import lancedb
import os
import gradio as gr
import numpy as np
from sentence_transformers import SentenceTransformer, CrossEncoder


db = lancedb.connect(".lancedb")

TABLE = db.open_table(os.getenv("TABLE_NAME"))
VECTOR_COLUMN = os.getenv("VECTOR_COLUMN", "vector")
TEXT_COLUMN = os.getenv("TEXT_COLUMN", "text")
BATCH_SIZE = int(os.getenv("BATCH_SIZE", 32))

retriever = SentenceTransformer(os.getenv("EMB_MODEL"))
reranker = CrossEncoder(os.getenv("RERANK_MODEL"), max_length=512)


def retrieve(query, n):
    query_vec = retriever.encode(query)
    try:
        documents = TABLE.search(query_vec, vector_column_name=VECTOR_COLUMN).limit(n).to_list()
        documents = [doc[TEXT_COLUMN] for doc in documents]

        return documents

    except Exception as e:
        raise gr.Error(str(e))

def rerank(query, documents, k):
    query_doc_pairs = [[query, doc] for doc in documents]
    similarity_scores = reranker.predict(query_doc_pairs)
    sim_scores_argsort = np.argsort(similarity_scores)[::-1]

    rerank_documents = []

    for idx in sim_scores_argsort[:k]:
        rerank_documents.append(documents[idx])
    return rerank_documents