def test_strict_cf_convention(self):
a = self._new_array(standard_name='a',
ancillary_variables='a_stdev b c')
a_stdev = self._new_array(standard_name='a standard_error')
a_uncert = self._new_array(standard_name='a standard_error')
a_count = self._new_array(standard_name='a number_of_observations')
b = self._new_array(standard_name='b')
b_stdev = self._new_array(standard_name='b standard_error')
c = self._new_array(standard_name='c')
ds = xr.Dataset(dict(a=a))
self.assertEqual({}, find_ancillary_var_names(ds, 'a'))
ds = xr.Dataset(
dict(a=a,
a_stdev=a_stdev,
a_uncert=a_uncert,
a_count=a_count,
b=b,
b_stdev=b_stdev,
c=c))
self.assertEqual({
'': {'c', 'b'},
'standard_error': {'a_stdev'}
}, find_ancillary_var_names(ds, 'a'))
def test_xcube_prefix_convention(self):
a = self._new_array()
a_std = self._new_array()
a_count = self._new_array()
b = self._new_array()
b_std = self._new_array()
ds = xr.Dataset(dict(a=a))
self.assertEqual({}, find_ancillary_var_names(ds, 'a'))
ds = xr.Dataset(dict(a=a, a_std=a_std, a_count=a_count, b=b, b_std=b_std))
self.assertEqual({'number_of_observations': {'a_count'},
'standard_error': {'a_std'}},
find_ancillary_var_names(ds, 'a'))
def _get_time_series_for_point(dataset: xr.Dataset,
var_name: str,
point: shapely.geometry.Point,
start_date: np.datetime64 = None,
end_date: np.datetime64 = None,
max_valids: int = None) -> Dict:
var_names = [var_name]
ancillary_var_names = find_ancillary_var_names(dataset,
var_name,
same_shape=True,
same_dims=True)
uncert_var_name = None
if 'standard_error' in ancillary_var_names:
uncert_var_name = next(iter(ancillary_var_names['standard_error']))
var_names.append(uncert_var_name)
ts_ds = timeseries.get_time_series(dataset,
point,
var_names,
start_date=start_date,
end_date=end_date)
if ts_ds is None:
return {'results': []}
return _collect_ts_result(ts_ds,
var_name,
uncert_var_name=uncert_var_name,
count_var_name=None,
max_valids=max_valids)
def _get_time_series_for_point(dataset: xr.Dataset,
var_name: str,
point: shapely.geometry.Point,
agg_methods: Set[str],
start_date: np.datetime64 = None,
end_date: np.datetime64 = None,
max_valids: int = None,
incl_ancillary_vars=False) -> TimeSeries:
var_key = None
if timeseries.AGG_MEAN in agg_methods:
var_key = timeseries.AGG_MEAN
elif timeseries.AGG_MEDIAN in agg_methods:
var_key = timeseries.AGG_MEDIAN
elif timeseries.AGG_MIN in agg_methods or timeseries.AGG_MAX in agg_methods:
var_key = timeseries.AGG_MIN
if not var_key:
raise RuntimeError(
'Aggregation methods must include one of "mean", "median", "min", "max"'
)
roles_to_anc_var_names = dict()
if incl_ancillary_vars:
roles_to_anc_var_name_sets = find_ancillary_var_names(dataset,
var_name,
same_shape=True,
same_dims=True)
for role, roles_to_anc_var_name_sets in roles_to_anc_var_name_sets.items(
):
if role:
roles_to_anc_var_names[role] = roles_to_anc_var_name_sets.pop()
var_names = [var_name] + list(set(roles_to_anc_var_names.values()))
time_series_ds = timeseries.get_time_series(dataset,
point,
var_names,
start_date=start_date,
end_date=end_date,
cube_asserted=True)
if time_series_ds is None:
return []
key_to_var_names = {var_key: var_name}
for role, anc_var_name in roles_to_anc_var_names.items():
key_to_var_names[role] = anc_var_name
return _collect_timeseries_result(time_series_ds,
key_to_var_names,
max_valids=max_valids)
def _get_time_series_for_geometry(dataset: xr.Dataset,
var_name: str,
geometry: shapely.geometry.base.BaseGeometry,
start_date: np.datetime64 = None,
end_date: np.datetime64 = None,
include_count=True,
include_stdev=False,
max_valids: int = None) -> Dict:
if isinstance(geometry, shapely.geometry.Point):
return _get_time_series_for_point(dataset,
var_name,
geometry,
start_date=start_date,
end_date=end_date)
agg_methods = set()
agg_methods.add('mean')
if include_stdev:
agg_methods.add('std')
if include_count:
agg_methods.add('count')
ts_ds = timeseries.get_time_series(dataset,
geometry, [var_name],
agg_methods=agg_methods,
start_date=start_date,
end_date=end_date,
cube_asserted=True)
if ts_ds is None:
return {'results': []}
ancillary_var_names = find_ancillary_var_names(ts_ds, var_name)
uncert_var_name = None
if 'standard_error' in ancillary_var_names:
uncert_var_name = next(iter(ancillary_var_names['standard_error']))
count_var_name = None
if 'number_of_observations' in ancillary_var_names:
count_var_name = next(
iter(ancillary_var_names['number_of_observations']))
return _collect_ts_result(ts_ds,
var_name,
uncert_var_name=uncert_var_name,
count_var_name=count_var_name,
max_valids=max_valids)