New method: apply-chain-rules

[matentzn]

Despite some overlap with boomer, I would like to have a function that basically implements https://mapping-commons.github.io/sssom/chaining_rules/ after all.

Basically, the method should work just like invert, but create new mappings by applying the chaining rules above - once. For example, if we have (:A)-[predicate_id]->(:B)-[predicate_id]->(:C), we should infer: (:A)-[predicate_id]->(:C).

Chris will be against this adding all these methods here, but they are needed for our vision of mapping commons.

[cmungall]

I would tend towards keeping the core sssom py light on logic and focus on access, logic in oak

…

On Sun, Apr 9, 2023 at 10:24 AM Nico Matentzoglu ***@***.***> wrote: Despite some overlap with boomer, I would like to have a function that basically implements https://mapping-commons.github.io/sssom/chaining_rules/ after all. Basically, the method should work just like invert, but create new mappings by applying the chaining rules above - once. For example, if we have (:A)-[predicate_id]->(:B)-[predicate_id]->(:C), we should infer: (:A)-[predicate_id]->(:C). Chris will be against this adding all these methods here, but they are needed for our vision of mapping commons. — Reply to this email directly, view it on GitHub <#365>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAMMOIVGAEMINZJTLA3IKTXALWEDANCNFSM6AAAAAAWYHI2DM> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[cthoyt]

I've implemented these functionalities in semra. I will be giving Nico a walkthrough next week.

[cmungall]

I'm OK with implementing in sssom-py. If it's not urgent then we could wait til whelk.rs is ready and avoid writing our own reasoner

[matentzn]

Ok, thanks. I would suggest not to call simple rule execution reasoning to avoid mistaking it for Tableau or CBR; I would suggest we keep the scope down to a few well defined rules: https://mapping-commons.github.io/sssom/chaining_rules/