report to list complex licenses in definitions

[elrayle]

Description

Complex licenses may reflect the actual license, but also may reflect false positives coming from the tools generating the license data. Identifying complex licenses reveals licenses that may benefit from human review and possible curation.

Write a utility to generate a report from ClearlyDefined data listing complex licenses.

Considerations:

• define complex (e.g. any license with an AND, any license with more than one connector, etc.) - are there other characteristics that can be used to identify complex licenses?
• consider presentation (e.g. group licenses by the number of connectors with the most appearing first)
• determine a way to mark a license as verified to prevent it from showing up on a future report if a human has reviewed it and finds it to be correct

[elrayle]

Backfill

The proposed work to create a report addresses the need to identify existing definitions that could benefit from curation. This would be resource intensive and is more of a backfill procedure that would not be repeated.

Operational

To assist with the ongoing need for curation, the following can be done at definition compute time.

When a complex license is identified during compute definition, check if there were curations for other versions.

• IF a curation doesn't exist
  • create an issue requesting human review and possible curation
• ELSE
  • IF all versions have the same license, create a PR with that license for the new version reducing the work to review, approve, and merge
  • ELSE create an issue and link to the existing curation file to assist human exploration and possible curation