junshan-kit 2.2.9__py2.py3-none-any.whl → 2.3.0__py2.py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
junshan_kit/DataSets.py
CHANGED
@@ -12,7 +12,7 @@ import junshan_kit.kit
|
|
12
12
|
from sklearn.preprocessing import StandardScaler
|
13
13
|
|
14
14
|
#----------------------------------------------------------
|
15
|
-
def
|
15
|
+
def _download_data(data_name):
|
16
16
|
from junshan_kit.kit import JianguoyunDownloaderFirefox, JianguoyunDownloaderChrome
|
17
17
|
|
18
18
|
# User selects download method
|
@@ -36,12 +36,12 @@ def download_data(data_name):
|
|
36
36
|
else:
|
37
37
|
print("❌ Invalid choice. Please enter 1 or 2.\n")
|
38
38
|
|
39
|
-
def
|
39
|
+
def _run(csv_path, data_name, drop_cols, label_col, label_map, print_info):
|
40
40
|
if not os.path.exists(csv_path):
|
41
41
|
print('\n' + '*'*60)
|
42
42
|
print(f"Please download the data.")
|
43
43
|
print(csv_path)
|
44
|
-
|
44
|
+
_download_data(data_name)
|
45
45
|
junshan_kit.kit.unzip_file(f'./exp_data/{data_name}/{data_name}.zip', f'./exp_data/{data_name}')
|
46
46
|
|
47
47
|
cleaner = junshan_kit.DataProcessor.CSV_TO_Pandas()
|
@@ -59,10 +59,10 @@ def credit_card_fraud_detection(data_name = "Credit Card Fraud Detection", print
|
|
59
59
|
|
60
60
|
csv_path = f'./exp_data/{data_name}/creditcard.csv'
|
61
61
|
drop_cols = []
|
62
|
-
label_col = '
|
62
|
+
label_col = 'Class'
|
63
63
|
label_map = {0: -1, 1: 1}
|
64
64
|
|
65
|
-
df =
|
65
|
+
df = _run(csv_path, data_name, drop_cols, label_col, label_map, print_info)
|
66
66
|
|
67
67
|
return df
|
68
68
|
|
@@ -70,10 +70,10 @@ def credit_card_fraud_detection(data_name = "Credit Card Fraud Detection", print
|
|
70
70
|
def diabetes_health_indicators_dataset(data_name = "Diabetes Health Indicators Dataset", print_info = False):
|
71
71
|
csv_path = f'./exp_data/{data_name}/diabetes_dataset.csv'
|
72
72
|
drop_cols = []
|
73
|
-
label_col = '
|
73
|
+
label_col = 'diagnosed_diabetes'
|
74
74
|
label_map = {0: -1, 1: 1}
|
75
75
|
|
76
|
-
df =
|
76
|
+
df = _run(csv_path, data_name, drop_cols, label_col, label_map, print_info)
|
77
77
|
|
78
78
|
return df
|
79
79
|
|
@@ -87,7 +87,7 @@ def electric_vehicle_population_data(data_name = "Electric Vehicle Population Da
|
|
87
87
|
'Plug-in Hybrid Electric Vehicle (PHEV)': -1
|
88
88
|
}
|
89
89
|
|
90
|
-
df =
|
90
|
+
df = _run(csv_path, data_name, drop_cols, label_col, label_map, print_info)
|
91
91
|
|
92
92
|
return df
|
93
93
|
|
@@ -97,7 +97,7 @@ def global_house_purchase_dataset(data_name = "Global House Purchase Dataset", p
|
|
97
97
|
label_col = 'Electric Vehicle Type'
|
98
98
|
label_map = {0: -1, 1: 1}
|
99
99
|
|
100
|
-
df =
|
100
|
+
df = _run(csv_path, data_name, drop_cols, label_col, label_map, print_info)
|
101
101
|
|
102
102
|
return df
|
103
103
|
|
@@ -108,7 +108,7 @@ def health_lifestyle_dataset(data_name = "Health_lifestyle_dataset", print_info
|
|
108
108
|
label_col = 'decision'
|
109
109
|
label_map = {0: -1, 1: 1}
|
110
110
|
|
111
|
-
df =
|
111
|
+
df = _run(csv_path, data_name, drop_cols, label_col, label_map, print_info)
|
112
112
|
|
113
113
|
return df
|
114
114
|
|
@@ -0,0 +1,7 @@
|
|
1
|
+
junshan_kit/DataProcessor.py,sha256=eryVmS5BFZj8wjDN2QWVHqkbFgFuWU0HXV9s6TGf9QM,4442
|
2
|
+
junshan_kit/DataSets.py,sha256=rf5AVlA9DxP7wBpXjSO1_xznCMuxEoK50TqExafwHhc,3972
|
3
|
+
junshan_kit/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
4
|
+
junshan_kit/kit.py,sha256=tB1TpW9hW1EweK1RQwHOdUo7uG1QU4vSeyR0fdaSydo,9569
|
5
|
+
junshan_kit-2.3.0.dist-info/METADATA,sha256=9NlU4YOD0zx5F5tLIMREKKxn-LwYD8-7IFVtp7DvMNM,329
|
6
|
+
junshan_kit-2.3.0.dist-info/WHEEL,sha256=tkmg4JIqwd9H8mL30xA7crRmoStyCtGp0VWshokd1Jc,105
|
7
|
+
junshan_kit-2.3.0.dist-info/RECORD,,
|
@@ -1,7 +0,0 @@
|
|
1
|
-
junshan_kit/DataProcessor.py,sha256=eryVmS5BFZj8wjDN2QWVHqkbFgFuWU0HXV9s6TGf9QM,4442
|
2
|
-
junshan_kit/DataSets.py,sha256=BNlXUbsautITelxp35uHSYSq2bTbIbqLyQnmqapaeDc,3963
|
3
|
-
junshan_kit/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
4
|
-
junshan_kit/kit.py,sha256=tB1TpW9hW1EweK1RQwHOdUo7uG1QU4vSeyR0fdaSydo,9569
|
5
|
-
junshan_kit-2.2.9.dist-info/METADATA,sha256=oIytayt1Z9OpFOJl8_7cYMl3li3KsphQfZDRqsqat8g,329
|
6
|
-
junshan_kit-2.2.9.dist-info/WHEEL,sha256=tkmg4JIqwd9H8mL30xA7crRmoStyCtGp0VWshokd1Jc,105
|
7
|
-
junshan_kit-2.2.9.dist-info/RECORD,,
|
File without changes
|