INNER CODE UNIT · Python
_read_parquet_files
paperswithbacktest/pwb-toolbox · pwb_toolbox/datasets/__init__.py:61
def _read_parquet_files(paths: list[str], desc: str, symbols=None) -> pd.DataFrame:
if not paths:
raise ValueError("No parquet files provided.")
iterator = tqdm(paths, desc=desc, unit="file") if len(paths) > 1 else paths
frames = []
for path in iterator:
if isinstance(path, str) and path.startswith(("http://", "https://")):
resp = requests.get(path, timeout=120)
resp.raise_for_status()
frames.append(_read_one_parquet(BytesIO(resp.content), symbols))
else:
frames.append(_read_one_parquet(path, symbols))
return pd.concat(frames, ignore_index=True) if len(frames) > 1 else frames[0]
def _load_dataset_from_pwb(dataset_name: str, split: str, pwb_api_key: str, symbols=None) -> pd.DataFrame:
base_url = "https://data.paperswithbacktest.com/v1/datasets"