Skip to content

README.md

Latest commit

 

History

History
192 lines (134 loc) · 7.24 KB

README.md

File metadata and controls

192 lines (134 loc) · 7.24 KB

Integration Tests

This directory contains integration tests for the Express MongoDB application. Unlike unit tests, these tests make actual HTTP requests to the Express app using supertest and connect to a real MongoDB instance to verify end-to-end functionality.

Overview

The integration tests are organized into three main categories:

  1. Movie CRUD Integration Tests (movie.integration.test.ts)

    • Tests basic CRUD API endpoints using supertest
    • Makes actual HTTP requests to the Express app
    • Tests pagination, filtering, and sorting via API
    • Automatically skipped if MONGODB_URI is not set

  2. MongoDB Search Integration Tests (mongodbSearch.integration.test.ts)

    • Tests the MongoDB Atlas Search API endpoint using supertest
    • Makes actual HTTP requests to test the /api/movies/search endpoint
    • Tests search by plot, directors, cast, pagination, and search operators
    • Automatically skipped if ENABLE_SEARCH_TESTS is not set

  3. Advanced Endpoints Integration Tests (advancedEndpoints.integration.test.ts)

    • Tests advanced API endpoints using supertest
    • Makes actual HTTP requests to test:
      • Aggregation endpoints: Movies with comments, statistics by year, directors with most movies
      • Atlas Search endpoint: Compound queries, phrase matching, fuzzy matching
      • Vector Search endpoint: Semantic similarity search using embeddings
    • Aggregation tests automatically run if MONGODB_URI is set
    • Atlas Search tests require ENABLE_SEARCH_TESTS=true
    • Vector Search tests require VOYAGE_API_KEY to be set

Note: Tests use describeIntegration, describeSearch, and describeVectorSearch wrappers (from setup.ts and test files) that automatically skip entire test suites when the required environment variables are not set.

Testing Approach

These integration tests use supertest to make actual HTTP requests to the Express application, testing the complete request/response cycle including:

  • Routing
  • Request parsing
  • Controller logic
  • Database operations
  • Response formatting
  • Error handling

This approach ensures that the API endpoints work correctly from the client's perspective.

Requirements

Basic Integration Tests (CRUD and Aggregations)

  • MONGODB_URI environment variable must be set
  • MongoDB instance must be accessible (can be local MongoDB or Atlas)

Atlas Search Tests

  • MongoDB instance with Search enabled (local MongoDB or Atlas)
  • MONGODB_URI environment variable
  • ENABLE_SEARCH_TESTS=true environment variable to enable tests

Vector Search Tests

  • MONGODB_URI environment variable must be set
  • VOYAGE_API_KEY environment variable must be set with a valid Voyage AI API key
  • MongoDB instance must have the embedded_movies collection with vector embeddings
  • Vector search index must be configured on the embedded_movies collection

Running the Tests

Run All Integration Tests

npm run test:integration

Run Specific Test File

# Run only CRUD tests
npx jest --config jest.integration.config.json tests/integration/movie.integration.test.ts

# Run only Search tests (requires Search-enabled MongoDB)
npx jest --config jest.integration.config.json tests/integration/mongodbSearch.integration.test.ts

# Run only Advanced Endpoints tests (aggregations, search, vector search)
npx jest --config jest.integration.config.json tests/integration/advancedEndpoints.integration.test.ts

Set Environment Variables

Using .env file

Create a .env file in the server/js-express directory:

MONGODB_URI=mongodb://localhost:27017/sample_mflix
# or for Atlas
MONGODB_URI=mongodb+srv://username:password@cluster.mongodb.net/sample_mflix?retryWrites=true&w=majority

# Optional: Enable Search tests (requires Search-enabled MongoDB)
ENABLE_SEARCH_TESTS=true

# Optional: Enable Vector Search tests (requires Voyage AI API key)
VOYAGE_API_KEY=your-voyage-ai-api-key

Test Structure

Setup and Teardown

All integration tests follow an idempotent pattern to ensure they can be run multiple times without side effects:

  • Global Setup (beforeAll in setup.ts): Connects to MongoDB once before all tests
  • Global Teardown (afterAll in setup.ts): Closes MongoDB connection after all tests
  • Pre-Test Cleanup (beforeAll in test files): Removes any orphaned test data from previous runs
  • Post-Test Cleanup (afterEach in test files): Removes test data created during each test

This ensures:

  • Tests can be run multiple times without conflicts
  • The dataset is left in an untouched state after test execution
  • Tests are isolated from each other
  • Failed test runs don't affect subsequent runs

Troubleshooting

Tests are Skipped

If you see "⚠️ Skipping integration tests - MONGODB_URI not set":

  1. Make sure MONGODB_URI environment variable is set
  2. Verify the connection string is correct
  3. Check that MongoDB is running and accessible

Search Tests are Skipped

If you see "⚠️ Skipping MongoDB Search tests - ENABLE_SEARCH_TESTS not set":

  1. Set ENABLE_SEARCH_TESTS=true environment variable
  2. Verify your MongoDB instance has Search enabled (available in both local MongoDB and Atlas)
  3. For local MongoDB, ensure you're running a version that supports Search

Vector Search Tests are Skipped

If you see "⚠️ Vector Search tests skipped - VOYAGE_API_KEY not set":

  1. Set VOYAGE_API_KEY environment variable with your Voyage AI API key
  2. Verify your MongoDB instance has the embedded_movies collection
  3. Ensure the vector search index is configured on the embedded_movies collection
  4. The embedded_movies collection should have documents with plot_embedding_voyage_3_large field

Connection Errors

If tests fail with connection errors:

  1. Verify MONGODB_URI is correct
  2. Check that MongoDB is running and accessible
  3. For Atlas: Verify your IP address is whitelisted
  4. For Atlas: Verify database user credentials are correct
  5. For local MongoDB: Verify the connection string format is correct

Index Creation Timeout

If search tests fail with "Search index did not become ready within 120 seconds":

  1. Check that your MongoDB instance has Search enabled
  2. Verify the search index is being created (check logs or admin UI)
  3. The index creation time varies by instance type and data size
  4. For Atlas free tier clusters, index creation may take longer
  5. For local MongoDB, ensure Search is properly configured

Test Timeouts

Integration tests have a 2-minute timeout (120 seconds) to accommodate:

  • Network latency
  • Index creation and polling
  • Document indexing delays

If tests timeout, you may need to:

  1. Check your network connection
  2. Use a faster MongoDB instance
  3. Increase the timeout in jest.integration.config.json

Differences from Unit Tests

Aspect Unit Tests Integration Tests
HTTP Requests Mocked with Jest Real HTTP requests via supertest
Database Mocked with Jest Real MongoDB instance
Speed Fast (milliseconds) Slower (seconds)
Dependencies None Requires MongoDB
Isolation Complete Requires cleanup
CI/CD Always run Conditional (requires MONGODB_URI)
Purpose Test business logic Test end-to-end API functionality
⚡