yujunjun/autogen
1
0
Fork 0
mirror of https://github.com/microsoft/autogen.git synced 2025-07-23 08:52:56 +00:00
Code Issues Packages Projects Releases Wiki Activity

Sklearn api x (#405)

Browse Source 
* changed signature of automl.predict and automl.predict_proba to X

* XGBoostEstimator

* changed signature of Prophet predict to X

* changed signature of ARIMA predict to X

* changed signature of TS_SKLearn_Regressor predict to X
This commit is contained in:
MichaelMarien 2022-01-16 23:37:56 +01:00 committed by GitHub
parent a6d70efff7
commit 1c911da9f8
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 52 additions and 56 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
flaml/automl.py
View File

@ -714,13 +714,11 @@ class AutoML(BaseEstimator):
"""Time taken to find best model in seconds.""" """Time taken to find best model in seconds."""
return self.__dict__.get("_time_taken_best_iter") return self.__dict__.get("_time_taken_best_iter")
def predict( def predict(self, X: Union[np.array, pd.DataFrame, List[str], List[List[str]]]):
self, X_test: Union[np.array, pd.DataFrame, List[str], List[List[str]]]
):
"""Predict label from features. """Predict label from features.
Args: Args:
X_test: A numpy array of featurized instances, shape n * m, X: A numpy array of featurized instances, shape n * m,
or for 'ts_forecast' task: or for 'ts_forecast' task:
a pandas dataframe with the first column containing a pandas dataframe with the first column containing
timestamp values (datetime type) or an integer n for timestamp values (datetime type) or an integer n for
@ -748,8 +746,8 @@ class AutoML(BaseEstimator):
"No estimator is trained. Please run fit with enough budget." "No estimator is trained. Please run fit with enough budget."
) )
return None return None
X_test = self._preprocess(X_test) X = self._preprocess(X)
y_pred = estimator.predict(X_test) y_pred = estimator.predict(X)
if ( if (
isinstance(y_pred, np.ndarray) isinstance(y_pred, np.ndarray)
and y_pred.ndim > 1 and y_pred.ndim > 1
@ -763,12 +761,12 @@ class AutoML(BaseEstimator):
else: else:
return y_pred return y_pred
def predict_proba(self, X_test): def predict_proba(self, X):
"""Predict the probability of each class from features, only works for """Predict the probability of each class from features, only works for
classification problems. classification problems.
Args: Args:
X_test: A numpy array of featurized instances, shape n * m. X: A numpy array of featurized instances, shape n * m.
Returns: Returns:
A numpy array of shape n * c. c is the # classes. Each element at A numpy array of shape n * c. c is the # classes. Each element at
@ -780,8 +778,8 @@ class AutoML(BaseEstimator):
"No estimator is trained. Please run fit with enough budget." "No estimator is trained. Please run fit with enough budget."
) )
return None return None
X_test = self._preprocess(X_test) X = self._preprocess(X)
proba = self._trained_estimator.predict_proba(X_test) proba = self._trained_estimator.predict_proba(X)
return proba return proba
def _preprocess(self, X): def _preprocess(self, X):

90
flaml/model.py
View File

@ -197,32 +197,32 @@ class BaseEstimator:
train_time = self._fit(X_train, y_train, **kwargs) train_time = self._fit(X_train, y_train, **kwargs)
return train_time return train_time
def predict(self, X_test): def predict(self, X):
"""Predict label from features. """Predict label from features.
Args: Args:
X_test: A numpy array or a dataframe of featurized instances, shape n*m. X: A numpy array or a dataframe of featurized instances, shape n*m.
Returns: Returns:
A numpy array of shape n*1. A numpy array of shape n*1.
Each element is the label for a instance. Each element is the label for a instance.
""" """
if self._model is not None: if self._model is not None:
X_test = self._preprocess(X_test) X = self._preprocess(X)
return self._model.predict(X_test) return self._model.predict(X)
else: else:
logger.warning( logger.warning(
"Estimator is not fit yet. Please run fit() before predict()." "Estimator is not fit yet. Please run fit() before predict()."
) )
return np.ones(X_test.shape[0]) return np.ones(X.shape[0])
def predict_proba(self, X_test): def predict_proba(self, X):
"""Predict the probability of each class from features. """Predict the probability of each class from features.
Only works for classification problems Only works for classification problems
Args: Args:
X_test: A numpy array of featurized instances, shape n*m. X: A numpy array of featurized instances, shape n*m.
Returns: Returns:
A numpy array of shape n*c. c is the # classes. A numpy array of shape n*c. c is the # classes.
@ -231,8 +231,8 @@ class BaseEstimator:
""" """
assert self._task in CLASSIFICATION, "predict_proba() only for classification." assert self._task in CLASSIFICATION, "predict_proba() only for classification."
X_test = self._preprocess(X_test) X = self._preprocess(X)
return self._model.predict_proba(X_test) return self._model.predict_proba(X)
def cleanup(self): def cleanup(self):
del self._model del self._model
@ -708,18 +708,18 @@ class TransformersEstimator(BaseEstimator):
) )
return test_dataset, training_args return test_dataset, training_args
def predict_proba(self, X_test): def predict_proba(self, X):
assert ( assert (
self._task in CLASSIFICATION self._task in CLASSIFICATION
), "predict_proba() only for classification tasks." ), "predict_proba() only for classification tasks."
test_dataset, _ = self._init_model_for_predict(X_test) test_dataset, _ = self._init_model_for_predict(X)
predictions = self._trainer.predict(test_dataset) predictions = self._trainer.predict(test_dataset)
self._trainer = None self._trainer = None
return predictions.predictions return predictions.predictions
def predict(self, X_test): def predict(self, X):
test_dataset, training_args = self._init_model_for_predict(X_test) test_dataset, training_args = self._init_model_for_predict(X)
if self._task not in NLG_TASKS: if self._task not in NLG_TASKS:
predictions = self._trainer.predict(test_dataset) predictions = self._trainer.predict(test_dataset)
else: else:
@ -1108,12 +1108,12 @@ class XGBoostEstimator(SKLearnEstimator):
train_time = time.time() - start_time train_time = time.time() - start_time
return train_time return train_time
def predict(self, X_test): def predict(self, X):
import xgboost as xgb import xgboost as xgb
if not issparse(X_test): if not issparse(X):
X_test = self._preprocess(X_test) X = self._preprocess(X)
dtest = xgb.DMatrix(X_test) dtest = xgb.DMatrix(X)
return super().predict(dtest) return super().predict(dtest)
@classmethod @classmethod
@ -1598,22 +1598,22 @@ class Prophet(SKLearnEstimator):
self._model = model self._model = model
return train_time return train_time
def predict(self, X_test): def predict(self, X):
if isinstance(X_test, int): if isinstance(X, int):
raise ValueError( raise ValueError(
"predict() with steps is only supported for arima/sarimax." "predict() with steps is only supported for arima/sarimax."
" For Prophet, pass a dataframe with the first column containing" " For Prophet, pass a dataframe with the first column containing"
" the timestamp values." " the timestamp values."
) )
if self._model is not None: if self._model is not None:
X_test = self._preprocess(X_test) X = self._preprocess(X)
forecast = self._model.predict(X_test) forecast = self._model.predict(X)
return forecast["yhat"] return forecast["yhat"]
else: else:
logger.warning( logger.warning(
"Estimator is not fit yet. Please run fit() before predict()." "Estimator is not fit yet. Please run fit() before predict()."
) )
return np.ones(X_test.shape[0]) return np.ones(X.shape[0])
class ARIMA(Prophet): class ARIMA(Prophet):
@ -1678,30 +1678,30 @@ class ARIMA(Prophet):
self._model = model self._model = model
return train_time return train_time
def predict(self, X_test): def predict(self, X):
if self._model is not None: if self._model is not None:
if isinstance(X_test, int): if isinstance(X, int):
forecast = self._model.forecast(steps=X_test) forecast = self._model.forecast(steps=X)
elif isinstance(X_test, DataFrame): elif isinstance(X, DataFrame):
start = X_test[TS_TIMESTAMP_COL].iloc[0] start = X[TS_TIMESTAMP_COL].iloc[0]
end = X_test[TS_TIMESTAMP_COL].iloc[-1] end = X[TS_TIMESTAMP_COL].iloc[-1]
if len(X_test.columns) > 1: if len(X.columns) > 1:
X_test = self._preprocess(X_test.drop(columns=TS_TIMESTAMP_COL)) X = self._preprocess(X.drop(columns=TS_TIMESTAMP_COL))
regressors = list(X_test) regressors = list(X)
print(start, end, X_test.shape) print(start, end, X.shape)
forecast = self._model.predict( forecast = self._model.predict(
start=start, end=end, exog=X_test[regressors] start=start, end=end, exog=X[regressors]
) )
else: else:
forecast = self._model.predict(start=start, end=end) forecast = self._model.predict(start=start, end=end)
else: else:
raise ValueError( raise ValueError(
"X_test needs to be either a pandas Dataframe with dates as the first column" "X needs to be either a pandas Dataframe with dates as the first column"
" or an int number of periods for predict()." " or an int number of periods for predict()."
) )
return forecast return forecast
else: else:
return np.ones(X_test if isinstance(X_test, int) else X_test.shape[0]) return np.ones(X if isinstance(X, int) else X.shape[0])
class SARIMAX(ARIMA): class SARIMAX(ARIMA):
@ -1873,42 +1873,40 @@ class TS_SKLearn_Regressor(SKLearnEstimator):
train_time = time.time() - current_time train_time = time.time() - current_time
return train_time return train_time
def predict(self, X_test): def predict(self, X):
if self._model is not None: if self._model is not None:
X_test = self.transform_X(X_test) X = self.transform_X(X)
X_test = self._preprocess(X_test) X = self._preprocess(X)
if isinstance(self._model, list): if isinstance(self._model, list):
assert len(self._model) == len( assert len(self._model) == len(
X_test X
), "Model is optimized for horizon, length of X_test must be equal to `period`." ), "Model is optimized for horizon, length of X must be equal to `period`."
preds = [] preds = []
for i in range(1, len(self._model) + 1): for i in range(1, len(self._model) + 1):
( (
X_pred, X_pred,
_, _,
) = self.hcrystaball_model._transform_data_to_tsmodel_input_format( ) = self.hcrystaball_model._transform_data_to_tsmodel_input_format(
X_test.iloc[:i, :] X.iloc[:i, :]
) )
preds.append(self._model[i - 1].predict(X_pred)[-1]) preds.append(self._model[i - 1].predict(X_pred)[-1])
forecast = DataFrame( forecast = DataFrame(
data=np.asarray(preds).reshape(-1, 1), data=np.asarray(preds).reshape(-1, 1),
columns=[self.hcrystaball_model.name], columns=[self.hcrystaball_model.name],
index=X_test.index, index=X.index,
) )
else: else:
( (
X_pred, X_pred,
_, _,
) = self.hcrystaball_model._transform_data_to_tsmodel_input_format( ) = self.hcrystaball_model._transform_data_to_tsmodel_input_format(X)
X_test
)
forecast = self._model.predict(X_pred) forecast = self._model.predict(X_pred)
return forecast return forecast
else: else:
logger.warning( logger.warning(
"Estimator is not fit yet. Please run fit() before predict()." "Estimator is not fit yet. Please run fit() before predict()."
) )
return np.ones(X_test.shape[0]) return np.ones(X.shape[0])
class LGBM_TS_Regressor(TS_SKLearn_Regressor): class LGBM_TS_Regressor(TS_SKLearn_Regressor):