backend/api/browsecomp.py

from flask import Blueprint, jsonify
from datasets import load_dataset
import json

bp = Blueprint("browsecomp", __name__, url_prefix="/api/browsecomp")

HF_REPO = "timchen0618/browsecomp-plus-benchmark"

_cache: list | None = None


def _load() -> list:
    global _cache
    if _cache is not None:
        return _cache
    ds = load_dataset(HF_REPO, split="train")
    rows = []
    for row in ds:
        rows.append({
            "query_id":     row["query_id"],
            "query":        row["query"],
            "answer":       row["answer"],
            "evidence_docs": json.loads(row["evidence_docs"]),
            "gold_docs":    json.loads(row["gold_docs"]),
        })
    _cache = rows
    return rows


@bp.get("/")
def get_list():
    try:
        rows = _load()
        # Return lightweight list (just query_id + query preview)
        items = [{"query_id": r["query_id"], "query": r["query"]} for r in rows]
        return jsonify({"items": items, "total": len(items)})
    except Exception as e:
        return jsonify({"error": str(e)}), 500


@bp.get("/<query_id>")
def get_example(query_id: str):
    try:
        rows = _load()
        for row in rows:
            if str(row["query_id"]) == query_id:
                return jsonify(row)
        return jsonify({"error": "not found"}), 404
    except Exception as e:
        return jsonify({"error": str(e)}), 500


@bp.post("/reload")
def reload():
    global _cache
    _cache = None
    try:
        rows = _load()
        return jsonify({"status": "ok", "count": len(rows)})
    except Exception as e:
        return jsonify({"error": str(e)}), 500