def __init__(self, params, data):
BaseValidator.__init__(self, params, data)
self.train_ratio = self.params.get("train_ratio", 0.8)
self.shuffle = self.params.get("shuffle", True)
self.stratify = self.params.get("stratify", False)
self.random_seed = self.params.get("random_seed", 1706)
log.debug("SplitValidator, train_ratio: {0}".format(self.train_ratio))
def __init__(self, params, data):
BaseValidator.__init__(self, params, data)
if self.data.get("validation") is None:
msg = "Missing validation data"
raise WithDatasetValidatorException(msg)
for i in ["X", "y"]:
if self.data["validation"].get(i) is None:
msg = "Missing {0} in validation data".format(i)
raise WithDatasetValidatorException(msg)
def __init__(self, params):
BaseValidator.__init__(self, params)
cv_path = self.params.get("cv_path")
if cv_path is None:
raise AutoMLException(
"You need to specify `cv` as list or iterable")
self.cv = joblib.load(cv_path)
self.cv = list(self.cv)
self._results_path = self.params.get("results_path")
self._X_path = self.params.get("X_path")
self._y_path = self.params.get("y_path")
self._sample_weight_path = self.params.get("sample_weight_path")
if self._X_path is None or self._y_path is None:
raise AutoMLException("No data path set in CustomValidator params")
folds_path = os.path.join(self._results_path, "folds")
if not os.path.exists(folds_path):
os.mkdir(folds_path)
print("Custom validation strategy")
for fold_cnt, (train_index,
validation_index) in enumerate(self.cv):
print(f"Split {fold_cnt}.")
print(f"Train {train_index.shape[0]} samples.")
print(f"Validation {validation_index.shape[0]} samples.")
train_index_file = os.path.join(
self._results_path,
"folds",
f"fold_{fold_cnt}_train_indices.npy",
)
validation_index_file = os.path.join(
self._results_path,
"folds",
f"fold_{fold_cnt}_validation_indices.npy",
)
np.save(train_index_file, train_index)
np.save(validation_index_file, validation_index)
else:
log.debug("Folds split already done, reuse it")
def __init__(self, params):
BaseValidator.__init__(self, params)
self.train_ratio = self.params.get("train_ratio", 0.8)
self.shuffle = self.params.get("shuffle", True)
self.stratify = self.params.get("stratify", False)
self.random_seed = self.params.get("random_seed", 1234)
log.debug("SplitValidator, train_ratio: {0}".format(self.train_ratio))
self._results_path = self.params.get("results_path")
self._X_path = self.params.get("X_path")
self._y_path = self.params.get("y_path")
if self._X_path is None or self._y_path is None:
raise AutoMLException("No data path set in SplitValidator params")
def __init__(self, params):
BaseValidator.__init__(self, params)
self.train_ratio = self.params.get("train_ratio", 0.8)
self.shuffle = self.params.get("shuffle", True)
self.stratify = self.params.get("stratify", False)
self.random_seed = self.params.get("random_seed", 1234)
self.repeats = self.params.get("repeats", 1)
if not self.shuffle and self.repeats > 1:
warnings.warn(
"Disable repeats in validation because shuffle is disabled")
self.repeats = 1
self._results_path = self.params.get("results_path")
self._X_path = self.params.get("X_path")
self._y_path = self.params.get("y_path")
if self._X_path is None or self._y_path is None:
raise AutoMLException("No data path set in SplitValidator params")
def __init__(self, params, data):
BaseValidator.__init__(self, params, data)
self.k_folds = self.params.get("k_folds", 5)
self.shuffle = self.params.get("shuffle", True)
self.stratify = self.params.get("stratify", False)
self.random_seed = self.params.get("random_seed", 1706)
if self.stratify:
self.skf = StratifiedKFold(
n_splits=self.k_folds,
shuffle=self.shuffle,
random_state=self.random_seed,
)
else:
self.skf = KFold(
n_splits=self.k_folds,
shuffle=self.shuffle,
random_state=self.random_seed,
)
def __init__(self, params):
BaseValidator.__init__(self, params)
self.k_folds = self.params.get("k_folds", 5)
self.shuffle = self.params.get("shuffle", True)
self.stratify = self.params.get("stratify", False)
self.random_seed = self.params.get("random_seed", 1906)
if self.stratify:
if self.shuffle:
self.skf = StratifiedKFold(
n_splits=self.k_folds,
shuffle=self.shuffle,
random_state=self.random_seed if self.shuffle else None,
)
else:
self.skf = StratifiedKFold(n_splits=self.k_folds,
shuffle=self.shuffle)
else:
self.skf = KFold(
n_splits=self.k_folds,
shuffle=self.shuffle,
random_state=self.random_seed if self.shuffle else None,
)
self._results_path = self.params.get("results_path")
self._X_path = self.params.get("X_path")
self._y_path = self.params.get("y_path")
if self._X_path is None or self._y_path is None:
raise AutoMLException("No data path set in KFoldValidator params")
folds_path = os.path.join(self._results_path, "folds")
if not os.path.exists(folds_path):
os.mkdir(folds_path)
X = pd.read_parquet(self._X_path)
y = pd.read_parquet(self._y_path)
y = y["target"]
if isinstance(y[0], bytes):
# see https://github.com/scikit-learn/scikit-learn/issues/16980
y = y.astype(str)
for fold_cnt, (train_index,
validation_index) in enumerate(self.skf.split(X,
y)):
train_index_file = os.path.join(
self._results_path, "folds",
f"fold_{fold_cnt}_train_indices.npy")
validation_index_file = os.path.join(
self._results_path,
"folds",
f"fold_{fold_cnt}_validation_indices.npy",
)
np.save(train_index_file, train_index)
np.save(validation_index_file, validation_index)
del X
del y
gc.collect()
else:
log.debug("Folds split already done, reuse it")
def __init__(self, params):
BaseValidator.__init__(self, params)
self.k_folds = self.params.get("k_folds", 5)
self.shuffle = self.params.get("shuffle", True)
self.stratify = self.params.get("stratify", False)
self.random_seed = self.params.get("random_seed", 1906)
self.repeats = self.params.get("repeats", 1)
if not self.shuffle and self.repeats > 1:
warnings.warn(
"Disable repeats in validation because shuffle is disabled")
self.repeats = 1
self.skf = []
for r in range(self.repeats):
random_seed = self.random_seed + r if self.shuffle else None
if self.stratify:
if self.shuffle:
self.skf += [
StratifiedKFold(
n_splits=self.k_folds,
shuffle=self.shuffle,
random_state=random_seed,
)
]
else:
self.skf += [
StratifiedKFold(
n_splits=self.k_folds,
# shuffle=self.shuffle,
random_state=random_seed,
)
]
else:
self.skf += [
KFold(
n_splits=self.k_folds,
shuffle=self.shuffle,
random_state=random_seed,
)
]
self._results_path = self.params.get("results_path")
self._X_path = self.params.get("X_path")
self._y_path = self.params.get("y_path")
self._sample_weight_path = self.params.get("sample_weight_path")
if self._X_path is None or self._y_path is None:
raise AutoMLException("No data path set in KFoldValidator params")
folds_path = os.path.join(self._results_path, "folds")
if not os.path.exists(folds_path):
os.mkdir(folds_path)
X = pd.read_parquet(self._X_path)
y = pd.read_parquet(self._y_path)
y = y["target"]
if isinstance(y[0], bytes):
# see https://github.com/scikit-learn/scikit-learn/issues/16980
y = y.astype(str)
for repeat_cnt, skf in enumerate(self.skf):
for fold_cnt, (train_index,
validation_index) in enumerate(skf.split(X, y)):
repeat_str = f"_repeat_{repeat_cnt}" if len(
self.skf) > 1 else ""
train_index_file = os.path.join(
self._results_path,
"folds",
f"fold_{fold_cnt}{repeat_str}_train_indices.npy",
)
validation_index_file = os.path.join(
self._results_path,
"folds",
f"fold_{fold_cnt}{repeat_str}_validation_indices.npy",
)
np.save(train_index_file, train_index)
np.save(validation_index_file, validation_index)
del X
del y
gc.collect()
else:
log.debug("Folds split already done, reuse it")