`swaths` dataset missing in IW SLC example

[agrouaze]

issue appers with the compilation of the notebook examples/xsar_tops_slc.ipynb

traceback

KeyError                                  Traceback (most recent call last)
/opt/conda-envs/dev/lib/python3.10/site-packages/xarray/core/dataset.py in ?(self, name)
   1447             variable = self._variables[name]
   1448         except KeyError:
-> 1449             _, name, variable = _get_virtual_variable(self._variables, name, self.dims)
   1450 

KeyError: 'swaths'

During handling of the above exception, another exception occurred:

KeyError                                  Traceback (most recent call last)
/tmp/ipykernel_1620036/3271660200.py in ?()
----> 1 s1ds.add_high_resolution_variables()
      2 s1ds.dataset

/home1/datahome/agrouaze/sources/git/xsar/src/xsar/sentinel1_dataset.py in ?(self, luts, patch_variable, skip_variables, load_luts, lazy_loading)
    500             self._dataset = self._dataset.merge(
    501                 self._load_from_geoloc(geoloc_vars, lazy_loading=lazy_loading))
    502 
    503 
--> 504             self.add_swath_number()
    505 
    506             if self.apply_recalibration:
    507                 self.add_gains(config["auxiliary_names"][self.sar_meta.short_name.split(":")[-2][0:3]][self.aux_config_name]["AUX_CAL"],

/home1/datahome/agrouaze/sources/git/xsar/src/xsar/sentinel1_dataset.py in ?(self)
    556         flag_tab = xr.DataArray(np.zeros_like(self._dataset.elevation, dtype=int), coords={
    557                                 'line': self._dataset.coords["line"], 'sample': self._dataset.coords["sample"]})
    558 
    559         # Supposons que dataset.swaths ait 45 éléments comme mentionné
--> 560         for i in range(len(self.datatree['swath_merging'].ds['swaths'])):
    561             y_min, y_max = self.datatree['swath_merging']['firstAzimuthLine'][
    562                 i], self.datatree['swath_merging']['lastAzimuthLine'][i]
    563             x_min, x_max = self.datatree['swath_merging']['firstRangeSample'][

/opt/conda-envs/dev/lib/python3.10/site-packages/datatree/datatree.py in ?(self, key)
    167     def __getitem__(self, key) -> DataArray:
    168         # TODO call the `_get_item` method of DataTree to allow path-like access to contents of other nodes
    169         # For now just call Dataset.__getitem__
--> 170         return Dataset.__getitem__(self, key)

/opt/conda-envs/dev/lib/python3.10/site-packages/xarray/core/dataset.py in ?(self, key)
   1541         """
   1542         if utils.is_dict_like(key):
   1543             return self.isel(**key)
   1544         if utils.hashable(key):
-> 1545             return self._construct_dataarray(key)
   1546         if utils.iterable_of_hashable(key):
   1547             return self._copy_listed(key)
   1548         raise ValueError(f"Unsupported key-type {type(key)}")

/opt/conda-envs/dev/lib/python3.10/site-packages/xarray/core/dataset.py in ?(self, name)
   1445 
   1446         try:
   1447             variable = self._variables[name]
   1448         except KeyError:
-> 1449             _, name, variable = _get_virtual_variable(self._variables, name, self.dims)
   1450 
   1451         needed_dims = set(variable.dims)
   1452 

/opt/conda-envs/dev/lib/python3.10/site-packages/xarray/core/dataset.py in ?(variables, key, dim_sizes)
    210         raise KeyError(key)
    211 
    212     split_key = key.split(".", 1)
    213     if len(split_key) != 2:
--> 214         raise KeyError(key)
    215 
    216     ref_name, var_name = split_key
    217     ref_var = variables[ref_name]

KeyError: 'swaths'

[agrouaze]

@vincelhx 2 questions:

1. I can't see where in the xsar source code the swaths dataset is defined in the datatree, is there a piece of code missing?
2. Are you OK with the fact that such swaths dataset aimaing at giving information on the swath merging, is only relevant for GRD product?

[agrouaze]

I found explanation for point 1.: https://github.com/umr-lops/xarray-safe-s1/blob/2fc774caf1281e4e402b675b273f046fb087720b/safe_s1/reader.py#L106 and https://github.com/umr-lops/xarray-safe-s1/blob/2fc774caf1281e4e402b675b273f046fb087720b/safe_s1/reader.py#L515 For is means that dataset['swath_merging'] is supposed to be present only for GRD. This security was not added in xsar so far.

[agrouaze]

The reason of the bug is now understood,GRD SAFE path given with a trailing slash leads to self.product='XXX' then swath_merging dataset is an empty xarray.Dataset.

[agrouaze]

fixed with https://github.com/umr-lops/xarray-safe-s1/pull/28