Closed acrellin closed 8 years ago
@bnaul, did your featurization refactor address this issue?
No, it just changed in the other direction: now instead of datasets disappearing instantly, they just stick around forever. We aren't storing the paths in the database or anything.
On Wed, Dec 2, 2015 at 4:08 PM acrellin notifications@github.com wrote:
@bnaul https://github.com/bnaul, did your featurization refactor address this issue?
— Reply to this email directly or view it on GitHub https://github.com/mltsp/mltsp/issues/104#issuecomment-161472872.
Ok, so the paths just need to be stored in the db now. I'll do a quick fix soon.
Just realized I have 500MB of extracted time series hiding out various tempfile
directories (/var/folders/ts/...
). Figure this is the same issue, just thought I'd make a note that whatever we do for file handling should also remember to remove these temporary directories.
EDIT: Oh...I guess this was already changed in 4b6c5846fd1a0978c1e1a203cf6d3583eb1af132? Maybe those were all just leftover from a long time ago.
We do store the URI in the database now so I'm closing this; #99 contains other data handling improvements.
Add checks such that a dataset cannot be delete if there are results that depend on that dataset, or at least warn the user about the consequences before proceeding.