Method initialize

src/benchmark/leetcode.py:802–893

Initialize the benchmark by loading dataset, filtering finished tasks, and starting browser.

(self)

        self._queue_lock = asyncio.Lock()  # 初始化队列锁

    async def initialize(self):
        """Initialize the benchmark by loading dataset, filtering finished tasks, and starting browser."""
        try:
            from src.data.leetcode import LeetCodeDataset
            # 1. 加载数据集
            dataset = LeetCodeDataset(
                path=self.path,
                split=self.split,
                name=self.subset if self.subset else None
            )
            self._id_to_record_map = {}
            
            # 获取原始数据记录
            if hasattr(dataset, 'data'):
                self._data_records = dataset.data.to_dict(orient="records")
            else:
                self._data_records = []

            # ================= [新增逻辑：断点续跑过滤] =================
            
            # A. 动态定位结果文件路径 (逻辑需与 CodeSubmitter 保持完全一致)
            current_file_path = os.path.abspath(__file__)
            project_root_name = "AgentWorld"
            
            if project_root_name in current_file_path:
                root_path = current_file_path.split(project_root_name)[0] + project_root_name
                output_dir = os.path.join(
                    root_path, 
                    "workdir", "tool_calling_agent", "benchmark", "leetcode"
                )
            else:
                output_dir = os.path.join(os.getcwd(), "results")
            
            # 固定文件名，不再带时间戳
            result_file = os.path.join(output_dir, "results.jsonl")

            # B. 读取已完成的任务 ID
            finished_ids = set()
            if os.path.exists(result_file):
                logger.info(f"| 🔄 Found existing result file: {result_file}, checking for finished tasks...")
                try:
                    with open(result_file, 'r', encoding='utf-8') as f:
                        for line in f:
                            line = line.strip()
                            if not line: continue
                            try:
                                record = json.loads(line)
                                if "task_id" in record:
                                    finished_ids.add(str(record["task_id"]))
                            except json.JSONDecodeError:
                                continue
                except Exception as e:
                    logger.warning(f"| ⚠️ Error reading result file for filtering: {e}")
            
            # C. 过滤 self._data_records
            original_count = len(self._data_records)
            self._data_records = [
                r for r in self._data_records