, reranker, top_k=25, min_score_threshold=0.5):
    if not nodes or not reranker:
        return nodes[:top_k]
    
    try:
        log_message(f"Переранжирую {len(nodes)} узлов")
        
        pairs = [[query, node.text] for node in nodes]
        scores = reranker.predict(pairs)
        scored_nodes = list(zip(nodes, scores))
        
        scored_nodes.sort(key=lambda x: x[1], reverse=True)
        
        # Apply threshold
        filtered = [(node, score) for node, score in scored_nodes if score >= min_score_threshold]
        
        if not filtered:
            # Lower threshold if nothing passes
            filtered = scored_nodes[:top_k]
        
        log_message(f"Выбрано {min(len(filtered), top_k)} узлов")
        
        return [node for node, score in filtered[:top_k]]
        
    except Exception as e:
        log_message(f"Ошибка переранжировки: {str(e)}")
        return nodes[:top_k]

def create_query_engine(vector_index):
    try:
        from config import CUSTOM_PROMPT
        
        bm25_retriever = BM25Retriever.from_defaults(
            docstore=vector_index.docstore, 
            similarity_top_k=70 
        )
        
        vector_retriever = VectorIndexRetriever(
            index=vector_index, 
            similarity_top_k=70,
            similarity_cutoff=0.55
        )
        
        hybrid_retriever = QueryFusionRetriever(
            [vector_retriever, bm25_retriever],
            similarity_top_k=50,
            num_queries=1
        )
        
        custom_prompt_template = PromptTemplate(CUSTOM_PROMPT)
        response_synthesizer = get_response_synthesizer(
            response_mode=ResponseMode.TREE_SUMMARIZE,
            text_qa_template=custom_prompt_template
        )
        
        query_engine = RetrieverQueryEngine(
            retriever=hybrid_retriever,
            response_synthesizer=response_synthesizer
        )
        
        log_message("Query engine успешно создан")
        return query_engine
        
    except Exception as e:
        log_message(f"Ошибка создания query engine: {str(e)}")
        raise