The PMP conda environment is held at pcmdi_metrics 3.9.2 while the provider pip installs a fork of 4.2.0 over the top with --no-deps. The two should match. They do not, because bumping the conda pin to 4.2.0 drags in enso_metrics 2.0.0, and that release breaks enso_proc. #901
enso_metrics 2.0.0 is a full migration off the retired CDAT stack onto xarray (CLIVAR-PRP/ENSO_metrics#95). Its new compat shim raises inside EnsoFbSstTaux.
File "EnsoMetrics/EnsoUvcdatToolsLib.py", line 5260, in get_year_by_year
tmp = MV2concatenate((tmp, zeros))
File "EnsoMetrics/XarrayCompat.py", line 413, in _validate_axes_shape
raise ValueError(
ValueError: CDATVariable(): axis 0 (time) length 38 does not match data shape 39
The metric then goes missing from the output bundle, so climate_ref_pmp/drivers/enso_driver.py:248 fails with a secondary KeyError: 'units' while plotting.
Seen on mint run 33730826484, where enso_proc/cmip6 and enso_proc/cmip7 failed and the other 20 cases minted cleanly.
Once fixed bump the conda pin to pcmdi_metrics 4.2.0 and enso_metrics 2.0.0, then re-mint.
The PMP conda environment is held at
pcmdi_metrics3.9.2 while the provider pip installs a fork of 4.2.0 over the top with--no-deps. The two should match. They do not, because bumping the conda pin to 4.2.0 drags inenso_metrics2.0.0, and that release breaksenso_proc. #901enso_metrics2.0.0 is a full migration off the retired CDAT stack onto xarray (CLIVAR-PRP/ENSO_metrics#95). Its new compat shim raises insideEnsoFbSstTaux.The metric then goes missing from the output bundle, so
climate_ref_pmp/drivers/enso_driver.py:248fails with a secondaryKeyError: 'units'while plotting.Seen on mint run 33730826484, where
enso_proc/cmip6andenso_proc/cmip7failed and the other 20 cases minted cleanly.Once fixed bump the conda pin to
pcmdi_metrics4.2.0 andenso_metrics2.0.0, then re-mint.