Preserving full-strength validation & error reporting

[joncison]

One of many issues around GitHub-based content management for bio.tools.

[joncison]

Implementation (or restructuring) of logic to provide full-strength validation of tool information (including validation against biotoolsSchema and other checks not encodable in the schema) with human-readable error reporting, invocable as part of GitHub integration.

• validation in bio.tools goes beyond what biotoolsSchema provides (e.g. unique tool identifiers based on tool name, non-deprecated EDAM concepts etc.)
• meaningful / human-readable error reporting is essential (in 1st instance / in "toy" implementation, pointing users to XSD/XML validation log could be a starting point)
• validation as part of CI such that commits must pass validation checks to be accepted / before pull request is issued / accepted
• need for standalone validator, see biotoolsLint
• provide the validator as a Github app / bot, which must pull everything it needs (not stateful!) as it will be spun up in it's own VM on demand - cacheing handled by bot-hosting service

[joncison]

UPDATE

Made some major progress on this folks - biotoolsLint is more or less fully specified from a functional perspective, i.e.. the verifications that it should perform - which includes everything that bio.tools currently does and more besides. See the biotoolsLint README. Plan for the implementation will follow.

cc @piotrgithub1 @hansioan @hmenager @matuskalas @bgruening @JenHarrow