Gregory Lee

consolidated metadata storage for Zarr v3

12

In the current draft of the v3 spec in zarr-developers/zarr-python#898, consolidated metadata has been implemented similarly to v2. Currently this metadata gets stored in the group `meta/root/consolidated`, but this is...

add zarr-python v3 spec tests cases

1

marked as a draft since this is using a branch corresponding to unmerged upstream PRs Note that the environment file here installs from the branch corresponding to zarr-developers/zarr-python#898

In #24, zarrita was added to `environment.yml`, but is installed from a development branch. Once alimanfoo/zarrita#36 is merged, this should be updated to use the main branch again

multithreading issues

7

I only recently downloaded the library and am not sure what the expected result is, but currently the benchmarks involving multiple threads do not show any improvement for me. possibly...

adding a test suite and continuous integration

4

**Is your feature request related to a problem? Please describe.** In considering the JOSS review criteria, I think the lack of a test suite or any type of continuous integration...

add axes support in ndimage filters module

7

This MR updates several `cupyx.scipy.ndimage` filters with support for the `axes` arguments that was added in SciPy 1.11.

This MR includes the commits from #20809, so that one should be reviewed and merged first. #### What does this implement/fix? This is the second of two MRs that finish...

ENH: ndimage: extend ndimage filter axes support to correlate and convolve

5

#### What does this implement/fix? This is the first of two MRs that finish extending axes support to the remaining functions in the `_filters` submodule of ndimage. We added this...

proposal: cast small integer types to single precision

5

## Description Currently helpers like `_supported_float_type` and the utilities `convert_to_float` and `img_as_float` convert any integer type to `float64`. Now that we have single precision support almost everywhere, I think it...

Improve performance of `peak_local_max` and feature detectors that rely on it

1

## Description 1.) The inner `cdist` call in `ensure_spacing` is currently very inefficient because it is taking a code path that uses a Python loop. Specifing "minkowski" instead of the...