Matt McCormick
Matt McCormick
Great, thanks for the update @jo-mueller ! I'll leave this open until we have better default behavior -- ideally refined chunk specification that works downstream with zarr-python and dask. Or,...
@jo-mueller > Maybe a quick check there could make sense to see whether data already comes in chunks, and take the chunksize as-is, if so? After all, chunk sizes may...
@jo-mueller interested to know whether these errors are still observed with recent fixes.
> The changes introduce new data structures for coordinate systems and transformations, update the NgffImage class, and revise functions for creating and writing NGFF Zarr images to use the new...
@jo-mueller thanks for the follow-up and explanations! And your hard work on this. Yes, could please shoot me an email at [email protected] I'll send a calendar scheduler to chat. We...
CC @ntustison
pygccxml should support it per: https://github.com/CastXML/pygccxml/blob/26f63dc1f473114f836a43c309eaf5f65e5c5554/src/pygccxml/parser/source_reader.py#L159-L170 but our (vendored) pygccxml may need updates, fixes made, etc. In ITK, `--castxml-gccxml` is passed in our CMakeLists.txt CC @iMichka
Hi Flo, Great to hear ngff-zarr is helpful in your work! Re: the y-axis: ngff-zarr uses lower-left for 0,0. The may be visualization tool differences here. Visualization tools often flip...
Hi Flo, My pleasure :-) The y-flip can be confusing. Especially since there is both the consideration of how the data is loaded in memory and also what direction is...
Currently missing scipy dep.