def make_features(
values,
params,
fallback_func=param_processor.choose_default_placeholder,
converters={},
convert_fallback=True,
):
"""Return an iterator of the values for the parameters in sorted order with the given fallback function.
If passed, converters must map funcs to functions from (value, *args) -> new_value which will be run
on the resulting value for that func (but only on fallbacks if convert_fallback)."""
for name, (func, args, kwargs) in sorted_items(params):
# determine feature
fallback = False
_coconut_match_to = values
_coconut_match_check = False
if _coconut.isinstance(_coconut_match_to, _coconut.abc.Mapping):
_coconut_match_temp_0 = _coconut_match_to.get(
name, _coconut_sentinel)
if _coconut_match_temp_0 is not _coconut_sentinel:
feature = _coconut_match_temp_0
_coconut_match_check = True
if _coconut_match_check:
pass
else:
_coconut_match_to = kwargs
_coconut_match_check = False
if _coconut.isinstance(_coconut_match_to, _coconut.abc.Mapping):
_coconut_match_temp_0 = _coconut_match_to.get(
"placeholder_when_missing", _coconut_sentinel)
if _coconut_match_temp_0 is not _coconut_sentinel:
placeholder_value = _coconut_match_temp_0
_coconut_match_check = True
if _coconut_match_check:
feature = placeholder_value
else:
fallback = True
feature = fallback_func(name, func, *args, **kwargs)
# run converters
if not fallback or convert_fallback:
_coconut_match_to = converters
_coconut_match_check = False
if _coconut.isinstance(_coconut_match_to, _coconut.abc.Mapping):
_coconut_match_temp_0 = _coconut_match_to.get(
func, _coconut_sentinel)
if _coconut_match_temp_0 is not _coconut_sentinel:
converter_func = _coconut_match_temp_0
_coconut_match_check = True
if _coconut_match_check:
feature = converter_func(feature, *args)
yield feature
def __init__(self, examples, params, base_estimator="gp", **options):
self.init_fallback_backend()
if not examples:
self.current_values = {}
return
data_points, losses = split_examples(examples, params)
dimensions = [
create_dimension(name, func, *args)
for name, (func, args, kwargs) in sorted_items(params)
]
if isinstance(base_estimator, str):
base_estimator = py_str(base_estimator)
optimizer = Optimizer(dimensions, base_estimator, **options)
optimizer.tell(data_points, losses)
current_point = optimizer.ask()
self.current_values = make_values(params, current_point)
def __init__(self,
examples,
params,
algo=tpe.suggest,
rstate=np.random.RandomState(),
show_progressbar=False,
**options):
self.init_fallback_backend()
if not examples:
self.current_values = {}
return
space = (as_apply)(dict(
((name), (create_space(name, func, *args)))
for name, (func, args, kwargs) in sorted_items(params)))
domain = Domain(self.set_current_values, space)
trial_list = examples_to_trials(examples, params)
trials = Trials()
trials.insert_trial_docs(trial_list)
self.fmin_iter = FMinIter(algo,
domain,
trials,
rstate,
show_progressbar=show_progressbar,
**options)
# run one iteration of hyperparameter optimization, with values saved
# to the self.set_current_values callback passed to Domain
next(self.fmin_iter)
assert self.current_values is not None, self.current_values
assert set(
self.current_values.keys()) == set(params), self.current_values
def setup_backend(self,
params,
algo=tpe.suggest,
rstate=np.random.RandomState(),
show_progressbar=False,
**options):
"""Special method to initialize the backend from params."""
self.params = params
space = (as_apply)(dict(
((name), (create_space(name, func, *args)))
for name, (func, args, kwargs) in sorted_items(params)))
domain = Domain(self.set_current_values, space)
self.trials = Trials()
self.fmin_iter = FMinIter(algo,
domain,
self.trials,
rstate,
show_progressbar=show_progressbar,
**options)
def create_dimensions(params):
"""Construct the full optimization space for the given parameters."""
return [
create_space(name, func, *args)
for name, (func, args, kwargs) in sorted_items(params)
]
def get_names_and_features(
values,
params,
fallback_func=param_processor.choose_default_placeholder,
converters={},
convert_fallback=True,
):
"""Return an iterator of (name, feature) for the parameters in sorted order with the given fallback function.
If passed, converters must map funcs to functions from (value, *args) -> new_value which will be run
on the resulting value for that func (but only on fallbacks if convert_fallback)."""
for name, (func, args, kwargs) in sorted_items(params):
# determine feature
fallback = False
_coconut_match_to_1 = values
_coconut_match_check_3 = False
_coconut_match_set_name_feature = _coconut_sentinel
if _coconut.isinstance(_coconut_match_to_1, _coconut.abc.Mapping):
_coconut_match_temp_12 = _coconut_match_to_1.get(
name, _coconut_sentinel)
if _coconut_match_temp_12 is not _coconut_sentinel:
_coconut_match_set_name_feature = _coconut_match_temp_12
_coconut_match_check_3 = True
if _coconut_match_check_3:
if _coconut_match_set_name_feature is not _coconut_sentinel:
feature = _coconut_match_set_name_feature
if _coconut_match_check_3:
pass
else:
_coconut_match_to_0 = kwargs
_coconut_match_check_2 = False
_coconut_match_set_name_placeholder_value = _coconut_sentinel
if _coconut.isinstance(_coconut_match_to_0, _coconut.abc.Mapping):
_coconut_match_temp_11 = _coconut_match_to_0.get(
"placeholder_when_missing", _coconut_sentinel)
if _coconut_match_temp_11 is not _coconut_sentinel:
_coconut_match_set_name_placeholder_value = _coconut_match_temp_11
_coconut_match_check_2 = True
if _coconut_match_check_2:
if _coconut_match_set_name_placeholder_value is not _coconut_sentinel:
placeholder_value = _coconut_match_set_name_placeholder_value
if _coconut_match_check_2:
feature = placeholder_value
else:
fallback = True
feature = fallback_func(name, func, *args, **kwargs)
# run converters
if not fallback or convert_fallback:
_coconut_match_to_2 = converters
_coconut_match_check_4 = False
_coconut_match_set_name_converter_func = _coconut_sentinel
if _coconut.isinstance(_coconut_match_to_2, _coconut.abc.Mapping):
_coconut_match_temp_13 = _coconut_match_to_2.get(
func, _coconut_sentinel)
if _coconut_match_temp_13 is not _coconut_sentinel:
_coconut_match_set_name_converter_func = _coconut_match_temp_13
_coconut_match_check_4 = True
if _coconut_match_check_4:
if _coconut_match_set_name_converter_func is not _coconut_sentinel:
converter_func = _coconut_match_set_name_converter_func
if _coconut_match_check_4:
feature = converter_func(feature, *args)
yield name, feature
def __init__(self, params):
self.params = params
self.dim = len(params)
self.got_values = None
# lower and upper bounds; len must match self.dim
self.lb = []
self.ub = []
# lists of indices that are discrete vs. continuous
self.int_var = []
self.cont_var = []
self.names = []
self.choices = {}
for i, (name, (func, args, kwargs)) in enumerate(sorted_items(params)):
self.names.append(name)
_coconut_case_match_to_0 = func, args
_coconut_case_match_check_0 = False
_coconut_match_set_name_choices = _coconut_sentinel
if (_coconut.isinstance(
_coconut_case_match_to_0, _coconut.abc.Sequence)) and (
_coconut.len(_coconut_case_match_to_0) == 2) and (
_coconut_case_match_to_0[0]
== "choice") and (_coconut.isinstance(
_coconut_case_match_to_0[1],
_coconut.abc.Sequence)) and (_coconut.len(
_coconut_case_match_to_0[1]) == 1):
_coconut_match_set_name_choices = _coconut_case_match_to_0[1][
0]
_coconut_case_match_check_0 = True
if _coconut_case_match_check_0:
if _coconut_match_set_name_choices is not _coconut_sentinel:
choices = _coconut_case_match_to_0[1][0]
if _coconut_case_match_check_0:
self.lb.append(0)
self.ub.append(len(choices) - 1)
self.int_var.append(i)
self.choices[name] = choices
if not _coconut_case_match_check_0:
_coconut_match_set_name_start = _coconut_sentinel
_coconut_match_set_name_stop = _coconut_sentinel
_coconut_match_set_name_step = _coconut_sentinel
if (_coconut.isinstance(
_coconut_case_match_to_0, _coconut.abc.Sequence)) and (
_coconut.len(_coconut_case_match_to_0) == 2) and (
_coconut_case_match_to_0[0]
== "randrange") and (_coconut.isinstance(
_coconut_case_match_to_0[1],
_coconut.abc.Sequence)) and (_coconut.len(
_coconut_case_match_to_0[1]) == 3):
_coconut_match_set_name_start = _coconut_case_match_to_0[
1][0]
_coconut_match_set_name_stop = _coconut_case_match_to_0[1][
1]
_coconut_match_set_name_step = _coconut_case_match_to_0[1][
2]
_coconut_case_match_check_0 = True
if _coconut_case_match_check_0:
if _coconut_match_set_name_start is not _coconut_sentinel:
start = _coconut_case_match_to_0[1][0]
if _coconut_match_set_name_stop is not _coconut_sentinel:
stop = _coconut_case_match_to_0[1][1]
if _coconut_match_set_name_step is not _coconut_sentinel:
step = _coconut_case_match_to_0[1][2]
if _coconut_case_match_check_0:
rng = range(start, stop, step)
self.lb.append(0)
self.ub.append(len(rng) - 1)
self.int_var.append(i)
self.choices[name] = rng
if not _coconut_case_match_check_0:
_coconut_match_set_name_start = _coconut_sentinel
_coconut_match_set_name_stop = _coconut_sentinel
if (_coconut.isinstance(
_coconut_case_match_to_0, _coconut.abc.Sequence)) and (
_coconut.len(_coconut_case_match_to_0) == 2) and (
_coconut_case_match_to_0[0]
== "uniform") and (_coconut.isinstance(
_coconut_case_match_to_0[1],
_coconut.abc.Sequence)) and (_coconut.len(
_coconut_case_match_to_0[1]) == 2):
_coconut_match_set_name_start = _coconut_case_match_to_0[
1][0]
_coconut_match_set_name_stop = _coconut_case_match_to_0[1][
1]
_coconut_case_match_check_0 = True
if _coconut_case_match_check_0:
if _coconut_match_set_name_start is not _coconut_sentinel:
start = _coconut_case_match_to_0[1][0]
if _coconut_match_set_name_stop is not _coconut_sentinel:
stop = _coconut_case_match_to_0[1][1]
if _coconut_case_match_check_0:
self.lb.append(start)
self.ub.append(stop)
self.cont_var.append(i)
if not _coconut_case_match_check_0:
raise TypeError(
"insufficiently specified parameter {_coconut_format_0}".
format(_coconut_format_0=(name)))
self.lb = (np.array)(self.lb)
self.ub = (np.array)(self.ub)
self.int_var = (np.array)(self.int_var)
self.cont_var = (np.array)(self.cont_var)