huggingface / huggingface/datasets

`from_parquet` return type annotation

Open
#7,202 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
22k
Forks
3.4k
Avg merge
5d 7h
Merged PRs (30d)
17

Description

### Describe the bug

As already posted in https://github.com/microsoft/pylance-release/issues/6534, the correct type hinting fails when building a dataset using the `from_parquet` constructor.
Their suggestion is to comprehensively annotate the method's return type to better align with the docstring information.

### Steps to reproduce the bug

```python
from datasets import Dataset

dataset = Dataset.from_parquet(path_or_paths="file")
dataset.map(lambda x: {"new": x["old"]}, batched=True)
```

### Expected behavior

map is a [valid](https://huggingface.co/docs/datasets/v3.0.1/en/package_reference/main_classes#datasets.Dataset.map), no error should be thrown.

### Environment info

- `datasets` version: 3.0.1
- Platform: macOS-15.0.1-arm64-arm-64bit
- Python version: 3.12.6
- `huggingface_hub` version: 0.25.1
- PyArrow version: 17.0.0
- Pandas version: 2.2.3
- `fsspec` version: 2024.6.1

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.