You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Was wondering what the status was on releasing 2.2. There seem to be several things scheduled for 2.2, but am unsure if they are all required. Are there things you need help with?
Thank you for the offer, apologies I'm under a deadline and have had to put OSS on hold for the moment.
FWIW The next thing I wanted to do for the 2.2 release was to set up a data fixture (#138) which would provide some protection against any regressions that accidentally break backwards compatibility with the data format. That would then allow safe merging of the numcodecs refactoring work (#139, #74).
Apart from that, the following I would classify as high priority for 2.2. These are mostly boring but important bug fixes, except for the tree printing which is very exciting :-) ...
There are also a set of issues relating to indexing (#89, #93, #78, #116, #119). It would be nice to resolve all these together with some reworking of the indexing code. The code you have in kenjutsu is nice and I'd really like to draw on that, but I think there are some performance bottlenecks in there which could manifest in Zarr if we just depend on kenjutsu as-is, and I wanted to explore some options. This may all be best left for a subsequent release, it's not a small piece of work.
Activity
jakirkham commentedon Apr 6, 2017
Any thoughts on this @alimanfoo?
alimanfoo commentedon Apr 6, 2017
Thank you for the offer, apologies I'm under a deadline and have had to put OSS on hold for the moment.
FWIW The next thing I wanted to do for the 2.2 release was to set up a data fixture (#138) which would provide some protection against any regressions that accidentally break backwards compatibility with the data format. That would then allow safe merging of the numcodecs refactoring work (#139, #74).Apart from that, the following I would classify as high priority for 2.2. These are mostly boring but important bug fixes, except for the tree printing which is very exciting :-) ...
Group print hierarchy (Add tree #140, Group print hierarchy #82)Index with numpy int64 Index with numpy int64 #143Various repr tweaks (Group and Array repr performance #132, ENH: read_only in repr #115, Bring back "zarr" into repr #83)Specification for datetime dtypes is ambiguous Specification for datetime dtypes is ambiguous #85Don't include partial files when counting initialized chunks BUG: Don't include partial files when counting initialized chunks #97The following would be nice to have but could live without for this release:
Usability tweaks to zarr.open (https://github.com/alimanfoo/zarr/issues/141, https://github.com/alimanfoo/zarr/issues/105)Add a save() function similar to numpy.save/numpy.savez ENH: Add a save() function similar to numpy.save/numpy.savez #104Tweak and doc chunks kwarg (ENH: chunks=None #106, DOC: chunks=True #107)Key encoding for dbm-style databases Key encoding for dbm-style databases #133Some doc improvements (DOC: Integrate docs on saving existing array to disk #103, DOC: Pickle support #91, Add persistence examples using shelve/dbm, lmdb #131)There are also a set of issues relating to indexing (#89, #93, #78, #116, #119). It would be nice to resolve all these together with some reworking of the indexing code. The code you have in kenjutsu is nice and I'd really like to draw on that, but I think there are some performance bottlenecks in there which could manifest in Zarr if we just depend on kenjutsu as-is, and I wanted to explore some options. This may all be best left for a subsequent release, it's not a small piece of work.alimanfoo commentedon Nov 20, 2017
Tracking release 2.2 via: https://github.com/alimanfoo/zarr/milestone/5