File size: 4,021 Bytes
f1d873d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
# Imports

import pickle
import os
from sklearn.metrics import classification_report, ConfusionMatrixDisplay
from sklearn.ensemble import RandomForestClassifier
from sklearn.impute import SimpleImputer, KNNImputer
from sklearn.pipeline import Pipeline
from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import StandardScaler, RobustScaler, OneHotEncoder
from sklearn.model_selection import train_test_split
import pandas as pd
from ydata_profiling import ProfileReport
from sklearn import datasets
from subprocess import call

# PATHS
DIRPATH = os.path.dirname(os.path.realpath(__file__))
ml_fp = os.path.join(DIRPATH, "assets", "ml", "ml_components.pkl")
req_fp = os.path.join(DIRPATH, "assets", "ml", "requirements.txt")
eda_report_fp = os.path.join(DIRPATH, "assets", "ml", "eda-report.html")

# import some data to play with
iris = datasets.load_iris(return_X_y=False, as_frame=True)

df = iris['frame']
target_col = 'target'
# pandas profiling
profile = ProfileReport(df, title="Dataset", html={
                        'style': {'full_width': True}})
profile.to_file(eda_report_fp)

# Dataset Splitting
# Please specify
to_ignore_cols = [
    "ID",  # ID
    "Id", "id",
    target_col
]


num_cols = list(set(df.select_dtypes('number')) - set(to_ignore_cols))
cat_cols = list(set(df.select_dtypes(exclude='number')) - set(to_ignore_cols))
print(f"\n[Info] The '{len(num_cols)}' numeric columns are : {num_cols}\nThe '{len(cat_cols)}' categorical columns are : {cat_cols}")

X, y = df.iloc[:, :-1], df.iloc[:, -1].values


X_train, X_eval, y_train, y_eval = train_test_split(
    X, y, test_size=0.2, random_state=0, stratify=y)

print(
    f"\n[Info] Dataset splitted : (X_train , y_train) = {(X_train.shape , y_train.shape)}, (X_eval y_eval) = {(X_eval.shape , y_eval.shape)}. \n")

y_train

# Modeling

# Imputers
num_imputer = SimpleImputer(strategy="mean").set_output(transform="pandas")
cat_imputer = SimpleImputer(
    strategy="most_frequent").set_output(transform="pandas")

# Scaler & Encoder
if len(cat_cols) > 0:
    df_imputed_stacked_cat = cat_imputer.fit_transform(
        df
        .append(df)
        .append(df)
        [cat_cols])
    cat_ = OneHotEncoder(sparse=False, drop="first").fit(
        df_imputed_stacked_cat).categories_
else:
    cat_ = 'auto'

encoder = OneHotEncoder(categories=cat_, sparse=False, drop="first")
scaler = StandardScaler().set_output(transform="pandas")


# feature pipelines
num_pipe = Pipeline(steps=[("num_imputer", num_imputer), ("scaler", scaler)])
cat_pipe = Pipeline(steps=[("cat_imputer", cat_imputer), ("encoder", encoder)])

# end2end features preprocessor

transformers = []

transformers.append(("numerical", num_pipe, num_cols)) if len(
    num_cols) > 0 else None
transformers.append(("categorical", cat_pipe, cat_cols,)) if len(
    cat_cols) > 0 else None
#  ("date", date_pipe, date_cols,),

preprocessor = ColumnTransformer(
    transformers=transformers).set_output(transform="pandas")

print(
    f"\n[Info] Features Transformer : {transformers}. \n")


# end2end pipeline
end2end_pipeline = Pipeline([
    ('preprocessor', preprocessor),
    ('model', RandomForestClassifier(random_state=10))
]).set_output(transform="pandas")

# Training
print(
    f"\n[Info] Training.\n[Info] X_train : columns( {X_train.columns.tolist()}), shape: {X_train.shape} .\n")

end2end_pipeline.fit(X_train, y_train)

# Evaluation
print(
    f"\n[Info] Evaluation.\n")
y_eval_pred = end2end_pipeline.predict(X_eval)

print(classification_report(y_eval, y_eval_pred,
      target_names=iris['target_names']))

# ConfusionMatrixDisplay.from_predictions(
#     y_eval, y_eval_pred, display_labels=iris['target_names'])

# Exportation
print(
    f"\n[Info] Exportation.\n")
to_export = {
    "labels": iris['target_names'],
    "pipeline": end2end_pipeline,
}


# save components to file
with open(ml_fp, 'wb') as file:
    pickle.dump(to_export, file)

# Requirements
# ! pip freeze > requirements.txt
call(f"pip freeze > {req_fp}", shell=True)