Exporter: Inconsistent Grid points between forecast steps

[Matteo-Salis]

What happened?

I tried to export an hourly ERA5 model (muly9d4v) on Santis using the following command: uv run export --run-id muly9d4v --stream ERA5 --output-dir ../export/km-scale_eval --format netcdf --regrid-degree 0.025 --regrid-type regular_ll

but I get this error:

Batch 1/2:   5%|███▎                                                        | 21/384 [00:03<00:37,  9.72it/s]Traceback (most recent call last):
  File "/users/msalis/project/weather_generator/WeatherGenerator/.venv/bin/export", line 10, in <module>
    sys.exit(export())
             ^^^^^^^^
  File "/users/msalis/project/weather_generator/WeatherGenerator/packages/evaluate/src/weathergen/evaluate/export/export_inference.py", line 235, in export
    export_from_args(sys.argv[1:])
  File "/users/msalis/project/weather_generator/WeatherGenerator/packages/evaluate/src/weathergen/evaluate/export/export_inference.py", line 303, in export_from_args
    export_model_outputs(dtype, config, **kwargs)
  File "/users/msalis/project/weather_generator/WeatherGenerator/packages/evaluate/src/weathergen/evaluate/export/export_core.py", line 399, in export_model_outputs
    processed = parser.process_sample(
                ^^^^^^^^^^^^^^^^^^^^^^
  File "/users/msalis/project/weather_generator/WeatherGenerator/packages/evaluate/src/weathergen/evaluate/export/parsers/netcdf_parser.py", line 90, in process_sample
    assert np.array_equal(get_grid_points(da_fs[1]), get_grid_points(da_fs[0])), (
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: Grid points between forecast steps are not consistent.Check that inference was not performed with masking

What are the steps to reproduce the bug?

I'm on the updated develop branch on Santis

Hedgedoc link to logs and more information. This ticket is public, do not attach files directly.

No response

[clessig]

@Matteo-Salis : do you have max_num_targets: -1 in inference

[Matteo-Salis]

yes, ERA5 max_num_targets was set to -1

[clessig]

@enssow @iluise : any ideas?