Add `--merge-pull-request` option for `convert_to_parquet`

Open klamike opened this issue 8 months ago • 1 comments

Closes #7527

Note that this implementation will only merge the last PR in the case that they get split up by push_to_hub. See https://github.com/huggingface/datasets/discussions/7555 for more details.

May 06 '25 18:05 klamike

This is ready for a review, happy to make any changes. The main question for maintainers is how this should interact with #7555. If my suggestion there is accepted, this PR can be kept as is. If not, more changes are required to merge all the PR parts.

May 07 '25 17:05 klamike

Closing since convert to parquet has been removed... https://github.com/huggingface/datasets/pull/7592#issuecomment-3073053138

Jul 18 '25 19:07 klamike