def test_array_param():
# Check that arrays are correctly handled
config_def = {
'creator': ParamReturnCreator(param.Array),
'val': np.arange(1, 6),
}
config_inst = process_config(config_def)
assert np.all(config_inst == np.array([1, 2, 3, 4, 5]))
# Check that non arrays cause errors
config_def['val'] = 10
try:
config_inst = process_config(config_def)
except param.ParamError:
assert True
else:
assert False
config_def['val'] = [1, 2, 3, 4, 5]
try:
config_inst = process_config(config_def)
except param.ParamError:
assert True
else:
assert False
def test_iterable_param():
# Check that iterables are correctly handled
config_def = {
'item1': {
'creator': ParamReturnCreator(param.Iterable),
'val': [1, 2, 3, 4, 5]
},
'item2': {
'creator': ParamReturnCreator(param.Iterable, val_type=np.int),
'val': [1, 2, 3, 4, 5]
},
}
config_inst = process_config(config_def)
assert np.all(np.array(config_inst['item1']) == np.arange(1, 6))
assert np.all(np.array(config_inst['item2']) == np.arange(1, 6))
# Check arrays are considered iterable
config_def['item1']['val'] = np.arange(1, 6)
config_inst = process_config(config_def)
assert np.all(np.array(config_inst['item1']) == np.arange(1, 6))
def test_param_choice():
class ChoiceCreator(creator.base.Creator):
some_val = param.Choice(param.Scalar(int), param.Scalar(float))
def create(self, **kwargs):
return self.param("some_val")
config_def = {
'item1': {
'creator': ChoiceCreator,
'some_val': 10,
},
'item2': {
'creator': ChoiceCreator,
'some_val': 5.0,
},
}
config_inst = process_config(config_def)
assert config_inst['item1'] == 10
assert config_inst['item2'] == 5.0
config_def['item1']['some_val'] = "ABC"
try:
config_inst = process_config(config_def)
except param.ParamError:
assert True
else:
assert False
def test_nested():
config_def = {
'item1': {
'creator': ParamReturnCreator(param.Scalar),
'val': 5,
},
'item2': {
'creator': ParamReturnCreator(param.Scalar),
'val': 10,
},
'item3': {
'creator': ParamReturnCreator(param.Scalar),
'val': {
'creator': ParamReturnCreator(param.Scalar),
'val': 15,
},
},
}
config_inst = process_config(config_def)
assert config_inst['item1'] == 5
assert config_inst['item2'] == 10
assert config_inst['item3'] == 15
def test_array_with_unit_param():
# Check that arrays are correctly handled
config_def = {
'creator': ParamReturnCreator(param.ArrayWithUnit),
'val': rf.ArrayWithUnit_double_1(np.arange(1, 6), "m"),
}
config_inst = process_config(config_def)
def test_callable():
config_def = {
'creator': ParamReturnCreator(param.Scalar),
'val': lambda **kwargs: 10,
}
config_inst = process_config(config_def)
assert config_inst == 10
def test_scalar_with_type():
# Check that scalar params are correctly checked
config_def = {
'creator': ParamReturnCreator(partial(param.Scalar, dtype=str)),
'val': "10",
}
config_inst = process_config(config_def)
assert config_inst == "10"
# Check that non scalars cause errors
config_def['val'] = 10
try:
config_inst = process_config(config_def)
except param.ParamError:
assert True
else:
assert False
def test_scalar_param():
# Check that scalar params are correctly checked
config_def = {
'creator': ParamReturnCreator(param.Scalar),
'val': 10,
}
config_inst = process_config(config_def)
assert config_inst == 10
# Check that non scalars cause errors
config_def['val'] = np.arange(1, 6)
try:
config_inst = process_config(config_def)
except param.ParamError:
assert True
else:
assert False
def config(self, sounding_id):
logging.debug("Loading configuration for sounding: %s" % sounding_id)
config_def = retrieval_config_definition(self.l1b_file, self.met_file,
sounding_id)
config_def['atmosphere']['ground']['child'] = self.ground_type.value
config_inst = process_config(config_def)
config_inst.config_def = config_def
return config_inst
def test_instanceof_param():
class TestClass(object):
pass
# Check that iterables are correctly handled
config_def = {
'creator': ParamReturnCreator(param.InstanceOf, val_type=TestClass),
'val': TestClass(),
}
config_inst = process_config(config_def)
assert isinstance(config_inst, TestClass)
# Force error on type
config_def['val'] = 5
try:
config_inst = process_config(config_def)
except param.ParamError:
assert True
else:
assert False
def test_common_store():
config_def = {
'order': ['common', 'use_common'],
'common': {
'creator': creator.base.SaveToCommon,
'x': 5,
'y': 6,
},
'use_common': {
'creator': AddCreator,
},
}
config_inst = process_config(config_def)
assert config_inst['use_common'] == 11
def test_load_example_config():
data_dir = os.path.realpath('../test/in')
l1b_file = os.path.join(
data_dir,
"oco2_L1bScND_16094a_170711_B7302r_171102090317-selected_ids.h5")
met_file = os.path.join(
data_dir,
"oco2_L2MetND_16094a_170711_B8000r_171017214714-selected_ids.h5")
sounding_id = "2017071110541471"
config_def = oco_config.retrieval_config_definition(
l1b_file, met_file, sounding_id)
config_inst = process_config(config_def)
pprint(config_inst, indent=2)
fm = config_inst.forward_model
atm = config_inst.atmosphere
sv = config_inst.retrieval.state_vector
solver = config_inst.retrieval.solver
def test_object_vector():
config_def = {
'item1': {
'creator': ParamReturnCreator(param.ObjectVector),
'val': rf.vector_double(),
},
'item2': {
'creator': ParamReturnCreator(param.ObjectVector,
vec_type="double"),
'val': rf.vector_double(),
}
}
config_inst = process_config(config_def)
assert isinstance(config_inst['item1'], rf.vector_double)
assert isinstance(config_inst['item2'], rf.vector_double)
def test_bound_params():
class BoundParamCreator(creator.base.Creator):
some_val = param.Choice(param.Scalar(int), param.Scalar(int))
def create(self, **kwargs):
return self.some_val()
config_def = {
'item1': {
'creator': BoundParamCreator,
'some_val': 10,
},
'item2': {
'creator': BoundParamCreator,
'some_val': 5,
},
}
config_inst = process_config(config_def)
assert config_inst['item1'] == 10
assert config_inst['item2'] == 5
'filename': covariance_file,
}
},
'solver_nlls_gsl': {
'creator': creator.retrieval.NLLSSolverGSLLMSDER,
'max_cost_function_calls': 10,
'dx_tol_abs': 1e-5,
'dx_tol_rel': 1e-5,
'g_tol_abs': 1e-5,
},
'solver': {
'creator': creator.retrieval.ConnorSolverMAP,
'max_cost_function_calls': 14,
'threshold': 2.0,
'max_iteration': 7,
'max_divergence': 2,
'max_chisq': 1.4,
'gamma_initial': 10.0,
},
},
}
if __name__ == "__main__":
logging.basicConfig(level=logging.DEBUG)
config_inst = process_config(config_def)
from pprint import pprint
pprint(config_inst, indent=4)