-
Notifications
You must be signed in to change notification settings - Fork 2
/
transform.py
119 lines (99 loc) · 3.87 KB
/
transform.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
import argparse
import os
import pandas as pd
from azureml.core import Dataset, Datastore, Workspace
from azureml.core.run import Run, _OfflineRun
from loguru import logger
from sklearn.model_selection import train_test_split
def str_to_bool(value):
"""
*Converts a string to a boolean
*Args:
*value (Str): The value to convert to a boolean
*Returns:
*vallue: True or False depending on value of input
"""
if value.lower() in ("yes", "true", "t", "y", "1"):
return True
elif value.lower() in ("no", "false", "f", "n", "0"):
return False
else:
raise argparse.ArgumentTypeError("Boolean value expected.")
parser = argparse.ArgumentParser(
description="getting inputs from the pipeline setup")
parser.add_argument("--input_path", type=str, default=os.path.join(
'data', 'regression_kaggle_retail_data_analytics', 'raw'))
parser.add_argument("--output_path", type=str, default=os.path.join(
'data', 'regression_kaggle_retail_data_analytics', 'processed'))
parser.add_argument("--register_dataset", type=str_to_bool, default=False)
parser.add_argument("--dataset_name", type=str,
default='retail_training_tabular_dataset')
args, unknown = parser.parse_known_args()
os.makedirs(args.output_path, exist_ok=True)
input_path = args.input_path
output_path = args.output_path
logger.debug(f'{input_path = }')
logger.debug(f'{output_path = }')
logger.debug(f'{args.register_dataset = }')
logger.debug(f'{args.dataset_name = }')
run = Run.get_context()
if type(run) == _OfflineRun:
data_folder_path = input_path
ws = Workspace.from_config()
else:
data_folder_path = os.path.join(
input_path, 'data', 'retail', 'train', 'raw')
ws = run.experiment.workspace
logger.debug(f'{data_folder_path = }')
'''
Source: https://www.kaggle.com/manjeetsingh/retaildataset
'''
features_df = pd.read_csv(
os.path.join(data_folder_path, 'Features data set.csv'))
sales_df = pd.read_csv(
os.path.join(data_folder_path, 'sales data-set.csv'))
stores_df = pd.read_csv(
os.path.join(data_folder_path, 'stores data-set.csv'))
# Merge data
merged_df = pd.merge(
left=sales_df,
right=features_df,
how='left',
on=['Store', 'Date', 'IsHoliday'])
merged_df = pd.merge(
left=merged_df,
right=stores_df,
how='left',
on=['Store'])
logger.debug(f'{merged_df.shape = }') # (421570, 16)
# Convert booleans into integers
merged_df['IsHoliday'] = merged_df['IsHoliday'].astype(int)
print(merged_df.head())
print(merged_df['Weekly_Sales'].describe())
# Split data into train and test
train_df, test_df = train_test_split(merged_df, test_size=0.2, random_state=0)
logger.debug(f'{train_df.shape = }') # (337256, 16)
logger.debug(f'{test_df.shape = }') # (84314, 16)
# Sample data
sample_100_df = test_df.sample(n=100, random_state=0)
logger.debug(f'{sample_100_df.shape = }') # (100, 16)
# Save train and test into csv files
train_df_path = os.path.join(output_path, 'train_data.csv')
test_df_path = os.path.join(output_path, 'test_data.csv')
sample_100_df_path = os.path.join(output_path, 'test_data_sample_100.csv')
train_df.to_csv(train_df_path, index=False)
test_df.to_csv(test_df_path, index=False)
sample_100_df.to_csv(sample_100_df_path, index=False)
logger.debug(f'train_df saved to {train_df_path}')
logger.debug(f'test_df saved to {test_df_path}')
logger.debug(f'sample_100_df saved to {sample_100_df_path}')
if args.register_dataset:
workspaceblobstore = Datastore.get(ws, 'workspaceblobstore')
logger.debug(f'Connected to datastore {workspaceblobstore.name}')
training_tabular_dataset = Dataset.Tabular.register_pandas_dataframe(
dataframe=train_df,
target=workspaceblobstore,
name=args.dataset_name,
show_progress=True)
logger.debug("Registered version {0} of dataset {1}".format(
training_tabular_dataset.version, training_tabular_dataset.name))