def saveAnomaly(data, new, compute=True):
"""
save deviation to processeddir
"""
filename = generateFileName(data.name, dataset=data.dataset,
processed='anom', suffix='nc')
path = join(processeddir, filename)
if exists(path) and not new:
print(f"{data.name} anomaly already computed")
else:
print(f"Compute {data.name} anomaly")
if compute:
print(f"Compute and save {data.name} anomaly")
anom = computeAnomaly(data)
else:
print(f"Save {data.name} anomaly")
anom = data
anom.name = ''.join([data.name, 'Anom'])
anom.attrs = data.attrs.copy()
anom.attrs['statistic'] = 'Substracted the monthly Mean.'
anom.attrs = _delete_some_attributes(anom.attrs)
anom.to_netcdf(path)
def computeMeanClimatology(data):
"""
Monthly means
"""
filename = generateFileName(data.name, dataset=data.dataset,
processed='meanclim', suffix='nc')
path = join(processeddir, filename)
if not exists(path):
print(f"- Compute {data.name} climatetology")
period = _get_period(data)
print(f"- Data has {period} period")
if reference_period:
meanclim = data.loc['1981-01-01':'2010-12-31']. \
groupby(f'time.{period}').mean(dim="time")
else:
print("Use the entire time series for the Mean Climatology")
meanclim = data.groupby(f'time.{period}').mean(dim="time")
meanclim.to_netcdf(path)
else:
print(f"- Read {data.name} climatetology")
meanclim = xr.open_dataarray(path)
return meanclim
def save(self):
self.data = pd.DataFrame(
{'global_transitivity': self.global_transitivity,
'avelocal_transmissivity': self.avglocal_transitivity,
'fraction_clusters_size_2': self.frac_cluster_size2,
'fraction_clusters_size_3': self.frac_cluster_size3,
'fraction_clusters_size_5': self.frac_cluster_size5,
'fraction_giant_component': self.frac_giant,
'average_path_length': self.avg_path_length,
'hamming_distance': self.hamming_distance,
'corrected_hamming_distance': self.corrected_hamming_distance,
'threshold': self.threshold_value,
'edge_density': self.edge_density_value
})
filename = generateFileName(self.variable,
self.dataset,
processed=self.processed,
suffix='csv')
filename = '-'.join(['network_metrics', filename])
if self.threshold is not None:
# TODO: dynamic naming depending on the methods used
pass
elif self.edge_density is not None:
pass
self.data.to_csv(join(processeddir, filename))
def save(self, extension='', filename=None):
"""
Saves the first three pca components to a csv-file.
"""
# save data to first day of month
save_index = self.time.to_index()
pca1 = pd.Series(np.matmul(self.EOFarr, self.components_[0, :]),
index=save_index)
pca2 = pd.Series(np.matmul(self.EOFarr, self.components_[1, :]),
index=save_index)
pca3 = pd.Series(np.matmul(self.EOFarr, self.components_[2, :]),
index=save_index)
self.df = pd.DataFrame({'pca1': pca1, 'pca2': pca2, 'pca3': pca3})
if filename is None:
filename = generateFileName(self.variable,
self.dataset,
''.join((self.processed, extension)),
suffix='csv')
else:
filename = '.'.join((filename, 'csv'))
filename = '-'.join(['pca', filename])
self.df.to_csv(join(processeddir, filename))
def read_statistic(self, statistic, variable, dataset='', processed=''):
filename = generateFileName(variable, dataset,
processed=processed, suffix="csv")
filename = '-'.join([statistic, filename])
data = pd.read_csv(join(processeddir, filename),
index_col=0, parse_dates=True)
self._check_dates(data, f"{variable} - {statistic}" )
return data.loc[self.startdate:self.enddate]
def toProcessedDir(data, new):
"""
Save the basic data to the processeddir.
"""
filename = generateFileName(data.name, dataset=data.dataset, suffix='nc')
path = join(processeddir, filename)
if exists(path) and not new:
print(f"{data.name} already saved in post directory")
else:
print(f"save {data.name} in post directory")
data.to_netcdf(path)
def read_netcdf(self, variable, dataset='', processed='', chunks=None):
"""
wrapper for xarray.open_dataarray.
:param variable: the name of the variable
:param dataset: the name of the dataset
:param processed: the postprocessing that was applied
:param chunks: same as for xarray.open_dataarray
"""
filename = generateFileName(variable,
dataset,
processed=processed,
suffix="nc")
data = xr.open_dataarray(join(processeddir, filename), chunks=chunks)
regrided = ['GODAS', 'ERSSTv5', 'ORAS4', 'NODC', 'NCAR']
if processed == 'meanclim':
return data
else:
self._check_dates(data, f'{filename[:-3]}')
if dataset not in regrided and dataset != 'ORAP5' and dataset != 'GFDL-CM3':
return data.loc[self.startdate:self.enddate,
self.lat_max:self.lat_min,
self.lon_min:self.lon_max]
elif dataset in regrided or dataset == 'GFDL-CM3':
return data.loc[self.startdate:self.enddate,
self.lat_min:self.lat_max,
self.lon_min:self.lon_max]
elif dataset == 'ORAP5':
return data.loc[self.startdate:self.enddate, :, :].where(
(data.nav_lat > self.lat_min) &
(data.nav_lat < self.lat_max) &
(data.nav_lon > self.lon_min) &
(data.nav_lon < self.lon_max),
drop=True)
def saveNormAnomaly(data, new):
"""
save deviation to processeddir
"""
filename = generateFileName(data.name, dataset=data.dataset,
processed='normanom', suffix='nc')
path = join(processeddir, filename)
if exists(path) and not new:
print(f"{data.name} normed anomaly already computed")
else:
print(f"Compute {data.name} normed anomaly")
normanom = computeNormAnomaly(data)
normanom.name = ''.join([data.name, 'NormAnom'])
normanom.attrs = data.attrs.copy()
normanom.attrs['statistic'] = 'Substracted the monthly Mean.\
Divided by the Monthly standard deviation'
normanom.attrs = _delete_some_attributes(normanom.attrs)
normanom.to_netcdf(path)
