different behavior in pixi v conda env #44
Description
Trying in a fresh pixi env there's some slight changes in behavior from when i run in an env created with mamba/conda-forge. haven't looked into it yet but first guess is its due to pkg versions (prob xarray)?
running list of differences here then will try to pin down cause
- in conda env, must specify engine='zarr' else:
KeyError: [<class 'netCDF4._netCDF4.Dataset'>, ('http://its-live-data.s3.amazonaws.com/datacubes/v2/N30E090/ITS_LIVE_vel_EPSG32646_G0120_X750000_Y3350000.zarr',), 'r', (('clobber', True), ('diskless', False), ('format', 'NETCDF4'), ('persist', False)), '77d84055-bd8e-4d42-aca8-4dbb3fdd1cc1']
During handling of the above exception, another exception occurred:
OSError: [Errno -72] NetCDF: Malformed or inaccessible DAP2 DDS or DAP4 DMR response: 'http://its-live-data.s3.amazonaws.com/datacubes/v2/N30E090/ITS_LIVE_vel_EPSG32646_G0120_X750000_Y3350000.zarr'
- in pixi env its fine
pixi env has xr 2025.1.2, conda has xr 2024.10.0
- on two cells in this section (
np.unique(single_glacier_raster.satellite_img1),single_glacier_raster = single_glacier_raster.compute(), get the following warning:
home/emmamarshall/Desktop/phd_research/ch1/playground_2/cloud-open-source-geospatial-datacube-workflows/.pixi/envs/default/lib/python3.11/site-packages/dask/array/core.py:133: SerializationWarning: Can't decode floating point timedelta to 's' without precision loss, decoding to 'ns' instead. To silence this warning use time_unit='ns' in call to decoding function.
Seems related to this