Closed rly closed 6 months ago
Attention: Patch coverage is 91.66667%
with 1 lines
in your changes are missing coverage. Please review.
Project coverage is 79.47%. Comparing base (
731fcb5
) to head (ffea54a
).
Files | Patch % | Lines |
---|---|---|
...ndi/LindiH5pyFile/LindiReferenceFileSystemStore.py | 66.66% | 1 Missing :warning: |
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Looks good to me!
Thanks. Released in version 0.3.9.
Fix #76.
For an HDF5 dataset with shape (0,0,0), no chunks should be written. This was the previous behavior before #68 and seems reasonable.
For an HDF5 dataset with shape (0,0,0), the
chunks
key in.zarray
was being set to (0,0,0). However, that is not allowed in Zarr. Accessing an array with such chunking results in aZeroDivisionError: division by zero
when trying to access the array. I set it to (1,1,1) instead which is what Zarr does for an array of that shape.I also fixed an issue flagged by pyright in
LindiReferenceFileSystemStore.__contains__
: