Define logic for suits with tools that have different bio.tools entries

[paulzierep]

Currently, we assume that all tools in a suite (i.e. github folder) have the same bio.tools ID.
However, there are some suites where the contained tools have different bio.tool IDs: For example: https://github.com/galaxyproject/tools-iuc/tree/main/tools/spades

As a consequence, our tool only parses one bio.tools entry, which is not correct for all tools.
We need to brainstorm how to handle these cases.

A) Aggregate the bio.tools entries and EDAM Terms (similar to the Galaxy tool ids row)
B) One row for each Galaxy tool ids (this would probably mean a good amount of restructuring, i.e. creating a tool list and an aggregated list)

Any vote for either way ?

[matuskalas]

Oh no, this is a nightmare I was expecting. Such cases are basically a wrong use of Bio.tools.

spades is a good example, with Bio.tools records generated from Galaxy@Pasteur, separate records for subtools, with identical homepage, citations, version, etc. The correct record is https://bio.tools/spades (and I'll be happy to delete the rest🐱‍👤)

What needs to be done, is to include into Bio.tools linter checks for identical homepage, repo & other links (download, documentation, ...), primary citation, ... 😨

On the side of Tool Extractor, I see 3 options, ordered from IMHO least favourable to most:
A) Aggregate, but this is working around a situation that probably shouldn't exist.
C) Fail and prompt to fix the mess (Should ideally be part of linting the tool wrappers)
D) As C, but just throw a warning and ignore the Bio.tools IDs

[matuskalas]

Could you output all the wrappers with this problem? Billion thanks! 🙏🏽🙏🏽🙇🏽‍♂️

[paulzierep]

We have no functionality to automatically see which tools have multiple entries, yet.
But I will collect the cases here:

https://github.com/galaxyproject/tools-iuc/tree/main/tools/bbtools

[paulzierep]

I will update the parser to collect all IDs if there are multiple

[paulzierep]

Now, we produce a column with bio.tools ids #72
Overall, there are only like 4 or 5 cases, so very minor issue ! And can be fixed by using the tool suite as xref for all tools in a suite.