Allow to add field on a dataset

Author: AntSimi

src/py_eddy_tracker/__init__.py

@@ -432,6 +432,19 @@ def parse_args(self, *args, **kwargs):
             longname='effective contour error',
         )
     ),
+    score=dict(
+        attr_name=None,
+        nc_name='score',
+        nc_type='f2',
+        output_type='u1',
+        scale_factor=0.4,
+        nc_dims=('obs',),
+        nc_attr=dict(
+            units='%',
+            description='score',
+            longname='score',
+        )
+    ),
     shape_error_s=dict(
         attr_name=None,
         nc_name='speed_contour_shape_error',

src/py_eddy_tracker/observations/observation.py

@@ -41,6 +41,7 @@
     array,
     empty,
     absolute,
+    concatenate,
 )
 from netCDF4 import Dataset
 from ..generic import distance_grid, distance
@@ -219,6 +220,25 @@ def obs_dimension(cls, handler):
             if candidate in handler.dimensions.keys():
                 return candidate
 
+    def add_fields(self, fields):
+        """
+        Add a new field
+        """
+        nb_obs = self.obs.shape[0]
+        new = self.__class__(
+            size=nb_obs,
+            track_extra_variables=list(concatenate((self.track_extra_variables, fields))),
+            track_array_variables=self.track_array_variables,
+            array_variables=self.array_variables,
+            raw_data=self.raw_data
+        )
+        new.sign_type = self.sign_type
+        for field in self.obs.dtype.descr:
+            logging.debug('Copy of field %s ...', field)
+            var = field[0]
+            new.obs[var] = self.obs[var]
+        return new
+
     @property
     def dtype(self):
         """Return dtype to build numpy array
@@ -368,8 +388,6 @@ def load_from_netcdf(cls, filename, raw_data=False, remove_vars=None):
         with Dataset(filename) as h_nc:
             var_list = list(h_nc.variables.keys())
             if remove_vars is not None:
-                print(var_list)
-                print(remove_vars)
                 var_list = [i for i in var_list if i not in remove_vars]
 
             nb_obs = len(h_nc.dimensions[cls.obs_dimension(h_nc)])