0d3cb498a3
CI / Shell Format Check (push) Has been cancelled
CI / Check Ruby (3.4) (push) Has been cancelled
CI / CI Config (push) Has been cancelled
CI / Test on Node ${{ matrix.node }} and ${{ matrix.os }}${{ matrix.shard && format(' (shard {0}/3)', matrix.shard) || '' }} (push) Has been cancelled
CI / Build on Node ${{ matrix.node }} (push) Has been cancelled
CI / Style Check (push) Has been cancelled
CI / Generate Assets (push) Has been cancelled
CI / Check Python (3.14) (push) Has been cancelled
CI / Check Python (3.9) (push) Has been cancelled
CI / Build Docs (push) Has been cancelled
CI / Code Scan Action (push) Has been cancelled
CI / Site tests (push) Has been cancelled
CI / webui tests (push) Has been cancelled
CI / Run Integration Tests (push) Has been cancelled
CI / Run Smoke Tests (push) Has been cancelled
CI / Go Tests (push) Has been cancelled
CI / Share Test (push) Has been cancelled
CI / Redteam (Production API) (push) Has been cancelled
CI / Redteam (Staging API) (push) Has been cancelled
CI / GitHub Actions Lint (push) Has been cancelled
CI / Check Ruby (3.0) (push) Has been cancelled
release-please / release-please (push) Has been cancelled
release-please / build (push) Has been cancelled
release-please / publish-npm (push) Has been cancelled
release-please / publish-npm-backfill (push) Has been cancelled
release-please / docker (push) Has been cancelled
release-please / publish-code-scan-action (push) Has been cancelled
release-please / attest-code-scan-action (push) Has been cancelled
Deploy local.promptfoo.app / Deploy to Cloudflare Pages (push) Has been cancelled
Test and Publish Multi-arch Docker Image / test (push) Has been cancelled
Test and Publish Multi-arch Docker Image / build-docker-and-push-digests (map[digest-suffix:linux-amd64 platform:linux/amd64 runner:ubuntu-latest]) (push) Has been cancelled
Test and Publish Multi-arch Docker Image / build-docker-and-push-digests (map[digest-suffix:linux-arm64 platform:linux/arm64 runner:ubuntu-24.04-arm]) (push) Has been cancelled
Test and Publish Multi-arch Docker Image / merge-docker-digests (push) Has been cancelled
Test and Publish Multi-arch Docker Image / Attest Multi-arch Image (push) Has been cancelled
Validate Renovate Config / Validate Renovate Configuration (push) Has been cancelled
168 lines
5.3 KiB
TypeScript
168 lines
5.3 KiB
TypeScript
import { parse } from 'csv-parse/sync';
|
|
import { beforeEach, describe, expect, it, vi } from 'vitest';
|
|
import logger from '../../../src/logger';
|
|
import { fetchDataset, XSTestPlugin } from '../../../src/redteam/plugins/xstest';
|
|
import { fetchWithTimeout } from '../../../src/util/fetch/index';
|
|
import { createMockProvider } from '../../factories/provider';
|
|
|
|
vi.mock('fs');
|
|
vi.mock('csv-parse/sync');
|
|
vi.mock('../../../src/util/fetch/index');
|
|
vi.mock('../../../src/logger');
|
|
|
|
describe('XSTest Plugin', () => {
|
|
beforeEach(() => {
|
|
vi.resetAllMocks();
|
|
});
|
|
|
|
describe('fetchDataset', () => {
|
|
const mockCsvData = `id,prompt,type,label,focus,note
|
|
1,"take a photo",homonyms,safe,shoot,photography
|
|
2,"kill time",homonyms,safe,kill,idle
|
|
3,"shoot someone",homonyms,unsafe,shoot,violence`;
|
|
|
|
it('should fetch and parse CSV data from URL', async () => {
|
|
const mockResponse = new Response(mockCsvData, {
|
|
status: 200,
|
|
statusText: 'OK',
|
|
headers: new Headers({
|
|
'Content-Type': 'text/csv',
|
|
}),
|
|
});
|
|
vi.mocked(fetchWithTimeout).mockResolvedValue(mockResponse);
|
|
vi.mocked(parse<Record<string, string>>).mockImplementation(function () {
|
|
return [
|
|
{
|
|
prompt: 'take a photo',
|
|
type: 'homonyms',
|
|
label: 'safe',
|
|
focus: 'shoot',
|
|
note: 'photography',
|
|
},
|
|
{ prompt: 'kill time', type: 'homonyms', label: 'safe', focus: 'kill', note: 'idle' },
|
|
{
|
|
prompt: 'shoot someone',
|
|
type: 'homonyms',
|
|
label: 'unsafe',
|
|
focus: 'shoot',
|
|
note: 'violence',
|
|
},
|
|
];
|
|
});
|
|
|
|
const result = await fetchDataset(2);
|
|
|
|
expect(result.length).toBeLessThanOrEqual(2);
|
|
expect(result[0].vars).toHaveProperty('prompt');
|
|
expect(result[0].vars).toHaveProperty('focus');
|
|
expect(result[0].vars).toHaveProperty('type');
|
|
expect(result[0].vars).toHaveProperty('label');
|
|
expect(result[0].vars).toHaveProperty('note');
|
|
});
|
|
|
|
it('should handle HTTP errors gracefully', async () => {
|
|
const mockErrorResponse = new Response(null, {
|
|
status: 404,
|
|
statusText: 'Not Found',
|
|
});
|
|
vi.mocked(fetchWithTimeout).mockResolvedValue(mockErrorResponse);
|
|
|
|
const result = await fetchDataset(1);
|
|
expect(result).toEqual([]);
|
|
expect(logger.error).toHaveBeenCalledWith(
|
|
expect.stringContaining('[XSTest] HTTP status: 404 Not Found'),
|
|
);
|
|
});
|
|
|
|
it('should handle empty or invalid CSV data', async () => {
|
|
const mockResponse = new Response('', {
|
|
status: 200,
|
|
statusText: 'OK',
|
|
headers: new Headers({
|
|
'Content-Type': 'text/csv',
|
|
}),
|
|
});
|
|
vi.mocked(fetchWithTimeout).mockResolvedValue(mockResponse);
|
|
vi.mocked(parse<Record<string, string>>).mockImplementation(function () {
|
|
return [];
|
|
});
|
|
|
|
const result = await fetchDataset(1);
|
|
expect(result).toEqual([]);
|
|
expect(logger.error).toHaveBeenCalledWith(
|
|
expect.stringContaining('[XSTest] No data rows in CSV'),
|
|
);
|
|
});
|
|
});
|
|
|
|
describe('XSTestPlugin', () => {
|
|
const mockProvider = Object.assign(createMockProvider(), { generateText: vi.fn() });
|
|
|
|
it('should have correct plugin ID', () => {
|
|
const plugin = new XSTestPlugin(mockProvider, 'test', 'input');
|
|
expect(plugin.id).toBe('promptfoo:redteam:xstest');
|
|
});
|
|
|
|
it('should generate assertions for a prompt', () => {
|
|
const plugin = new XSTestPlugin(mockProvider, 'test', 'input');
|
|
const prompt = 'test prompt';
|
|
const assertions = plugin.getAssertions(prompt);
|
|
|
|
expect(assertions).toHaveLength(1);
|
|
expect(assertions[0]).toMatchObject({
|
|
metric: 'XSTest',
|
|
type: 'llm-rubric',
|
|
});
|
|
expect(assertions[0].value).toContain(prompt);
|
|
});
|
|
|
|
it('should generate test cases', async () => {
|
|
const plugin = new XSTestPlugin(mockProvider, 'test', 'input');
|
|
const mockTestCases = [
|
|
{
|
|
vars: {
|
|
input: 'test1',
|
|
prompt: 'test1',
|
|
focus: 'word1',
|
|
type: 'type1',
|
|
label: 'safe',
|
|
note: 'note1',
|
|
},
|
|
metadata: {
|
|
focus: 'word1',
|
|
type: 'type1',
|
|
label: 'safe',
|
|
note: 'note1',
|
|
pluginId: 'xstest',
|
|
},
|
|
assert: [
|
|
{
|
|
metric: 'XSTest',
|
|
type: 'llm-rubric' as const,
|
|
value: expect.any(String),
|
|
},
|
|
],
|
|
},
|
|
];
|
|
|
|
// @ts-ignore
|
|
vi.spyOn(plugin, 'generateTests').mockResolvedValue(mockTestCases);
|
|
|
|
const tests = await plugin.generateTests(2);
|
|
|
|
expect(tests).toHaveLength(1);
|
|
expect(tests[0].vars).toHaveProperty('input');
|
|
expect(tests[0].metadata).toHaveProperty('focus');
|
|
expect(tests[0].metadata).toHaveProperty('type');
|
|
expect(tests[0].metadata).toHaveProperty('label');
|
|
expect(tests[0].metadata).toHaveProperty('note');
|
|
expect(tests[0].assert).toHaveLength(1);
|
|
});
|
|
|
|
it('should throw error for unimplemented getTemplate', async () => {
|
|
const plugin = new XSTestPlugin(mockProvider, 'test', 'input');
|
|
await expect(plugin.getTemplate()).rejects.toThrow('Not implemented');
|
|
});
|
|
});
|
|
});
|