def transform(self, X, y=None):
X = pd.DataFrame(X.toarray())
assert isinstance(X, pd.DataFrame)
X = df_optimized(X)
if self.verbose:
print(X.head())
return X
def transform(self, df, y=None):
df = df_optimized(X, self.verbose)
return df
def transform(self, X, y=None):
X_dataframe = pd.DataFrame(X.toarray())
pdb.set_trace()
X_optimized = df_optimized(X_dataframe)
pdb.set_trace()
return X_optimized
mlflow=True, # set to True to log params to mlflow
experiment_name="[Fed-up!]-Phi-TaxiFare",
pipeline_memory=None,
model_upload=False, # for automatic upload to gcp
distance_type="manhattan",
feateng=["distance_to_center", "direction", "distance", "time_features", "geohash"],
)
####################
# Get and clean data
####################
if __name__ == "__main__":
print("############ Loading Data ############")
df = get_data(**params)
df = clean_df(df)
df = df_optimized(df)
y_train = df["fare_amount"]
X_train = df.drop("fare_amount", axis=1)
del df
print("shape: {}".format(X_train.shape))
print("size: {} Mb".format(X_train.memory_usage().sum() / 1e6))
####################
# single model kaggle transmission
####################
if params["is_4_kaggle"] == True:
print("Auto-Kaggle-submit is challenge is active")
t = Trainer(X=X_train, y=y_train, **params)
del X_train, y_train
t.train()
def size_optimize(self, df):
return df_optimized(df)
