File size: 5,920 Bytes
4937cba | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 | """Training/inference preprocessing pipeline utilities."""
from __future__ import annotations
from pathlib import Path
from typing import Any
import joblib
import numpy as np
import pandas as pd
from imblearn.over_sampling import SMOTE
from sklearn.compose import ColumnTransformer
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.utils.class_weight import compute_class_weight
TARGET_COLUMN = "Class"
SCALE_COLUMNS = ["Time", "Amount"]
def split_data(
df: pd.DataFrame,
*,
target_column: str = TARGET_COLUMN,
test_size: float = 0.2,
random_state: int = 42,
) -> tuple[pd.DataFrame, pd.DataFrame, pd.Series, pd.Series]:
"""Split dataframe into train/test with class stratification."""
if target_column not in df.columns:
raise ValueError(f"Missing target column: {target_column}")
if not 0 < test_size < 1:
raise ValueError("test_size must be between 0 and 1")
X = df.drop(columns=[target_column])
y = df[target_column]
return train_test_split(
X,
y,
test_size=test_size,
random_state=random_state,
stratify=y,
)
def scale_features(
df: pd.DataFrame,
*,
columns: list[str] | None = None,
scaler: StandardScaler | None = None,
) -> tuple[pd.DataFrame, StandardScaler]:
"""Scale selected columns and return transformed dataframe and scaler."""
scale_columns = columns or SCALE_COLUMNS
missing = [column for column in scale_columns if column not in df.columns]
if missing:
raise ValueError(f"Columns not found for scaling: {missing}")
local_scaler = scaler or StandardScaler()
result = df.copy()
result[scale_columns] = local_scaler.fit_transform(df[scale_columns])
return result, local_scaler
def build_preprocessor(
feature_columns: list[str],
*,
scale_columns: list[str] | None = None,
) -> ColumnTransformer:
"""Build a column transformer for consistent training/inference transforms."""
chosen_scale_columns = scale_columns or SCALE_COLUMNS
missing = [column for column in chosen_scale_columns if column not in feature_columns]
if missing:
raise ValueError(f"Scale columns missing from features: {missing}")
preprocessor = ColumnTransformer(
transformers=[("scale", StandardScaler(), chosen_scale_columns)],
remainder="passthrough",
verbose_feature_names_out=False,
)
preprocessor.set_output(transform="pandas")
return preprocessor
def transform_features(
preprocessor: ColumnTransformer,
X: pd.DataFrame,
) -> pd.DataFrame:
"""Transform feature dataframe using a fitted preprocessor."""
transformed = preprocessor.transform(X)
if not isinstance(transformed, pd.DataFrame):
transformed = pd.DataFrame(transformed, columns=preprocessor.get_feature_names_out())
return transformed
def handle_imbalance(
X_train: pd.DataFrame,
y_train: pd.Series,
*,
method: str = "class_weight",
random_state: int = 42,
sampling_strategy: float = 0.5,
) -> tuple[pd.DataFrame, pd.Series, dict[str, Any]]:
"""Handle class imbalance using strategy selected by method."""
selected = method.lower()
if selected not in {"none", "class_weight", "smote"}:
raise ValueError("method must be one of: none, class_weight, smote")
if selected == "none":
return X_train, y_train, {"method": "none", "class_weight": None}
if selected == "class_weight":
classes = np.array(sorted(y_train.unique().tolist()))
weights = compute_class_weight(class_weight="balanced", classes=classes, y=y_train)
class_weight = {int(label): float(weight) for label, weight in zip(classes, weights)}
return X_train, y_train, {"method": "class_weight", "class_weight": class_weight}
smote = SMOTE(random_state=random_state, sampling_strategy=sampling_strategy)
X_resampled, y_resampled = smote.fit_resample(X_train, y_train)
X_balanced = pd.DataFrame(X_resampled, columns=X_train.columns)
y_balanced = pd.Series(y_resampled, name=y_train.name)
return X_balanced, y_balanced, {"method": "smote", "class_weight": None}
def save_preprocessor(preprocessor: ColumnTransformer, output_path: str | Path) -> Path:
"""Persist fitted preprocessor to disk."""
path = Path(output_path)
path.parent.mkdir(parents=True, exist_ok=True)
joblib.dump(preprocessor, path)
return path
def load_preprocessor(preprocessor_path: str | Path) -> ColumnTransformer:
"""Load persisted preprocessor from disk."""
return joblib.load(Path(preprocessor_path))
def preprocess_for_training(
df: pd.DataFrame,
*,
target_column: str = TARGET_COLUMN,
test_size: float = 0.2,
random_state: int = 42,
imbalance_method: str = "class_weight",
preprocessor_path: str | Path = "models/preprocessor.pkl",
) -> dict[str, Any]:
"""Run train/test split, fit/transform preprocessor, and handle imbalance."""
X_train_raw, X_test_raw, y_train, y_test = split_data(
df,
target_column=target_column,
test_size=test_size,
random_state=random_state,
)
preprocessor = build_preprocessor(feature_columns=X_train_raw.columns.tolist())
preprocessor.fit(X_train_raw)
X_train = transform_features(preprocessor, X_train_raw)
X_test = transform_features(preprocessor, X_test_raw)
X_train_final, y_train_final, imbalance_metadata = handle_imbalance(
X_train,
y_train,
method=imbalance_method,
random_state=random_state,
)
save_preprocessor(preprocessor, preprocessor_path)
return {
"X_train": X_train_final,
"X_test": X_test,
"y_train": y_train_final,
"y_test": y_test,
"preprocessor": preprocessor,
"imbalance_metadata": imbalance_metadata,
}
|