File size: 4,817 Bytes
1aa7fae
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
"""
Data preparation script for the predictive maintenance project.

Responsibilities:
- Load the raw engine dataset from the Hugging Face dataset repo (preferred)
  or from the local data folder as a fallback.
- Clean and preprocess the data (rename columns, handle missing values,
  drop duplicates, basic sanity checks).
- Split the cleaned data into train and test sets.
- Save train and test CSVs locally.
- Upload the resulting train and test CSVs back to the Hugging Face dataset repo.
"""

from __future__ import annotations

from pathlib import Path
from typing import Tuple

import numpy as np
import pandas as pd
from sklearn.model_selection import train_test_split

import config
from hf_data_utils import download_dataset_file, upload_dataset_file


def _load_raw_data_from_hf_or_local() -> pd.DataFrame:
    """
    Try to load the raw dataset from the Hugging Face dataset repo.
    If that fails (e.g., no token or repo yet), fall back to the local CSV.
    """
    # Preferred: load from HF dataset repo if token and repo are configured
    if config.HF_TOKEN and config.HF_DATASET_REPO:
        try:
            remote_path = download_dataset_file(
                filename="data/engine_data.csv",
                repo_id=config.HF_DATASET_REPO,
                token=config.HF_TOKEN,
                local_dir=config.DATA_DIR,
            )
            return pd.read_csv(remote_path)
        except Exception:
            # Fall back to local file
            pass

    # Local fallback
    if not config.RAW_DATA_FILE.exists():
        raise FileNotFoundError(
            f"Raw data file not found at {config.RAW_DATA_FILE}. "
            "Ensure engine_data.csv exists or upload it to the HF dataset repo."
        )

    return pd.read_csv(config.RAW_DATA_FILE)


def _clean_data(df: pd.DataFrame) -> pd.DataFrame:
    """
    Perform basic cleaning and feature engineering.
    """
    # Standardize column names
    df = df.rename(columns=config.RAW_COLUMN_RENAME_MAP)

    # Keep only the expected columns (drop any extras, if present)
    expected_cols = set(config.FEATURE_COLUMNS + [config.TARGET_COLUMN])
    df = df[[col for col in df.columns if col in expected_cols]]

    # Drop duplicate rows
    df = df.drop_duplicates().reset_index(drop=True)

    # Handle missing values: for this numeric dataset, fill with median
    if df.isna().any().any():
        df = df.fillna(df.median(numeric_only=True))

    # Ensure target is integer/binary
    df[config.TARGET_COLUMN] = df[config.TARGET_COLUMN].astype(int)

    return df


def _train_test_split(
    df: pd.DataFrame,
) -> Tuple[pd.DataFrame, pd.DataFrame]:
    """
    Split the cleaned dataframe into train and test sets.
    """
    X = df[config.FEATURE_COLUMNS]
    y = df[config.TARGET_COLUMN]

    X_train, X_test, y_train, y_test = train_test_split(
        X,
        y,
        test_size=config.TEST_SIZE,
        random_state=config.RANDOM_STATE,
        stratify=y,
    )

    train_df = X_train.copy()
    train_df[config.TARGET_COLUMN] = y_train

    test_df = X_test.copy()
    test_df[config.TARGET_COLUMN] = y_test

    return train_df, test_df


def main() -> None:
    """
    Execute the full data preparation pipeline.
    """
    print("Loading raw data...")
    raw_df = _load_raw_data_from_hf_or_local()
    print(f"Raw data shape: {raw_df.shape}")

    print("Cleaning data...")
    clean_df = _clean_data(raw_df)
    print(f"Clean data shape: {clean_df.shape}")

    print("Performing train/test split...")
    train_df, test_df = _train_test_split(clean_df)
    print(f"Train shape: {train_df.shape}, Test shape: {test_df.shape}")

    # Save locally
    config.PROCESSED_DIR.mkdir(parents=True, exist_ok=True)
    train_df.to_csv(config.TRAIN_FILE, index=False)
    test_df.to_csv(config.TEST_FILE, index=False)
    print(f"Saved train to {config.TRAIN_FILE}")
    print(f"Saved test to {config.TEST_FILE}")

    # Upload to HF dataset repo, if configured
    if config.HF_TOKEN and config.HF_DATASET_REPO:
        try:
            print("Uploading train and test splits to Hugging Face dataset repo...")
            upload_dataset_file(
                local_path=config.TRAIN_FILE,
                repo_id=config.HF_DATASET_REPO,
                repo_path="data/train.csv",
                token=config.HF_TOKEN,
            )
            upload_dataset_file(
                local_path=config.TEST_FILE,
                repo_id=config.HF_DATASET_REPO,
                repo_path="data/test.csv",
                token=config.HF_TOKEN,
            )
            print("Upload to Hugging Face completed.")
        except Exception as e:
            print(
                f"Warning: Failed to upload train/test to Hugging Face dataset repo: {e}"
            )


if __name__ == "__main__":
    main()