testutil

Testing Infrastructure

This package provides testing utilities for ModFetch unit and integration tests.

Overview

The testutil package provides:

• Mock HTTP servers with canned responses
• In-memory SQLite database for testing
• Test fixture loading helpers
• Mock HTTP round trippers
• Temporary file/directory helpers

Usage

Creating a Test Database

func TestSomething(t *testing.T) {
    db := testutil.TestDB(t)
    // Use db for testing
    // Automatically cleaned up when test completes
}

Mock HTTP Server

func TestHTTPFetcher(t *testing.T) {
    server := testutil.NewMockHTTPServer()
    defer server.Close()

    // Add canned responses
    server.AddJSONResponse("/api/models/123", 200, `{"id": 123, "name": "test"}`)

    // Use server.URL in your tests
    client := &http.Client{}
    resp, err := client.Get(server.URL + "/api/models/123")
    // ... assertions
}

Loading Fixtures

func TestWithFixture(t *testing.T) {
    json := testutil.LoadFixture(t, "huggingface_model.json")
    // Use fixture data in tests
}

Mock Round Tripper

func TestHTTPClient(t *testing.T) {
    mock := testutil.NewMockRoundTripper()
    mock.AddStringResponse("https://api.example.com/model", 200, `{"result": "ok"}`)

    client := &http.Client{Transport: mock}
    // Use client

    // Verify requests
    mock.AssertRequestMade(t, "https://api.example.com/model")
}

Test Fixtures

Fixtures are stored in testdata/fixtures/ and include:

• huggingface_model.json - Sample HuggingFace API response
• civitai_model.json - Sample CivitAI API response

Add new fixtures as needed for additional test cases.

Running Tests

Most tests run against temporary files, in-memory databases, local HTTP servers, or fixtures. Tests that need gated provider access read tokens from the environment and skip when those tokens are not set.

# Run all tests
go test ./...

# Run tests for specific package
go test ./internal/metadata/...

# Run with verbose output
go test -v ./internal/state/...

# Run specific test
go test -run TestDB_UpsertMetadata ./internal/state/...

Best Practices

1. Always use testutil.TestDB() for database tests - it creates an in-memory database that's automatically cleaned up
2. Use fixtures for API responses - keeps tests maintainable and realistic
3. Mock HTTP transports - don't make real network calls in tests
4. Use t.Helper() in utility functions - improves error reporting
5. Clean up resources - use t.Cleanup() for deferred cleanup

Coverage

Current test coverage includes:

Metadata Package

• ✅ HuggingFace fetcher with mocked API responses
• ✅ CivitAI fetcher with mocked API responses
• ✅ Model type inference from filenames and tags
• ✅ Quantization extraction from filenames
• ✅ URL pattern matching
• ✅ Error handling for API failures

State Package

• ✅ Metadata CRUD operations (create, read, update, delete)
• ✅ Filtering by source, type, rating, tags, favorites
• ✅ Full-text search across metadata fields
• ✅ Usage tracking (times_used, last_used)
• ✅ Tag JSON serialization/deserialization
• ✅ Timestamp handling

Current Coverage

Coverage now spans batch parsing, downloader behavior, resolver behavior, configuration validation, metrics, archive extraction, catalog import/export, scanner behavior, and TUI state transitions. Keep new helpers small and prefer real temporary filesystem, SQLite, archive, and HTTP execution where practical.