def setUp(self):
self.preprocessor = preprocessor.CMIP5Preprocessor('.' + os.sep,
'test_file',
'nc',
'model',
'experiment',
institute='MPI',
mip='Amon',
realm='atmos')
def _get_ensemble_filename(self, the_variable, mip, realm):
"""
get filename of ensemble mean file
if required, then all pre-processing steps are done
Parameters
----------
the_variable : str
variable name to be processed
Returns
-------
returns filename of file with multi-ensemble means
"""
# use model parser to generate a list of available institutes and
# models from data directory
data_dir = self.data_dir
if data_dir[-1] != os.sep:
data_dir += os.sep
CMP = preprocessor.CMIP5ModelParser(self.data_dir)
model_list = CMP.get_all_models()
# model name in configuration file is assumed to be INSTITUTE:MODEL
institute = self.model.split(':')[0]
model = self.model.split(':')[1]
# TODO why is the institute not in the model output name ???
output_file = get_temporary_directory(
) + the_variable + '_' + mip + '_' + model + '_' + self.experiment + '_ensmean.nc'
if institute not in model_list.keys():
raise ValueError('Data for this institute is not existing: %s' %
institute)
# do preprocessing of data from multiple ensembles if file
# already existing, then no processing is done
C5PP = preprocessor.CMIP5Preprocessor(data_dir,
output_file,
the_variable,
model,
self.experiment,
institute=institute,
mip=mip,
realm=realm)
# calculate the ensemble mean and store as file
# also the STDV is calculated on the fly calculated
# resulting filenames are available by C5PP.outfile_ensmean and C5PP.outfile_ensstd
C5PP.ensemble_mean(delete=False,
start_time=self.start_time,
stop_time=self.stop_time)
return C5PP.outfile_ensmean
def main():
from pycmbs.benchmarking import preprocessor
import datetime as dt
import sys
if len(sys.argv) == 3:
the_experiment = sys.argv[1]
the_variable = sys.argv[2]
else:
# specifiy experiment here
the_experiment = 'historical'
the_variable = 'rsds'
print('Experiment: %s' % the_experiment)
print('Variable: %s' % the_variable)
output_dir = '/data/share/mpiles/TRS/PROJECT_RESULTS/EvaClimod/CMIP5_RAWDATA_NEW/radiation/dummy_out/' + the_experiment + '_' + the_variable + '/'
data_dir = '/data/share/mpiles/TRS/PROJECT_RESULTS/EvaClimod/CMIP5_RAWDATA_NEW/radiation/'
# init parser that returns a list of institutes and models
CP = preprocessor.CMIP5ModelParser(data_dir)
model_list = CP.get_all_models()
# perform for each institute and model the calculation of ensemble means etc.
for institute in model_list.keys():
for model in model_list[institute]:
output_file = output_dir + the_variable + '_Amon_' + model + '_' + the_experiment + '_ensmean.nc'
E = preprocessor.CMIP5Preprocessor(data_dir,
output_file,
the_variable,
model,
the_experiment,
institute=institute)
E.get_ensemble_files()
E.ensemble_mean(delete=False,
start_time=dt.datetime(1979, 1, 1),
stop_time=dt.datetime(2012, 12, 31))