RowsPostProcessingError in the viewer of several datasets

[albertvillanova]

We find a RowsPostProcessingError in the viewer of several datasets:

Server error while post-processing the split rows. Please report the issue.

Error code:   RowsPostProcessingError

• https://huggingface.co/datasets/multi_woz_v22
• https://huggingface.co/datasets/qasper

[severo]

It seems like the issue is that we don't support a Sequence of dicts adequately:

A datasets.Sequence with a internal dictionary feature will be automatically converted into a dictionary of lists. This behavior is implemented to have a compatilbity layer with the TensorFlow Datasets library but may be un-wanted in some cases. If you don’t want this behavior, you can use a python list instead of the datasets.Sequence.

https://huggingface.co/docs/datasets/v2.5.2/en/package_reference/main_classes#datasets.Features

[severo]

#603 only fixed part of the datasets, ie https://huggingface.co/datasets/qasper

But not https://huggingface.co/datasets/multi_woz_v22

Working on it in #605

[severo]

OK, #605 fixed https://huggingface.co/datasets/multi_woz_v22 as well

Still 29 datasets failing with RowsPostProcessingError -> it seems to be due to another reason: missing fields (columns), or None values, in some rows. Created #606 to work on it

[severo]

With #603, #605 and #606, we have fixed all the RowsPostProcessingError occurrences!