Add default preprocessor, add check for existence of valid data

[phirework]

• Seems like we've added a whole bunch of languages since the last dataset release and instead of copying the same boilerplate file 20 times and keeping track of new languages I added a default preprocessor that just spits back the sentence. Also deleted all the language preprocessors that were just the boilerplate - let me know if there was a specific reason these were in separate files. Also amended documentation to explain this.
• We have 108 clips from vot locale (Votic) where none of it has any votes (because there are 20 living speakers in the world), which was breaking _calculate_data_set_sizes since the loop never runs and train_size et al are never instantiated, so I added a length check.
• Added an exception handler to skip Attribute errors to deal with the case of a Corpus not having buckets as per above.

I am not a python dev, so please do let me know if there's a more pythonic way of handling any of the above.

[johngian]

Hi @phirework ! Here are some comments for the PR after your feedback request:

• You are deleting some preprocessors (eg. br.py) completely but under preprocessors/__init__.py we still import them, which might raise an error.
• Is there a way I can test it locally? I am not really familiar with the tool (I only run it once with en dataset).
• There are some comments in-line. Might be a little nit-picky but I don't really know the context of the tool. Feel free to omit.
• flake8 linter raises a lot of errors. Most of them were not introduced in this PR but it might worth it to check what was introduced in the diff.

[phirework]

Thanks for the review! Removed unused preprocessors.

There's no easy way to test locally, you could generate a live clips.tsv using the bundler (https://github.com/Common-Voice/common-voice-bundler/) and then running the result of that through this. I can also send you the full clips.tsv file I was working off of for the 2019 H2 corpus.

Good call on the lint issues, I'm going to file a separate ticket to clean that up.