Validate max_seq_length in SquadProcessor

[francescocastelli]

Related Issue(s):
#2319

Proposed changes:

• What we need to validate max_seq_len is processor.tokenizer.model_max_length. I've put the assert in the SquadProcessor class because in FARMReader we would have to place it in multiple different methods. Let me know if I'm missing something!

• Should we validate the max length also on other processors? for instance TextSimilarityProcessor uses two tokenizers and has both max_seq_len_query and max_seq_len_passage.

Pre-flight checklist

• I have read the contributors guidelines
• I have enabled actions on my fork
• If this is a code change, I added tests or updated existing ones
• If this is a code change, I updated the docstrings

[julian-risch]

Hi @francescocastelli thank you for making this pull request! We will probably need a few days until we can have a look at it and give feedback. Stay tuned! 🙂

[ZanSara]

Hello @francescocastelli! I have a small style comment on this PR. I see you've been using the .format() method of strings, while we generally prefer to use f-strings, mostly for readability and consistency. Could you modify the code to use f-strings instead of .format()? Thanks 🙂

In addition, it would be great if you could add a small test for the new asserts. Reach out if you need help testing, unfortunately our test suite is not the most intuitive to deal with 😄

[francescocastelli]

@ZanSara Sure! I will fix everything next week :)

[francescocastelli]

@ZanSara done! I've also added a test for invalid parameters at the construction of a FARMReader

[ZanSara]

Really nice! I will merge it soon. Thank you for the contribution! 😊