update __getitem__ and __setitem__ for DataProto

[flishwang]

Checklist Before Starting

• Search for similar PR(s).
  I found nothing similar

What does this PR do?

Simplify the usage of DataProto

I modified __getitem__ and __setitem__ so that we can:
(1) use data['input_ids'] instead of data.non_tensor_batch['input_ids'] or data.batch['input_ids'] if 'input_ids' exists in data's batch, non_tensor_batch or meta_info.
(2) update data by data[1:5] = new_data[2:6] if there's no extra keys in new_data.

Specific Changes

(1) Update __getitem__ for DataProto
(2) Add __setitem__ for DataProto

API

• data[key] for DataProto data
  Before: data['key'] will raise a ValueError
  After: it will returns data.batch['key'] or data.non_tensor_batch['key'] or data.meta_info['key'] if key in any of them.

• data[index]=other_data for DataProto data and other_data
  Before: data[index]=other_data will raise an Exception as there's no __setitem__ method for DataProto
  After: the values in rows of index (which could be a integer, ndarray or slice) will updated by other_data.

Checklist Before Submitting

• Read the Contribute Guide.
• Apply pre-commit checks.
• Add [BREAKING] to the PR title if it breaks any API.
• Update the documentation about your changes in the docs.
• Add CI test(s) if neccessary.

[CLAassistant]

All committers have signed the CLA.