extract-medium-bugs.ts

import fs from 'fs';

// Configuration
const INPUT_FILE = 'bugs.json';
const OUTPUT_FILE = 'medium-bugs.json';
const DIFFICULTY_LEVEL = 'medium';

interface Bug {
    repo: string;
    LLMReviewPassed: boolean;
    difficultyLevel?: string;
    [key: string]: any;
}

// Get limit from command line argument, default to all bugs if not specified
const limit = process.argv[2] ? parseInt(process.argv[2], 10) : Infinity;
if (isNaN(limit) || limit <= 0) {
    console.error('Error: Limit must be a positive number');
    process.exit(1);
}

// Read and parse the bugs file
const bugs = JSON.parse(fs.readFileSync(INPUT_FILE, 'utf-8')) as Bug[];

// Filter for bugs that have "medium" difficulty
const mediumBugs = bugs.filter(bug => 
    bug.LLMReviewPassed === true && 
    bug.difficultyLevel === DIFFICULTY_LEVEL
);

console.log(`Found ${mediumBugs.length} medium bugs out of ${bugs.length} total bugs`);

// Apply limit if needed
const limitedBugs = mediumBugs.slice(0, limit);
console.log(`Using ${limitedBugs.length} medium bugs (limit: ${limit === Infinity ? 'none' : limit})`);

// Group bugs by repository to show distribution
const bugsByRepo = limitedBugs.reduce((acc: { [key: string]: any[] }, bug: any) => {
    if (!acc[bug.repo]) {
        acc[bug.repo] = [];
    }
    acc[bug.repo].push(bug);
    return acc;
}, {});

// Get list of repos with their bug counts
const repoStats = Object.entries(bugsByRepo)
    .map(([repo, bugs]) => ({
        repo,
        count: bugs.length
    }))
    .sort((a, b) => b.count - a.count); // Sort by count descending

console.log('\n=== Repositories with Medium Bugs (in output) ===');
repoStats.forEach(({repo, count}) => {
    console.log(`${repo}: ${count} medium bugs`);
});

// Write medium bugs to output file
fs.writeFileSync(OUTPUT_FILE, JSON.stringify(limitedBugs, null, 2));

console.log(`\n${limitedBugs.length} medium bugs have been saved to ${OUTPUT_FILE}`);