# SciPy 2024 - Zarr Sprints ## Ideas * Xarray/Dask integration tests and documenting what doesn't work today (https://github.com/zarr-developers/zarr-python/issues/1762) * :point_up_2: this issue needs additional detail * port consolidate_metadata from v2 (https://github.com/zarr-developers/zarr-python/issues/1161) * :point_up_2: this issue needs an update * CPU and memory profiling for different IO loads. @d-v-b did some basic comparisons with tensorstore here but only looked at run time * https://github.com/d-v-b/zarr-workbench/tree/main * https://github.com/zarr-developers/zarr-python/issues/2034 * https://github.com/zarr-developers/zarr-benchmark * an http proxy that presents a zarr array API over http, potentially wrapping a non-zarr data source, e.g. hdf5 or png or whatever * https://github.com/zarr-developers/zarr-python/issues/2033 * a sweet html repr for groups and arrays * https://github.com/zarr-developers/zarr-python/issues/2026 * explore whether now is the time to spin off the xarray lazy slicing array interface into its own package, so zarr can use it * https://github.com/zarr-developers/zarr-python/discussions/1603 * explore batched IO operations, like creating 10 arrays with a single API call * https://github.com/zarr-developers/zarr-python/issues/1806, https://github.com/zarr-developers/zarr-python/issues/1398 :point_left: I don't think either of these are right? * * zipstore (https://github.com/zarr-developers/zarr-python/issues/2010)