Decoder2704/python-chatbot-jay

1#!/usr/bin/env python32"""3Build AnswersV2.csv from Answers.csv (Preprocessing.ipynb logic).4 5Adds columns: latest_score, alternate, date — unchanged from the notebook.6"""7 8from __future__ import annotations9 10import argparse11import sys12from datetime import date13from pathlib import Path14 15import pandas as pd16 17_BACKEND_DIR = Path(__file__).resolve().parent.parent18if str(_BACKEND_DIR) not in sys.path:19    sys.path.insert(0, str(_BACKEND_DIR))20 21from app.paths import DATA_DIR, PYBOT_DATA_DIR, ensure_data_and_artifacts_dirs22 23DEFAULT_INPUT = PYBOT_DATA_DIR / "Answers.csv"24DEFAULT_OUTPUT = DATA_DIR / "AnswersV2.csv"25CSV_ENCODING = "latin-1"26 27 28def parse_args() -> argparse.Namespace:29    p = argparse.ArgumentParser(description="Build AnswersV2.csv from Answers.csv.")30    p.add_argument(31        "--input",32        type=Path,33        default=DEFAULT_INPUT,34        help=f"Path to Answers.csv (default: {DEFAULT_INPUT})",35    )36    p.add_argument(37        "--output",38        type=Path,39        default=DEFAULT_OUTPUT,40        help=f"Output path (default: {DEFAULT_OUTPUT})",41    )42    return p.parse_args()43 44 45def build_answers_v2(answers: pd.DataFrame) -> pd.DataFrame:46    """Mirror Preprocessing.ipynb: add latest_score, alternate, date; preserve all else."""47    out = answers.copy()48    out["latest_score"] = out["Score"]49    out["alternate"] = out["Score"]50    out["date"] = date.today()51    return out52 53 54def main() -> int:55    ensure_data_and_artifacts_dirs()56    args = parse_args()57    input_path = args.input.resolve()58    output_path = args.output.resolve()59 60    if not input_path.is_file():61        print(f"ERROR: input file not found: {input_path}", file=sys.stderr)62        return 163 64    try:65        answers = pd.read_csv(input_path, encoding=CSV_ENCODING)66    except UnicodeDecodeError as exc:67        print(68            f"ERROR: failed to decode {input_path} with encoding={CSV_ENCODING!r}: {exc}",69            file=sys.stderr,70        )71        return 172    except OSError as exc:73        print(f"ERROR: cannot read {input_path}: {exc}", file=sys.stderr)74        return 175 76    if "Score" not in answers.columns:77        print("ERROR: Answers.csv must contain a 'Score' column.", file=sys.stderr)78        return 179 80    if answers.empty:81        print("WARNING: Answers.csv is empty; writing empty AnswersV2.csv.", file=sys.stderr)82 83    built = build_answers_v2(answers)84 85    try:86        output_path.parent.mkdir(parents=True, exist_ok=True)87        built.to_csv(output_path, index=False)88    except OSError as exc:89        print(f"ERROR: cannot write {output_path}: {exc}", file=sys.stderr)90        return 191 92    print(f"input:  {input_path}")93    print(f"output: {output_path}")94    print(f"rows:   {len(built):,}")95    return 096 97 98if __name__ == "__main__":99    raise SystemExit(main())100