create_metadata.py

#!/usr/bin/env python3
"""
Script to create metadata.json files for all exercism exercises
by counting TEST_CASE and REQUIRE statements in test files.
"""

import os
import re
import json

def count_test_cases_and_assertions(file_path):
    """Count TEST_CASE and REQUIRE statements in a test file."""
    try:
        with open(file_path, 'r', encoding='utf-8') as f:
            content = f.read()

        # Count TEST_CASE occurrences
        test_cases = len(re.findall(r'TEST_CASE\s*\(', content))

        # Count REQUIRE and REQUIRE_THROWS assertions
        require_matches = re.findall(r'REQUIRE(?:_THROWS(?:_AS)?)?(?:_FALSE)?\s*\(', content)
        assertions = len(require_matches)

        return test_cases, assertions
    except Exception as e:
        print(f"Error processing {file_path}: {e}")
        return 0, 0

def create_metadata_for_all_exercises():
    """Process all exercises and create metadata.json files."""
    practice_dir = "/Users/marcvanduyn/Projects/Microsoft/llm-cpp-eval/benchmark/practice"

    # Get all test files
    test_files = []
    for root, dirs, files in os.walk(practice_dir):
        for file in files:
            if file.endswith('_test.cpp'):
                test_files.append(os.path.join(root, file))

    print(f"Found {len(test_files)} test files")

    processed = 0
    for test_file in test_files:
        # Extract exercise name from path
        exercise_dir = os.path.dirname(test_file)
        exercise_name = os.path.basename(exercise_dir)

        # Check if metadata.json already exists
        metadata_path = os.path.join(exercise_dir, "metadata.json")
        if os.path.exists(metadata_path):
            print(f"Skipping {exercise_name} - metadata.json already exists")
            continue

        # Count test cases and assertions
        test_cases, assertions = count_test_cases_and_assertions(test_file)

        # Create metadata.json
        metadata = {
            "number_of_test_cases": test_cases,
            "number_of_assertions": assertions
        }

        with open(metadata_path, 'w') as f:
            json.dump(metadata, f, indent=2)

        print(f"Created metadata for {exercise_name}: {test_cases} test cases, {assertions} assertions")
        processed += 1

    print(f"Processed {processed} exercises")

if __name__ == "__main__":
    create_metadata_for_all_exercises()