-
Notifications
You must be signed in to change notification settings - Fork 0
/
helpers.py
100 lines (87 loc) · 3.29 KB
/
helpers.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
"""Some helper functions for project 1."""
import csv
import numpy as np
import os
def load_csv_data(data_path, sub_sample=False, selected_cols=None):
"""
This function loads the data and returns the respectinve numpy arrays.
Remember to put the 3 files in the same folder and to not change the names of the files.
Args:
data_path (str): datafolder path
sub_sample (bool, optional): If True the data will be subsempled. Default to False.
Returns:
x_train (np.array): training data
x_test (np.array): test data
y_train (np.array): labels for training data in format (-1,1)
train_ids (np.array): ids of training data
test_ids (np.array): ids of test data
"""
with open(os.path.join(data_path, "x_train.csv"), "r") as f:
header = f.readline().strip().split(",")
y_train = np.genfromtxt(
os.path.join(data_path, "y_train.csv"),
delimiter=",",
skip_header=1,
dtype=int,
usecols=1,
)
x_train = np.genfromtxt(
os.path.join(data_path, "x_train.csv"), delimiter=",", skip_header=1
)
x_test = np.genfromtxt(
os.path.join(data_path, "x_test.csv"), delimiter=",", skip_header=1
)
col_names_train = np.genfromtxt(
data_path + "/x_train.csv", delimiter=",", max_rows=1, dtype=str
).tolist()
col_names_test = np.genfromtxt(
data_path + "/x_test.csv", delimiter=",", max_rows=1, dtype=str
).tolist()
train_ids = x_train[:, 0].astype(dtype=int)
test_ids = x_test[:, 0].astype(dtype=int)
x_train = x_train[:, 1:]
x_test = x_test[:, 1:]
final_columns = []
# Select only the specified columns
if selected_cols is not None:
selected_indices = []
for col_name in selected_cols:
if col_name in header:
selected_indices.append(header.index(col_name) - 1)
final_columns.append(col_name)
x_train = x_train[:, selected_indices]
x_test = x_test[:, selected_indices]
# sub-sample
if sub_sample:
y_train = y_train[::50]
x_train = x_train[::50]
train_ids = train_ids[::50]
return (
x_train,
x_test,
y_train,
train_ids,
test_ids,
col_names_train,
col_names_test,
final_columns,
)
def create_csv_submission(ids, y_pred, name):
"""
This function creates a csv file named 'name' in the format required for a submission in Kaggle or AIcrowd.
The file will contain two columns the first with 'ids' and the second with 'y_pred'.
y_pred must be a list or np.array of 1 and -1 otherwise the function will raise a ValueError.
Args:
ids (list,np.array): indices
y_pred (list,np.array): predictions on data correspondent to indices
name (str): name of the file to be created
"""
# Check that y_pred only contains -1 and 1
if not all(i in [-1, 1] for i in y_pred):
raise ValueError("y_pred can only contain values -1, 1")
with open(name, "w", newline="") as csvfile:
fieldnames = ["Id", "Prediction"]
writer = csv.DictWriter(csvfile, delimiter=",", fieldnames=fieldnames)
writer.writeheader()
for r1, r2 in zip(ids, y_pred):
writer.writerow({"Id": int(r1), "Prediction": int(r2)})