Skip to content

FBCM Testing Strategy

Current State

Modules With Tests (8/18 source modules)

Module Test File Coverage Level
metadata.py (functions) test_base.py High — playoff conversion, week parsing
config.py test_config.py High — all from_cli_and_config paths
browser_retry.py test_browser_retry.py High — retry logic, error recovery
file_namer.py test_file_namer.py High — NFL/NCAA/Series naming
constants.py test_constants.py Basic validation
utils.py test_utils.py generate_episode_metadata_xml
logging_config.py test_logging_config.py setup_logging
parsers/* test_parsers/* High — all 5 parsers with HTML fixtures

Modules Without Tests

Module LOC Risk Testability
models.py 393 HIGH Excellent — pure dataclasses, no I/O
draft_buzz.py 409 HIGH Mixed — static methods testable, browser logic needs mocks
docx/word_gen.py 920 HIGH Mixed — helpers testable, DOCX generation needs fixtures
file_operations.py 234 HIGH Good — filesystem ops mockable with tmp_path
nfl.py 499 HIGH Good — injectable deps, mockable API client
metadata.py (MetaDataCreator) 137 MEDIUM Good — filesystem + string logic
mcmillen.py 30 MEDIUM Good — mock requests
downloaders/base_downloader.py 100 MEDIUM Low — yt-dlp wrapper
fbcm.py (CLI) 626 LOW Can use Click's CliRunner

Testing Strategy

Framework & Tools

  • pytest (already in use) — test runner and fixtures
  • pytest-cov — coverage reporting (added as dev dependency)
  • unittest.mock — mocking external dependencies (Playwright, requests, ffmpeg, yt-dlp)
  • tmp_path — pytest built-in fixture for filesystem tests
  • HTML fixtures — already established pattern in tests/test_fbcm/test_parsers/conftest.py

Testing Approach by Module Category

1. Pure Logic (Unit Tests — No Mocking Required)

These modules have no external dependencies and can be tested directly:

  • models.py — from_dict(), to_dict(), _convert_value(), position-aware type resolution
  • docx/word_gen.py helpers — blend_colors(), darken_color(), hex_to_rgb(), skill_bar(), get_primary_position()
  • metadata.py MetaDataCreator._create_title_string()

2. I/O-Dependent (Mock External Dependencies)

  • file_operations.py — Use tmp_path for filesystem, mock mutagen.mp4.MP4 and ffmpeg
  • nfl.py — Mock GriddyNFL client, use injectable collaborators
  • draft_buzz.py — Mock Playwright browser/page objects, test parsing orchestration with HTML fixtures
  • mcmillen.py — Mock requests.get, test HTML parsing with fixture strings

3. Integration-Heavy (Defer or Mark as @pytest.mark.integration)

  • fbcm.py CLI commands — Click's CliRunner for smoke tests
  • downloaders/base_downloader.py — yt-dlp wrapper, low value to unit test
  • docx/word_gen.py DOCX generation — complex document rendering, consider golden-file approach later

Fixture Strategy

  • HTML fixtures: Continue the established pattern in conftest.py for parser and scraper tests
  • Factory fixtures: Use @pytest.fixture factories (like existing make_soup) for creating test data
  • Model fixtures: Create reusable prospect data fixtures for word_gen.py and draft_buzz.py tests
  • tmp_path: Use pytest's built-in tmp_path for all filesystem operations

Coverage Target

Given the I/O-heavy nature of the codebase:

  • Target: 70% line coverage as initial goal
  • Pure logic modules: Target 90%+ (models, helpers, metadata functions)
  • I/O modules: Target 60%+ (mock boundaries, test business logic)
  • CLI/integration: Defer to future tickets, mark with @pytest.mark.integration

Snapshot/Golden-File Testing

Not recommended at this stage. The parsers already use HTML fixture strings effectively. Golden-file testing could be considered later for DOCX output validation if visual regression becomes a concern.

Browser-Dependent Tests

  • Mock Playwright objects (Browser, Page, Locator) for unit tests
  • Static methods in PageFetcher (_make_absolute_url, _get_image_type, _should_skip_image) are testable without mocks
  • Full browser integration tests should be marked @pytest.mark.integration and @pytest.mark.slow

Pytest Configuration

The following additions were made to pyproject.toml:

  • pytest-cov added to dev dependencies
  • Coverage configuration updated with --cov flags in addopts
  • Coverage report configured with fail_under threshold

Priority Order for Test Implementation

  1. models.py — Used everywhere, pure logic, highest ROI
  2. docx/word_gen.py helpers — Pure functions, easy wins
  3. file_operations.py — Safety-critical filesystem operations
  4. nfl.py (GameExtractor, MetadataWriter) — Core business logic
  5. metadata.py (MetaDataCreator) — String construction + filesystem
  6. draft_buzz.py (static methods + ProspectParser) — Scraping orchestration
  7. mcmillen.py — Small module, quick to cover