search

RobokopU24 / Feedback

Feedback on the ROBOKOP project
https://robokop.renci.org
0 stars 0 forks source link

Text Mining Provider #190

Closed karafecho closed 5 months ago

karafecho commented 10 months ago

This issue is to report an issue with automat-text-mining-provider. Specifically, Bill Baumgartner mentioned that the Automat version of Text Mining Provider is significantly outdated and should not be used. I clarified after today's ROBOKOP call that the issue is NOT related to targeted vs co-occurrence, but rather related issues that have since been fixed. We will need to update our source data.

- id: infores:automat-text-mining-provider
    status: deprecated
    name: Automat Textmining KP
    knowledge level: curated
    agent type: not_provided
    description: A literature co-occurrence graph created by the Translator Text Mining KP.
cbizon commented 5 months ago

@EvanDietzMorris @eKathleenCarter what's the status here?

EvanDietzMorris commented 5 months ago

I suspect this issue was initially due to us having an old version on the prod deployment, when we had already gotten the latest one up on dev. Previous discussions about this resulted in the conclusion that we did have the right thing. However, the long term updating process for tmkp is not great. We attempt to get the latest textmining kp from their google storage, and use the modified date as the version. If they change where they're putting the latest, or the file name changes, our pipeline will break and we won't have the latest anymore.

We use "edges.tsv.gz" from: https://console.cloud.google.com/storage/browser/translator-text-workflow-dev-public/kgx/UniProt

cbizon commented 5 months ago

So can this issue be closed?

EvanDietzMorris commented 5 months ago

I think so, the potential issue with changing file name/location is not really specific to textminingkp, per say.