| """Train a small EC-frequency baseline using only the training split.""" |
| import argparse |
| from collections import Counter |
| import json |
| from pathlib import Path |
|
|
| import pandas as pd |
|
|
|
|
| def train(path, output): |
| root = Path(__file__).parent |
| task = json.loads((root / 'task.json').read_text()) |
| valid = set(json.loads((root / 'ec_vocabulary.json').read_text())) |
| frame = pd.read_parquet(path) |
| unique = frame.drop_duplicates(task['input_columns'] + ['ec_number']) |
| if not set(unique.ec_number) <= valid: |
| raise ValueError('Training data contains ECs outside the pinned vocabulary.') |
| counts = Counter(unique.ec_number) |
| ranked = sorted(counts, key=lambda ec: (-counts[ec], ec)) |
| if not ranked: |
| raise ValueError('Training data is empty.') |
| limit = 1 |
| result = {'prediction': ranked[:limit], 'training_pairs': len(unique), |
| 'baseline': 'EC frequency; ignores the input'} |
| target = Path(output) |
| target.parent.mkdir(parents=True, exist_ok=True) |
| target.write_text(json.dumps(result, indent=2) + '\n') |
| print(f'Trained on {len(unique)} input–EC pairs; saved {target}') |
|
|
|
|
| if __name__ == '__main__': |
| parser = argparse.ArgumentParser(description=__doc__) |
| parser.add_argument('--train', required=True) |
| parser.add_argument('--output', default='weights/baseline.json') |
| args = parser.parse_args() |
| train(args.train, args.output) |
|
|
