Support single value for attribute list in doc.to_array

[ramananbalakrishnan]

Description

• Updates Doc.to_array(attr_ids) to accept a single value in attr_ids
• Partial fix for discussion in issue to_array(scalar) has bad error message #1410.

Types of changes

• New feature (non-breaking change adding functionality to spaCy)

Checklist:

• My change requires a change to spaCy's documentation.
• I have updated the documentation accordingly.
• I have added tests to cover my changes.
• All new and existing tests passed.

[ramananbalakrishnan]

I am not a Cython expert, so could not get the reshape command working as-is. If there is a better to way to achieve the same result, do let me know.

Also, can I directly use intify_attrs to map a string name ("LEMMA") in the input list to the attribute id (spacy.attrs.lemma) ?

[explosion-bot]

Hi @ramananbalakrishnan, thanks for your pull request! 👍 It looks like you haven't filled in the spaCy Contributor Agreement (SCA) yet. The agrement ensures that we can use your contribution across the project. Once you've filled in the template, put it in the .github/contributors directory and add it to this pull request. Thanks a lot!

[ramananbalakrishnan]

I have ported the changes over to the develop branch as well, for merging into v2. ( PR #1440 )
I should be able to keep track of suggestions from either PR and fix across both branches.

[honnibal]

Thanks for this! It will be merged in time for the new v2 release.

[honnibal]

Also, can I directly use intify_attrs to map a string name ("LEMMA") in the input list to the attribute id (spacy.attrs.lemma) ?

I think so. It's possible there's some quirks, but yes that's what we would want to be using.

[ramananbalakrishnan]

Done. to_array now supports string names in attr_ids.
Updated tests and the v2 PR as well.

[honnibal]

🎉

Thanks for this! Made some small changes, especially changing use of type in the conditionals. This is a little bit like the weird equality checking stuff in Javascript --- it's best to avoid type-based checks in Python if possible, because the language makes it hard to make the logic fully correct. hasattr checks are ugly, but it's a bit easier to cover all the cases this way.

[ramananbalakrishnan]

got it! looks much cleaner now.

Also, since you are directly returning output.reshape(), looks like the output_1D variable is redundant now.