|
1
|
|
|
''' |
|
2
|
|
|
Utilities and auxiliary functions. |
|
3
|
|
|
|
|
4
|
|
|
:author: Andreas Kanz |
|
5
|
|
|
|
|
6
|
|
|
''' |
|
7
|
|
|
|
|
8
|
|
|
# Imports |
|
9
|
|
|
import numpy as np |
|
10
|
|
|
import pandas as pd |
|
11
|
|
|
|
|
12
|
|
|
|
|
13
|
|
|
def _corr_selector(corr, split=None, threshold=0): |
|
14
|
|
|
''' |
|
15
|
|
|
Select correlations based on the provided parameters. |
|
16
|
|
|
|
|
17
|
|
|
Parameters |
|
18
|
|
|
---------- |
|
19
|
|
|
corr: List or matrix of correlations. |
|
20
|
|
|
|
|
21
|
|
|
split: {None, 'pos', 'neg', 'high', 'low'}, default None |
|
22
|
|
|
Type of split to be performed. |
|
23
|
|
|
|
|
24
|
|
|
threshold: float, default 0 |
|
25
|
|
|
Value between 0 <= threshold <= 1 |
|
26
|
|
|
|
|
27
|
|
|
Returns: |
|
28
|
|
|
------- |
|
29
|
|
|
corr: List or matrix of (filtered) correlations. |
|
30
|
|
|
''' |
|
31
|
|
|
|
|
32
|
|
|
if split == 'pos': |
|
33
|
|
|
corr = corr.where((corr >= threshold) & (corr > 0)) |
|
34
|
|
|
print('Displaying positive correlations. Use "threshold" to further limit the results.') |
|
35
|
|
|
elif split == 'neg': |
|
36
|
|
|
corr = corr.where((corr <= threshold) & (corr < 0)) |
|
37
|
|
|
print('Displaying negative correlations. Use "threshold" to further limit the results.') |
|
38
|
|
|
elif split == 'high': |
|
39
|
|
|
corr = corr.where(np.abs(corr) >= threshold) |
|
40
|
|
|
print(f'Displaying absolute correlations above the threshold ({threshold}).') |
|
41
|
|
|
elif split == 'low': |
|
42
|
|
|
corr = corr.where(np.abs(corr) <= threshold) |
|
43
|
|
|
print(f'Displaying absolute correlations below the threshold ({threshold}).') |
|
44
|
|
|
else: |
|
45
|
|
|
corr = corr |
|
46
|
|
|
|
|
47
|
|
|
return corr |
|
48
|
|
|
|
|
49
|
|
|
|
|
50
|
|
|
def _drop_duplicates(data): |
|
51
|
|
|
''' |
|
52
|
|
|
Provides information and drops duplicate rows. |
|
53
|
|
|
|
|
54
|
|
|
Parameters |
|
55
|
|
|
---------- |
|
56
|
|
|
data: 2D dataset that can be coerced into Pandas DataFrame. |
|
57
|
|
|
|
|
58
|
|
|
Returns |
|
59
|
|
|
------- |
|
60
|
|
|
data: Deduplicated Pandas DataFrame |
|
61
|
|
|
rows_dropped: Index Object of rows dropped. |
|
62
|
|
|
''' |
|
63
|
|
|
|
|
64
|
|
|
data = pd.DataFrame(data).copy() |
|
65
|
|
|
rows_dropped = data[data.duplicated()].index |
|
66
|
|
|
data = data.drop_duplicates() |
|
67
|
|
|
|
|
68
|
|
|
return data, rows_dropped |
|
69
|
|
|
|
|
70
|
|
|
|
|
71
|
|
|
def _memory_usage(data): |
|
72
|
|
|
''' |
|
73
|
|
|
Gives the total memory usage in kilobytes. |
|
74
|
|
|
|
|
75
|
|
|
Parameters |
|
76
|
|
|
---------- |
|
77
|
|
|
data: 2D dataset that can be coerced into Pandas DataFrame. |
|
78
|
|
|
|
|
79
|
|
|
Returns |
|
80
|
|
|
------- |
|
81
|
|
|
memory_usage: float |
|
82
|
|
|
''' |
|
83
|
|
|
|
|
84
|
|
|
data = pd.DataFrame(data).copy() |
|
85
|
|
|
memory_usage = round(data.memory_usage(index=True, deep=True).sum()/1024, 2) |
|
86
|
|
|
|
|
87
|
|
|
return memory_usage |
|
88
|
|
|
|
|
89
|
|
|
|
|
90
|
|
|
def _missing_vals(data): |
|
91
|
|
|
''' |
|
92
|
|
|
Gives metrics of missing values in the dataset. |
|
93
|
|
|
|
|
94
|
|
|
Parameters |
|
95
|
|
|
---------- |
|
96
|
|
|
data: 2D dataset that can be coerced into Pandas DataFrame. |
|
97
|
|
|
|
|
98
|
|
|
Returns |
|
99
|
|
|
------- |
|
100
|
|
|
mv_total: float, number of missing values in the entire dataset |
|
101
|
|
|
mv_rows: float, number of missing values in each row |
|
102
|
|
|
mv_cols: float, number of missing values in each column |
|
103
|
|
|
mv_rows_ratio: float, ratio of missing values for each row |
|
104
|
|
|
mv_cols_ratio: float, ratio of missing values for each column |
|
105
|
|
|
''' |
|
106
|
|
|
|
|
107
|
|
|
data = pd.DataFrame(data).copy() |
|
108
|
|
|
mv_rows = data.isna().sum(axis=1) |
|
109
|
|
|
mv_cols = data.isna().sum(axis=0) |
|
110
|
|
|
mv_total = data.isna().sum().sum() |
|
111
|
|
|
mv_rows_ratio = mv_rows/data.shape[1] |
|
112
|
|
|
mv_cols_ratio = mv_cols/data.shape[0] |
|
113
|
|
|
|
|
114
|
|
|
return {'mv_total': mv_total, |
|
115
|
|
|
'mv_rows': mv_rows, |
|
116
|
|
|
'mv_cols': mv_cols, |
|
117
|
|
|
'mv_rows_ratio': mv_rows_ratio, |
|
118
|
|
|
'mv_cols_ratio': mv_cols_ratio} |
|
119
|
|
|
|
|
120
|
|
|
|
|
121
|
|
|
def _validate_input_range(value, desc, lower, upper): |
|
122
|
|
|
if value < lower or value > upper: |
|
123
|
|
|
raise ValueError( |
|
124
|
|
|
f'Input value for {desc} is {value} but should be in the range {lower} <= {desc} <= {upper}.') |
|
125
|
|
|
|
|
126
|
|
|
|
|
127
|
|
|
def _validate_input_int(value, desc): |
|
128
|
|
|
if type(value) != int: |
|
129
|
|
|
raise TypeError(f'Input value for {desc} is {type(value)} but should be an integer.') |
|
130
|
|
|
|
|
131
|
|
|
|
|
132
|
|
|
def _validate_input_bool(value, desc): |
|
133
|
|
|
if not(isinstance(value, bool)): |
|
134
|
|
|
raise ValueError(f'Input value for {desc} is {value} but should be boolean.') |
|
135
|
|
|
|
akanz1 /
klib
