fix #10811: add templatePrefix to Fileset

[mtbc]

Fixes https://trac.openmicroscopy.org.uk/ome/ticket/10811

To test, try doing some imports and check that the templateprefix column in the fileset table in the database matches the appropriate prefix of the path column in originalfile.

[joshmoore]

@mtbc: @qidane already used dbpatch 5, can you use 6 please?

[mtbc]

Hmmm, experiencing github weirdness, will delete the branch and re-push.

[mtbc]

Okay, maybe that fixed it...

[mtbc]

@joshmoore: okay, bumped to dbpatch 6, cc @chris-allan

[chris-allan]

Did you just rebase and/or rewrite the history for this? That will have wiped out the discussion we've just been having.

[mtbc]

Yeah, github seemed to get extremely buggily confused about further pushes and, though there were extra commits locally and git push said everything was up to date, they vanished from github's web interface. Very irritating.

The discussion is at mtbc@54829cf

[mtbc]

@joshmoore how do I structure this so this dbpatch bump to 6 doesn't conflict with #1133, should I rebase against it?

[mtbc]

I just cannot get this to rebase well against #1133. Please mark as excluded for now, I'll wait until #1133 is merged then try again.

[mtbc]

Please remove the exclude tag.

[joshmoore]

Done

[joshmoore]

@mtbc: sorry, @pwalczysko has asked for PURGE_DATA to not occur tomorrow in which case this must be excluded again.

[mtbc]

No problem, testing comes first; let me know when you unexclude it again and I'll add it to standup.

[joshmoore]

Purging re-activated on gretzky; label removed.

[bpindelski]

Imported a dv folder to gretzky as user-1. In the Postgres console I found the fileset id of one of the imported files. Then executing

select o.path, f.templateprefix from fileset f join filesetentry fe on f.id = fe.fileset join originalfile o on o.id = fe.originalfile where f.id = 45;

results in

               path                |          templateprefix
-----------------------------------+-----------------------------------
 user-1_8/2013-06/04/10-21-54.547/ | user-1_8/2013-06/04/10-21-54.547/

Looks good to merge.

[joshmoore]

It would require a significant refactoring (i.e. not a part of this PR), but I wonder if we wouldn't be better suited to have a third element (template+path+name) which is perhaps not even exposed to the user. But that's just thinking out loud...

@chris-allan: I would plan on merging this tomorrow morning if that won't cause you any headaches sans upgrade script.

[joshmoore]

👍 from @chris-allan. Merging.