Update NDTIFF reader #145
Update NDTIFF reader #145
Conversation
ziw-liu
added
μManager
| # TODO: try casting the dask array into a zarr array | ||
| # using `dask.array.to_zarr()`. | ||
| # Currently this call brings the data into memory |
There was a problem hiding this comment.
I believe now as_array returns a dask array without bring the data into memory, I think you can test and remove this comment
There was a problem hiding this comment.
I think this means that dask.array.to_zarr() brings data to memory if the store is a MemoryStore so this method cannot be consistent with other get_zarr() methods.
There was a problem hiding this comment.
Hopefully this won't be a problem anymore after #132 and we refactor this.
Yes, I can acquire such datasets. Where would we store them? Should we update the waveorder collection or create a new one? |
Could work either way. If we do create a new one we could also migrate the existing test datasets over too. |
|
@ieivanov do you think we should merge this as-is or wait for new test data? |
|
Opened #149 to track the test dataset issue. Merging. |
* inheriting basefov on ccfov * add deprecation warning to ccfov.scale * activating github pr workflow * Update NDTIFF reader (#145) use new ndtiff and stop sorting axes * add ccfov scales tests * Release timing requirement for I/O-heavy test (#147) * release timing requirement for I/O-heavy test * suppress data size check for arrays --------- Co-authored-by: Ziwen Liu <67518483+ziw-liu@users.noreply.github.com>
Resolves #124.
This bumps
ndtiffto 2.1.0 so that we can use the automatic axes sorting instead of pulling our own.Currently CI only tests for NDTIFFv2 datasets. @ieivanov Should we add some v3 datasets to the pool?