Enhance metadata of produce files

Author: AntSimi

src/py_eddy_tracker/__init__.py

@@ -99,17 +99,8 @@ def parse_args(self, *args, **kwargs):
             units='days since 1950-01-01 00:00:00',
             calendar='proleptic_gregorian',
             axis='T',
-            longname='time of gridded file',
-            description='date of this observation',
-            )
-        ),
-    ocean_time=dict(
-        attr_name=None,
-        nc_name='ocean_time',
-        nc_type='float64',
-        nc_dims=('obs',),
-        nc_attr=dict(
-            units='ROMS ocean_time (seconds)',
+            longname='Time',
+            description='Date of this observation',
             )
         ),
     type_cyc=dict(
@@ -121,7 +112,7 @@ def parse_args(self, *args, **kwargs):
         nc_attr=dict(
             longname='cyclonic',
             units='boolean',
-            description='cyclonic -1; anti-cyclonic +1',
+            description='Cyclonic -1; anti-cyclonic +1',
             )
         ),
     segment_size=dict(
@@ -147,13 +138,15 @@ def parse_args(self, *args, **kwargs):
         ),
     virtual=dict(
         attr_name=None,
-        nc_name='virtual',
+        nc_name='observation_flag',
+        old_nc_name=['virtual'],
         nc_type='byte',
         nc_dims=('obs',),
         nc_attr=dict(
             longname='virtual_position',
             units='boolean',
-            description='Virtual observation: 0 for real',
+            description='Flag indicating if the value is interpolated between two'
+                        ' observations or not (0: observed, 1: interpolated)',
             )
         ),
     cost_association=dict(
@@ -176,7 +169,7 @@ def parse_args(self, *args, **kwargs):
         nc_attr=dict(
             units='degrees_east',
             axis='X',
-            description='observation longitude',
+            description='Observation longitude',
             longname='longitude of measurement',
             standard_name='longitude',
             )
@@ -193,7 +186,7 @@ def parse_args(self, *args, **kwargs):
             axis='Y',
             longname='latitude of measurement',
             standard_name='latitude',
-            description='observation latitude',
+            description='Observation latitude',
             )
         ),
     lon_max=dict(
@@ -206,7 +199,7 @@ def parse_args(self, *args, **kwargs):
         nc_attr=dict(
             units='degrees_east',
             axis='X',
-            description='observation longitude',
+            description='Observation longitude',
             longname='longitude of amplitude max',
             standard_name='longitude',
             )
@@ -221,7 +214,7 @@ def parse_args(self, *args, **kwargs):
         nc_attr=dict(
             units='degrees_north',
             axis='Y',
-            description='observation latitude',
+            description='Observation latitude',
             longname='latitude of amplitude max',
             standard_name='latitude',
             )
@@ -237,7 +230,7 @@ def parse_args(self, *args, **kwargs):
         nc_attr=dict(
             longname='amplitude',
             units='m',
-            description='magnitude of the height difference between the extremum of ADT within '
+            description='Magnitude of the height difference between the extremum of ADT within '
                         'the eddy and the ADT around the contour defining the eddy perimeter',
             )
         ),
@@ -252,8 +245,7 @@ def parse_args(self, *args, **kwargs):
         nc_attr=dict(
             longname='maximum circum-averaged speed',
             units='m/s',
-            description='average speed of the contour defining the '
-                        'speed radius',
+            description='Average speed of the contour defining the radius scale “speed_radius”',
             )
         ),
     uavg_profile=dict(
@@ -266,7 +258,7 @@ def parse_args(self, *args, **kwargs):
         nc_attr=dict(
             longname='radial profile of uavg',
             units='m/s',
-            description='profile speed average values from effective contour inwards to smallest inner contour',
+            description='Speed average values from effective contour inwards to smallest inner contour',
         )
     ),
     i=dict(
@@ -276,7 +268,7 @@ def parse_args(self, *args, **kwargs):
         nc_dims=('obs',),
         nc_attr=dict(
             longname='longitude index in the grid of the detection',
-            description='longitude index in the grid of the detection',
+            description='Longitude index in the grid of the detection',
             )
         ),
     j=dict(
@@ -286,7 +278,7 @@ def parse_args(self, *args, **kwargs):
         nc_dims=('obs',),
         nc_attr=dict(
             longname='latitude index in the grid of the detection',
-            description='latitude index in the grid of the detection',
+            description='Latitude index in the grid of the detection',
             )
         ),
     eke=dict(
@@ -297,7 +289,7 @@ def parse_args(self, *args, **kwargs):
         nc_attr=dict(
             longname='sum EKE within contour Ceff',
             units='m^2/s^2',
-            description='sum of eddy kinetic energy within contour '
+            description='Sum of eddy kinetic energy within contour '
                         'defining the effective radius',
             )
         ),
@@ -312,7 +304,7 @@ def parse_args(self, *args, **kwargs):
         nc_attr=dict(
             longname='effective radius scale',
             units='m',
-            description='radius of a circle whose area is equal to that enclosed by the effective contour',
+            description='Radius of a circle whose area is equal to that enclosed by the effective contour',
             )
         ),
     radius_s=dict(
@@ -326,7 +318,7 @@ def parse_args(self, *args, **kwargs):
         nc_attr=dict(
             longname='speed radius scale',
             units='m',
-            description='radius of a circle whose area is equal to that enclosed by the contour of maximum circum-average speed',
+            description='Radius of a circle whose area is equal to that enclosed by the contour of maximum circum-average speed',
             )
         ),
     track=dict(
@@ -337,7 +329,7 @@ def parse_args(self, *args, **kwargs):
         nc_attr=dict(
             longname='track number',
             units='ordinal',
-            description='eddy identification number',
+            description='Eddy identification number',
             )
         ),
     n=dict(
@@ -349,7 +341,7 @@ def parse_args(self, *args, **kwargs):
         nc_attr=dict(
             longname='observation number',
             units='ordinal',
-            description='observation sequence number, days from eddy start',
+            description='Observation sequence number, days from eddy first detection',
             )
         ),
     contour_lon_e=dict(
@@ -363,7 +355,7 @@ def parse_args(self, *args, **kwargs):
         nc_dims=('obs', 'NbSample'),
         nc_attr=dict(
             longname='effective contour longitudes',
-            description='longitudes of effective contour',
+            description='Longitudes of effective contour',
             units='degrees_east',
             axis='X',
             )
@@ -378,7 +370,7 @@ def parse_args(self, *args, **kwargs):
         nc_dims=('obs', 'NbSample'),
         nc_attr=dict(
             longname='effective contour latitudes',
-            description='latitudes of effective contour',
+            description='Latitudes of effective contour',
             units='degrees_east',
             axis='X',
             )
@@ -394,7 +386,7 @@ def parse_args(self, *args, **kwargs):
         nc_dims=('obs', 'NbSample'),
         nc_attr=dict(
             longname='speed contour longitudes',
-            description='longitudes of speed contour',
+            description='Longitudes of speed contour',
             units='degrees_east',
             axis='X',
             )
@@ -409,7 +401,7 @@ def parse_args(self, *args, **kwargs):
         nc_dims=('obs', 'NbSample'),
         nc_attr=dict(
             longname='speed contour latitudes',
-            description='latitudes of speed contour',
+            description='Latitudes of speed contour',
             units='degrees_east',
             axis='X',
             )
@@ -480,11 +472,14 @@ def parse_args(self, *args, **kwargs):
         ),
     nb_contour_selected=dict(
         attr_name=None,
-        nc_name='nb_contour_selected',
+        nc_name='num_contours',
+        old_nc_name=['nb_contour_selected'],
         nc_type='u2',
         nc_dims=('obs',),
         nc_attr=dict(
-            units='1',
+            longname='number of contour',
+            units='ordinal',
+            description='Number of contour selected for this eddy',
             )
         ),
     )

src/py_eddy_tracker/observations.py

@@ -29,14 +29,14 @@
 """
 from numpy import zeros, empty, absolute, arange, where, unique, \
     ma, concatenate, cos, radians, isnan, ones, ndarray, \
-    interp, floor
+    interp, floor, array
 from netCDF4 import Dataset
 from .generic import distance_grid, distance
 from shapely.geometry import Polygon
 from shapely.geos import TopologicalError
 from . import VAR_DESCR, VAR_DESCR_inv
 import logging
-from datetime import datetime
+from datetime import datetime, timedelta
 from numba import njit, types as numba_types
 
 
@@ -627,8 +627,9 @@ def to_netcdf(self, handler):
             handler.array_variables = ','.join(self.array_variables)
         # Iter on variables to create:
         fields = [field[0] for field in self.observations.dtype.descr]
-        fields.sort()
-        for ori_name in fields:
+        fields_ = array([VAR_DESCR[field[0]]['nc_name'] for field in self.observations.dtype.descr])
+        i = fields_.argsort()
+        for ori_name in array(fields)[i]:
             # Patch for a transition
             name = ori_name
             #
@@ -643,20 +644,7 @@ def to_netcdf(self, handler):
                 scale_factor=VAR_DESCR[name].get('scale_factor', None),
                 add_offset=VAR_DESCR[name].get('add_offset', None)
             )
-        # Add cyclonic information
-        if self.sign_type is not None:
-            self.create_variable(
-                handler,
-                dict(varname=VAR_DESCR['type_cyc']['nc_name'],
-                     datatype=VAR_DESCR['type_cyc']['nc_type'],
-                     dimensions=VAR_DESCR['type_cyc']['nc_dims']),
-                VAR_DESCR['type_cyc']['nc_attr'],
-                self.sign_type)
-
-        handler.Metadata_Conventions = 'Unidata Dataset Discovery v1.0'
-        handler.comment = 'Surface product; mesoscale eddies'
-        handler.framework_used = 'https://bitbucket.org/emason/py-eddy-tracker'
-        handler.standard_name_vocabulary ='NetCDF Climate and Forecast (CF) Metadata Convention Standard Name Table'
+        self.set_global_attr_netcdf(handler)
 
     def create_variable(self, handler_nc, kwargs_variable, attr_variable,
                         data, scale_factor=None, add_offset=None):
@@ -692,41 +680,14 @@ def write_netcdf(self, path='./', filename='%(path)s/%(sign_type)s.nc'):
         eddy_size = len(self.observations)
         filename = filename % dict(path=path, sign_type=self.sign_legend, prod_time=datetime.now().strftime('%Y%m%d'))
         logging.info('Store in %s', filename)
-        with Dataset(filename, 'w', format='NETCDF4') as h_nc:
-            logging.info('Create file %s', filename)
-            # Create dimensions
-            logging.debug('Create Dimensions "obs" : %d', eddy_size)
-            h_nc.createDimension('obs', eddy_size)
-            if self.track_array_variables != 0:
-                h_nc.createDimension('NbSample', self.track_array_variables)
-            # Iter on variables to create:
-            for field in self.observations.dtype.descr:
-                name = field[0]
-                logging.debug('Create Variable %s', VAR_DESCR[name]['nc_name'])
-                self.create_variable(
-                    h_nc,
-                    dict(varname=VAR_DESCR[name]['nc_name'],
-                         datatype=VAR_DESCR[name]['output_type'],
-                         dimensions=VAR_DESCR[name]['nc_dims']),
-                    VAR_DESCR[name]['nc_attr'],
-                    self.observations[name],
-                    scale_factor=VAR_DESCR[name].get('scale_factor', None),
-                    add_offset=VAR_DESCR[name].get('add_offset', None)
-                    )
-
-            # Add cyclonic information
-            self.create_variable(
-                h_nc,
-                dict(varname=VAR_DESCR['type_cyc']['nc_name'],
-                     datatype=VAR_DESCR['type_cyc']['nc_type'],
-                     dimensions=VAR_DESCR['type_cyc']['nc_dims']),
-                VAR_DESCR['type_cyc']['nc_attr'],
-                self.sign_type)
-            # Global attr
-            self.set_global_attr_netcdf(h_nc)
+        with Dataset(filename, 'w', format='NETCDF4') as handler:
+            self.to_netcdf(handler)
 
     def set_global_attr_netcdf(self, h_nc):
-        pass
+        h_nc.Metadata_Conventions = 'Unidata Dataset Discovery v1.0'
+        h_nc.comment = 'Surface product; mesoscale eddies'
+        h_nc.framework_used = 'https://bitbucket.org/emason/py-eddy-tracker'
+        h_nc.standard_name_vocabulary ='NetCDF Climate and Forecast (CF) Metadata Convention Standard Name Table'
 
     def display(self, ax, ref=None, **kwargs):
         if ref is None:
@@ -875,7 +836,16 @@ def elements(self):
     def set_global_attr_netcdf(self, h_nc):
         """Set global attr
         """
-        if self.sign_type == -1:
-            h_nc.title = 'Cyclonic'
-        else:
-            h_nc.title = 'Anticyclonic' + ' eddy tracks'
+        h_nc.title = 'Cyclonic' if self.sign_type == -1 else 'Anticyclonic'
+        h_nc.Metadata_Conventions = 'Unidata Dataset Discovery v1.0'
+        h_nc.comment = 'Surface product; mesoscale eddies'
+        h_nc.framework_used = 'https://bitbucket.org/emason/py-eddy-tracker'
+        h_nc.standard_name_vocabulary = 'NetCDF Climate and Forecast (CF) Metadata Convention Standard Name Table'
+        h_nc.date_created = datetime.now().strftime('%Y-%m-%dT%H:%M:%SZ')
+        t = h_nc.variables[VAR_DESCR_inv['j1']]
+        delta = t.max - t.min + 1
+        h_nc.time_coverage_duration = 'P%dD' % delta
+        d_start = datetime(1950, 1, 1) + timedelta(int(t.min))
+        d_end = datetime(1950, 1, 1) + timedelta(int(t.max))
+        h_nc.time_coverage_start = d_start.strftime('%Y-%m-%dT00:00:00Z')
+        h_nc.time_coverage_end = d_end.strftime('%Y-%m-%dT00:00:00Z')