Spaces:
Sleeping
Sleeping
| """Export Gen_AI Dataset.xlsx into train.csv and test.csv at repo root.""" | |
| from pathlib import Path | |
| import csv | |
| import openpyxl | |
| ROOT = Path(__file__).resolve().parent.parent | |
| SRC = ROOT / "Gen_AI Dataset.xlsx" | |
| def export_sheet(ws, out_path: Path) -> int: | |
| rows = list(ws.iter_rows(values_only=True)) | |
| with out_path.open("w", newline="", encoding="utf-8") as f: | |
| w = csv.writer(f) | |
| for row in rows: | |
| w.writerow(["" if c is None else c for c in row]) | |
| return len(rows) - 1 | |
| def main() -> None: | |
| wb = openpyxl.load_workbook(SRC, data_only=True) | |
| train_n = export_sheet(wb["Train-Set"], ROOT / "train.csv") | |
| test_n = export_sheet(wb["Test-Set"], ROOT / "test.csv") | |
| print(f"train.csv: {train_n} labeled rows") | |
| print(f"test.csv: {test_n} queries") | |
| if __name__ == "__main__": | |
| main() | |