Open adkinsrs opened 4 weeks ago
To my knowledge the dataset_shares table isn't actually used at all. The only time shares are involved is if they're passed on the URL and they survive for that page load only. A solution to bring this back into use would be to insert into dataset_shares anytime a URL with a private share is accessed.
But since none of this is currently implemented anyway, we could drop dataset_shares and instead have something like "shares" which could generally store them for any object type.
shares
Certainly up for arguments.
To my knowledge the dataset_shares table isn't actually used at all.
https://github.com/IGS/gEAR/blob/c23f910506aed171f1e73fb353535a2a016bc4f6/www/cgi/search_datasets.cgi#L120-L127. (plus line 54's function which queries the dataset_shares table)
And as for updating the table to just be a generalized "shares", I have no arguments against it.
Oh, absolutely. In the older interface there was a control page to formally share with a user, the user accepted, it inserted into the database, etc. But that was removed some time ago and there are probably some queries laying around which still reference it.
Currently for "ownership" filtering, there is an option to include/exclude gene lists that have been shared with you. Right now I just have this block on the server-side code set to "pass" because we have no current way to implement it. For the dataset explorer page, there is code to retrieve the share ID for a dataset if a user has been shared that dataset (via the dataset_shares table). Currently, no such equivalent table exists for gene lists.
So how to handle this?
@jorvis thoughts? I can make the code changes.