INNER CODE UNIT · Python
bixbench
Future-House/BixBench · bixbench/generate_trajectories.py:142
bixbench = datasets.load_dataset(self.config.paths.hf_repo_id, split=self.config.dataset_split).to_list() # type: ignore[attr-defined]
# Process all capsule data concurrently
zip_filenames = {question["data_folder"] for question in bixbench}
tasks = [self.process_capsule_data(zip_fname) for zip_fname in zip_filenames]
await asyncio.gather(*tasks)
# Prepare all questions concurrently
tasks = [self.process_question(question) for question in bixbench]
await asyncio.gather(*tasks)
return bixbench
def _extract_and_process_files(self, zip_path: Path, extract_dir: Path) -> None:
"""
Extract and process zip files for a capsule.
Args: