Malformed link URL

[maxglycine]

In this report https://mines.legumeinfo.org/glycinemine/report.do?id=892000001, the data source URL is malformed. It goes to: https://legumeinfo.org/data/v2/. It is missing the part of the URL that specifies the data source. It came from the "Data Sets" section and the "Data Sets Name" link @StevenCannon-USDA @jd-campbell @ChenProm-USDA

[adf-ncgr]

Thanks @maxglycine looks like this is probably coming from: https://github.com/legumeinfo/lis-bio-sources/blob/6bc4b9d78c63c8dff06a61110a48374435555656/lis-datastore/src/main/java/org/intermine/bio/dataconversion/DatastoreFileConverter.java#L55 likely it was considered correct when the code was written. Easy enough to fix the code but might take a while before that change propagates into the databases; it seems like it should also be easy to update in the databases directly, but I suspect it is more complicated than you'd guess. In any case we can leave the issue open until it's fully fixed.

[StevenCannon-USDA]

Explicitly listing the options: Incorrect:

• https://legumeinfo.org/data/v2/

Correct:

1. https://data.legumeinfo.org/data/v2/
2. https://data.legumeinfo.org
3. https://data.soybase.org/data/v2/
4. https://data.soybase.org

Of those options, I think number 4 would be preferred in GlycineMine, as long as (1) it can be done specifically in that Mine, and (2) not mess up LegumeMine or other generic tools that assume the data.legumeinfo.org form. Pulling in @nathanweeks to see if there are any other considerations regarding the form of the Data Store URL.

[adf-ncgr]

It looks like it is probably possible to do this in a mine-specific way via properties in the project.xml file that controls mine building. I'll note that the individual datasets do also currently refer to data.legumeinfo.org URLs, but those could also be updated in the soymine project.xml file if we decide that's how we want it to be.

[adf-ncgr]

FYI, just as a proof of being able to do it, I went ahead and made a direct update to the glycinemine database to change the URL stored there after the fact. I knew that this would involve installing the famous Joe Carlson triggers but it also seemed to require a redeploy of the web app presumably due to some caching. We probably ought to also modify the name and description of this data source (e.g. it is still called "LIS Datastore" despite the change in URL) but at least this is a step in the right direction. We will also want to update the project.xml file so that we don't have to do this every time we rebuild it, but that should be straightforward I hope (though not very clearly documented as to specifics).

[StevenCannon-USDA]

Thank you, @adf-ncgr! What database gumption!