following the reasoning in the paths of the UI

sstemann commented 1 year ago

Doing the "math" as Andy calls it of the reasoning is cumbersome for users. We're not sure if this is something the UI can enhance.

sandrine-m commented 12 months ago

@sstemann is this issue linked with #282 and #342 or is that a completely different issue? I am trying to move along this "no status ticket". Thank you!

gglusman commented 12 months ago

I read this as being closer to #342, but it's a wider issue: it would be great if Translator had much stronger ways to explain the provenance of each assertion.

When a KP reports an assertion drawn from a database, it's quite sufficient to link out to that database, ideally to a page in that database that includes that assertion (not the homepage/search).
When a KP reports a text-mined assertion, linking to the paper is great, but frequently insufficient as the user may have a hard time finding how the paper connects to the assertion. It would be great to be able to provide a dynamic page that explains it fully, including important bits like differences between the identifier used in the assertion and the actual name used in the paper.
When a KP reports an assertion computed from raw data, it would be great to offer a dynamic page explaining how the specific assertion was computed from the raw data, with details on cohort size and such.
Finally and to Sarah's point, when an ARA reports a path it constructed, it would be useful to have a dynamic page explaining to the user the specific logic rule(s) that were used to do that reasoning.

To clarify, when I write 'a dynamic page', I mean a page that is tailored to the specific example being explained, as opposed to a page that explains what the KP/ARA does in generic terms.

sandrine-m commented 12 months ago

Thank you @gglusman ! Those are great points! If I can add to your first point, it could be nice to be able to link in a user readible format to the definitions of our categories and predicates from the biolink model as a way to express possible discrepencies in what is reported by a source and what predicate maps to this definition. That would explain (provenance) the choice of vocabulary presented to the user and internally represented.

gprice1129 commented 11 months ago

@sstemann @gglusman @sandrine-m we are not sure if this ticket has changed from the initial intent or what the feature request actually is. Please advise.

sierra-moxon commented 6 months ago

from TAQA: Matt: in the retrieval source object - source_url holds the direct link, but coordinating the implementation of this with KPs hasn't bubbled up yet. Gwenlyn: there is already lots of EPC on the edges from TMKP - we just need a link to that information (e.g. the sentence, provenance of edges -- Bill: the feedback interface built for SEMMED working group - consolidates EPC into a visual form). Gus: this is exactly what we need; need to get back to us on when to implement. seems high priority.

sierra-moxon commented 3 months ago

@gprice1129 - is there any priority being discussed for this issue in the UI team?

gprice1129 commented 3 months ago

@sierra-moxon yes, we have discussed this quite a bit. We will be working on this in the future, but it probably won't be implemented by Octopus.

NCATSTranslator / Feedback

following the reasoning in the paths of the UI #263