#!/usr/bin/env python
from __future__ import annotations

import argparse
import sys
from pathlib import Path

ROOT = Path(__file__).resolve().parents[1]
if str(ROOT) not in sys.path:
    sys.path.insert(0, str(ROOT))

import numpy as np

from spec_rag.faiss_index import load_index
from spec_rag.io import load_embeddings, load_smiles, save_jsonl
from spec_rag.retrieval import build_smiles_to_index, recall_at_k, search_index


def parse_args() -> argparse.Namespace:
    parser = argparse.ArgumentParser(description="Retrieve SMILES from FAISS index.")
    parser.add_argument("--index-path", required=True)
    parser.add_argument("--smiles-path", required=True)
    parser.add_argument("--query-embeddings", required=True)
    parser.add_argument("--out-jsonl", required=True)
    parser.add_argument("--k", type=int, default=100)
    parser.add_argument("--ground-truth-smiles", default=None)
    return parser.parse_args()


def main() -> None:
    args = parse_args()
    index = load_index(args.index_path)
    smiles = load_smiles(args.smiles_path)
    queries = load_embeddings(args.query_embeddings)
    scores, indices = search_index(index, queries, args.k)

    rows = []
    for i in range(indices.shape[0]):
        retrieved = [smiles[j] if j >= 0 else "" for j in indices[i].tolist()]
        rows.append(
            {
                "query_id": i,
                "indices": indices[i].tolist(),
                "scores": scores[i].tolist(),
                "smiles": retrieved,
            }
        )
    save_jsonl(args.out_jsonl, rows)

    if args.ground_truth_smiles:
        gt_smiles = load_smiles(args.ground_truth_smiles)
        lookup = build_smiles_to_index(smiles)
        gt_indices = [lookup.get(smi, -1) for smi in gt_smiles]
        if not any(i >= 0 for i in gt_indices):
            print("No ground-truth SMILES found in index.")
            return
        recall = recall_at_k(indices, gt_indices, args.k)
        print(f"Recall@{args.k}: {recall:.4f}")


if __name__ == "__main__":
    main()