Revisiting tool relations

[joncison]

Following suggestions (from review of JPR paper and discussions) the idea is to support the definition of relationships between tools (hence between bio.tools entries), building on a first shot from biotoolsSchema v2 (which didn't make the cut into v3).

The basic model will be like this:

Rather than go for a comprehensive list of relations, I propose to start with a basic set of relations - things that we can realistically curate to improve bio.tools user experience - and see how it pans out, before extending the model further.

I propose:

• isNewVersionOf The tool is a new version of an existing tool, typically providing new or improved functionality.

• hasNewVersion The tool has a new version, which typically provides new or improved functionality.

• includes A suite or workbench [1] includes some other, independently available, tool.

• includedIn The tool is included in some suite or workbench [1].

• uses The tool uses some other tool under the hood, e.g. invoking a command-line tool or calling a Web API, Web service or SPARQL endpoint to perform its function. This includes wrappers (Web wrappers) of command-line tools, and workflows of multiple tools.

• used_by The tool is used by some other tool, e.g. a command-line tool that is wrapped by a Web interface, used in a workflow or by another tool. Or a Web API, Web service or SPARQL endpoint that is invoked by another tool.

[1] maybe also Library, although I don't expect we'll be registering individual components of libraries.

Curation for all of these is very do-able, building on what we already have in collection annotations, and would be a big win for bio.tools users.

@matuskalas @albangaignard @hansioan @piotrgithub1 @magnuspalmblad @veitveit @hmenager could you please provide a comment - is this sensible? Bear in mind I want a light touch (I mean easily curatable things) to begin.

Thanks!

Let's keep this thread for the discussion of the modelling only. I'll create separate threads for bio.tools refactoring and curation, those and link them here. See also https://github.com/bio-tools/biotoolsSchema/issues/107 and https://github.com/bio-tools/biotoolsSchema/issues/106 (old threads).

[veitveit]

+1

[joncison]

spoken to @hansioan in details ... nudge-nudge @matuskalas @albangaignard @piotrgithub1 @magnuspalmblad @hmenager - implementation of schema changes is about to start. Any comments?

PS. this is the starting point - which we want to be stable, but also extensible

[joncison]

@hansioan

pls. take a quick look at

• https://biotoolsschema.readthedocs.io/en/latest/biotoolsschema_elements.html#relation-group
• https://biotoolsschema.readthedocs.io/en/latest/controlled_vocabularies.html#relation-type
• https://biotools.readthedocs.io/en/latest/curators_guide.html#relation-group
• https://biotools.readthedocs.io/en/latest/api_usage_guide.html#relation

and confirm all looks good

[albangaignard]

Thanks Jon for this proposal. This seems a very good starting point ! It can be very nice for tool providers to know that their contribution is used in another tools.

I quickly checked PAV https://pav-ontology.github.io/pav/ and I don't see anything missing to represent a set of relations between tools. May be a "used_with" could be helpful but this could be inferred from "includedIn" relations.

I'm not sure to understand the subtleties between isNewVersionOf an hasNewVersionOf. Is "has" aimed at representing tools compartment ? Can you provide an example ?

Le 28 mai 2019 à 13:43, Jon Ison notifications@github.com a écrit :

spoken to @hansioan https://github.com/hansioan in details ... nudge-nudge @matuskalas https://github.com/matuskalas @albangaignard https://github.com/albangaignard @piotrgithub1 https://github.com/piotrgithub1 @magnuspalmblad https://github.com/magnuspalmblad @hmenager https://github.com/hmenager - implementation of schema changes is about to start. Any comments?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/bio-tools/biotoolsSchema/issues/145?email_source=notifications&email_token=ACRMMEB4E4LP7HJNF6BYVXLPXULEPA5CNFSM4HNLEG5KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODWL3K2I#issuecomment-496481641, or mute the thread https://github.com/notifications/unsubscribe-auth/ACRMMEAZTLPT7BZTGIXNMLLPXULEPANCNFSM4HNLEG5A.

-- Alban Gaignard, PhD, CNRS

[joncison]

thanks @albangaignard

hasNewVersioOf should have been hasNewVersion and is simply the inverse relationship of isNewVersionOf ... sorry about that

we can indeed tweak more, once we see how this 1st pass pans out .... used_with is an interesting one, as there's value knowing what tools are used with what others. For workflows I'm imagining having workflow uses tool and thus inferring used_with

We'll have to really drill down on the precise semantics in due course ...