autocast number types

[jspaaks]

Related issues

Refs: #91

• This PR includes 148-sort-schema/PR Sort fields in schema.json alphabetically #218 and lint-schema / PR Lint schema #201 branches. Merging PR Sort fields in schema.json alphabetically #218 will clean up the diff for schema.json.
• Also awaiting better consensus after Auto-cast numbers to strings where needed #91 (comment).

Describe the changes made in this pull request

from #91 (comment)

Now that we have switched to JSONSchema, we should re-evaluate whether union types for some keys are useful. For example, it is currently (https://github.com/citation-file-format/citation-file-format/blob/38d9b1fb629373f6a3607de64aa3a6e21a76f750/schema.json) invalid to write the following in CITATION.cff:

version: 1

version: 1.11

number: 41

issue: 11

This is a bit unexpected and inconvenient for people particularly those who write CITATION.cff files by hand / unassisted.

I went over the keys and tried to identify a list of keys that could potentially have values that are interpreted as numbers by YAML, which would render them invalid. Here's the list I came up with:

1. #/definitions/post-code
2. issue
3. number
4. reference/version
5. section
6. version

This PR expands their "type": "string" to a str|number union type using JSONSchema "anyOf".

This also meant that some of the tests were no longer necessary, specifically the tests that verify that certain CITATION.cff files do not pass validation. This PR removes those tests (from examples/1.2.0/fail/**)

Instructions to review the pull request

cd $(mktemp -d --tmpdir cff.XXXXXX)
git clone https://github.com/citation-file-format/citation-file-format .
git checkout 91-autocast-number-types
python3 -m venv env
source env/bin/activate
pip install --upgrade pip setuptools wheel
pip install -r requirements.txt
pytest

pytest should pass with 68 (not 70 like before) items discovered

[sdruskat]

Think this needs a resolution of #91 per discussion before it can be reviewed, and may need more work after we've come to a decision there.

[sdruskat]

Thanks, excellent work.
Should we perhaps deduplicate version via a definition?

[jspaaks]

Should we perhaps deduplicate version via a definition?

Created #226

[jspaaks]

Thanks for reviewing!