def test_some_lookups(self):
x = cmip6.CMIP6_CVs()
self.assertEqual('NCC',
x.lookup('NorCPM1', 'source_id', 'institution_id'))
self.assertCountEqual([
'NorCPM1', 'NorESM2-LMEC', 'NorESM2-HH', 'NorESM1-F', 'NorESM2-MH',
'NorESM2-LM', 'NorESM2-MM', 'NorESM2-LME'
], x.lookup('NCC', 'institution_id', 'source_id'))
def test_table_id_lookup(self):
x = cmip6.CMIP6_CVs()
self.assertEqual(x.lookup('AERmon', 'table_id', 'frequency'),
set([dt_freq('mon')]))
self.assertEqual(x.lookup('AERmon', 'table_id', 'table_freq'),
set(['mon']))
self.assertEqual(
x.lookup(dt_freq('mon'), 'frequency', 'table_id'),
set([
'EmonZ', 'AERmon', 'SImon', 'Amon', 'CFmon', 'Omon', 'ImonGre',
'Emon', 'ImonAnt', 'Lmon', 'LImon', 'Oclim', 'AERmonZ'
]))
self.assertEqual(
x.lookup('mon', 'table_freq', 'table_id'),
set([
'EmonZ', 'AERmon', 'SImon', 'Amon', 'CFmon', 'Omon', 'ImonGre',
'Emon', 'ImonAnt', 'Lmon', 'LImon', 'Oclim', 'AERmonZ'
]))
def __init__(self, case_dict, DateFreqMixin=None):
# set root_dir
# from experiment and model, determine institution and mip
# set realization code = 'r1i1p1f1' unless specified
cmip = cmip6.CMIP6_CVs()
if 'activity_id' not in case_dict:
if 'experiment_id' in case_dict:
key = case_dict['experiment_id']
elif 'experiment' in case_dict:
key = case_dict['experiment']
else:
raise Exception("Can't determine experiment.")
self.experiment_id = key
self.activity_id = cmip.lookup(key, 'experiment_id', 'activity_id')
if 'institution_id' not in case_dict:
if 'source_id' in case_dict:
key = case_dict['source_id']
elif 'model' in case_dict:
key = case_dict['model']
else:
raise Exception("Can't determine model/source.")
self.source_id = key
self.institution_id = cmip.lookup(key, 'source_id', 'institution_id')
if 'member_id' not in case_dict:
self.member_id = 'r1i1p1f1'
case_dict['CASE_ROOT_DIR'] = os.path.join(
self._cmip6_root, self.activity_id, self.institution_id,
self.source_id, self.experiment_id, self.member_id)
# assign explicitly else linter complains
self.data_freq = None
self.table_id = None
self.grid_label = None
self.version_date = None
super(Gfdlcmip6abcDataManager,
self).__init__(case_dict, DateFreqMixin=cmip6.CMIP6DateFrequency)
if 'data_freq' in self.__dict__:
self.table_id = cmip.table_id_from_freq(self.data_freq)
def test_is_in_cv(self):
x = cmip6.CMIP6_CVs()
self.assertTrue(x.is_in_cv('table_id', 'IyrGre'))
def __post_init__(self, log=_log, model=None, experiment=None):
super(CMIP6DataSourceAttributes, self).__post_init__(log=log)
config = core.ConfigManager()
cv = cmip6.CMIP6_CVs()
def _init_x_from_y(source, dest):
if not getattr(self, dest, ""):
try:
source_val = getattr(self, source, "")
if not source_val:
raise KeyError()
dest_val = cv.lookup_single(source_val, source, dest)
log.debug("Set %s='%s' based on %s='%s'.", dest, dest_val,
source, source_val)
setattr(self, dest, dest_val)
except KeyError:
log.debug("Couldn't set %s from %s='%s'.", dest, source,
source_val)
setattr(self, dest, "")
if not self.CASE_ROOT_DIR and config.CASE_ROOT_DIR:
log.debug("Using global CASE_ROOT_DIR = '%s'.",
config.CASE_ROOT_DIR)
self.CASE_ROOT_DIR = config.CASE_ROOT_DIR
# verify case root dir exists
if not os.path.isdir(self.CASE_ROOT_DIR):
log.critical("Data directory CASE_ROOT_DIR = '%s' not found.",
self.CASE_ROOT_DIR)
util.exit_handler(code=1)
# should really fix this at the level of CLI flag synonyms
if model and not self.source_id:
self.source_id = model
if experiment and not self.experiment_id:
self.experiment_id = experiment
# validate non-empty field values
for field in dataclasses.fields(self):
val = getattr(self, field.name, "")
if not val:
continue
try:
if not cv.is_in_cv(field.name, val):
log.error((
"Supplied value '%s' for '%s' is not recognized by "
"the CMIP6 CV. Continuing, but queries will probably fail."
), val, field.name)
except KeyError:
# raised if not a valid CMIP6 CV category
continue
# currently no inter-field consistency checks: happens implicitly, since
# set_experiment will find zero experiments.
# Attempt to determine first few fields of DRS, to avoid having to crawl
# entire DRS structure
_init_x_from_y('experiment_id', 'activity_id')
_init_x_from_y('source_id', 'institution_id')
_init_x_from_y('institution_id', 'source_id')
# TODO: multi-column lookups
# set CATALOG_DIR to be further down the hierarchy if possible, to
# avoid having to crawl entire DRS strcture; CASE_ROOT_DIR remains the
# root of the DRS hierarchy
new_root = self.CASE_ROOT_DIR
for drs_attr in ("activity_id", "institution_id", "source_id",
"experiment_id"):
drs_val = getattr(self, drs_attr, "")
if not drs_val:
break
new_root = os.path.join(new_root, drs_val)
if not os.path.isdir(new_root):
log.error("Data directory '%s' not found; starting crawl at '%s'.",
new_root, self.CASE_ROOT_DIR)
self.CATALOG_DIR = self.CASE_ROOT_DIR
else:
self.CATALOG_DIR = new_root
