mirror of
https://github.com/multipleof4/lynchmark.git
synced 2026-01-14 00:27:55 +00:00
84 lines
3.1 KiB
JavaScript
84 lines
3.1 KiB
JavaScript
const fs = require('fs').promises;
|
|
const path = require('path');
|
|
const axios = require('axios');
|
|
const { exec } = require('child_process');
|
|
const util = require('util');
|
|
const { performance } = require('perf_hooks');
|
|
|
|
const execPromise = util.promisify(exec);
|
|
const README_PATH = path.join(__dirname, '..', 'README');
|
|
const TESTS_DIR = path.join(__dirname, '..', 'tests');
|
|
const TEMP_FILE = path.join(__dirname, 'temp_test.mjs');
|
|
|
|
const getLlmCode = async (prompt, model) => {
|
|
try {
|
|
const res = await axios.post(
|
|
'https://openrouter.ai/api/v1/chat/completions',
|
|
{ model, messages: [{ role: 'user', content: prompt }] },
|
|
{ headers: { Authorization: `Bearer ${process.env.OPENROUTER_KEY}` } }
|
|
);
|
|
const content = res.data.choices[0].message.content;
|
|
return content.match(/```(?:javascript|js)?\n([\s\S]+?)\n```/)?.[1].trim() ?? content.trim();
|
|
} catch (error) {
|
|
console.error(`API Error for ${model}: ${error.message}`);
|
|
return null;
|
|
}
|
|
};
|
|
|
|
const runTest = async (code) => {
|
|
const start = performance.now();
|
|
let passed = false;
|
|
try {
|
|
await fs.writeFile(TEMP_FILE, code);
|
|
await execPromise(`node ${TEMP_FILE}`);
|
|
passed = true;
|
|
} catch (error) { // eslint-disable-line no-empty
|
|
} finally {
|
|
await fs.unlink(TEMP_FILE).catch(() => {});
|
|
}
|
|
return { passed, duration: (performance.now() - start) / 1000 };
|
|
};
|
|
|
|
const main = async () => {
|
|
const readme = await fs.readFile(README_PATH, 'utf-8');
|
|
const models = readme.match(/<!-- MODELS_START -->\n([\s\S]+?)\n<!-- MODELS_END -->/)[1].trim().split('\n');
|
|
const percentage = parseInt(readme.match(/RUN_PERCENTAGE:\s*(\d+)/)?.[1] ?? '100', 10);
|
|
const allTestDirs = (await fs.readdir(TESTS_DIR, { withFileTypes: true }))
|
|
.filter(d => d.isDirectory()).map(d => d.name).sort();
|
|
const testsToRun = allTestDirs.slice(0, Math.ceil(allTestDirs.length * (percentage / 100)));
|
|
|
|
const results = [];
|
|
for (const model of models) {
|
|
results.push(`**${model}**`);
|
|
for (const dir of allTestDirs) {
|
|
if (!testsToRun.includes(dir)) {
|
|
results.push(`- ${dir}: ⚪ Not Run`);
|
|
continue;
|
|
}
|
|
const { prompt, harness } = require(path.join(TESTS_DIR, dir, 'test.js'));
|
|
console.log(`Running ${dir} for ${model}...`);
|
|
const llmCode = await getLlmCode(prompt, model);
|
|
if (llmCode) {
|
|
const outDir = path.join(TESTS_DIR, dir, 'outputs');
|
|
await fs.mkdir(outDir, { recursive: true });
|
|
const fname = `${model.replace(/[\/:]/g, '_')}_${new Date().toISOString().replace(/:/g, '-')}.js`;
|
|
await fs.writeFile(path.join(outDir, fname), llmCode);
|
|
const { passed, duration } = await runTest(`${llmCode}\n${harness}`);
|
|
results.push(`- ${dir}: ${passed ? '✅ Pass' : '❌ Fail'} (${duration.toFixed(3)}s)`);
|
|
} else {
|
|
results.push(`- ${dir}: ❌ API Error`);
|
|
}
|
|
}
|
|
results.push('');
|
|
}
|
|
|
|
const newReadme = readme.replace(
|
|
/<!-- RESULTS_START -->[\s\S]*<!-- RESULTS_END -->/,
|
|
`<!-- RESULTS_START -->\n${results.join('\n').trim()}\n<!-- RESULTS_END -->`
|
|
);
|
|
await fs.writeFile(README_PATH, newReadme);
|
|
console.log('Benchmark complete. README updated.');
|
|
};
|
|
|
|
main().catch(console.error);
|