huggingface / huggingface/datasets
URL query parameters are set as path in the compression hop for fsspec
- Dominant language
- Python
- Stars
- 22k
- Forks
- 3.4k
- Avg merge
- 5d 7h
- Merged PRs (30d)
- 17
Description
## Describe the bug
There is an ssue with `StreamingDownloadManager._extract`.
I don't know how the test `test_streaming_gg_drive_gzipped` passes:
For
```python
TEST_GG_DRIVE_GZIPPED_URL = "https://drive.google.com/uc?export=download&id=1Bt4Garpf0QLiwkJhHJzXaVa0I0H5Qhwz"
urlpath = StreamingDownloadManager().download_and_extract(TEST_GG_DRIVE_GZIPPED_URL)
```
gives `urlpath`:
```python
'gzip://uc?export=download&id=1Bt4Garpf0QLiwkJhHJzXaVa0I0H5Qhwz::https://drive.google.com/uc?export=download&id=1Bt4Garpf0QLiwkJhHJzXaVa0I0H5Qhwz'
```
The gzip path makes no sense: `gzip://uc?export=download&id=1Bt4Garpf0QLiwkJhHJzXaVa0I0H5Qhwz`
## Steps to reproduce the bug
```python
from datasets.utils.streaming_download_manager import StreamingDownloadManager
dl_manager = StreamingDownloadManager()
urlpath = dl_manager.extract("https://drive.google.com/uc?export=download&id=1Bt4Garpf0QLiwkJhHJzXaVa0I0H5Qhwz")
print(urlpath)
```
## Expected results
The query parameters should not be set as part of the path.
Contributor guide
Assessment
This issue has not been assessed yet.