Skip to content

Latest commit

 

History

History
80 lines (62 loc) · 3.33 KB

README.md

File metadata and controls

80 lines (62 loc) · 3.33 KB

Tests

Use Python 3.12, matching CI. From the project root:

requirements.txt contains runtime dependencies. requirements-test.txt includes that list plus pytest, coverage, Playwright, and dependency validation tools.

python3 -m venv .venv
source .venv/bin/activate
python -m pip install -r requirements-test.txt
python -m playwright install chromium
python -m pytest --cov --cov-report=term-missing

On Linux, use python -m playwright install --with-deps chromium to install browser system dependencies too. Browser tests use installed Google Chrome on macOS when available; otherwise they use Playwright's Chromium.

For Python tests without launching a browser:

python -m pytest -m 'not browser' --cov --cov-report=term-missing

For browser interaction checks only:

python -m pytest -m browser

Coverage includes branch coverage and all project Python modules, including the two image preparation scripts under resources. The minimum is 85%. GitHub Actions runs the complete suite and uploads coverage.xml.

Scope and isolation

  • AI: response parsing, failed requests, model/key forwarding, filling config.
  • Maps: real Pillow filling/compositing on small images, numeric range edges, text colors, optional features, output files, bundled fonts/coordinates.
  • CLI: help, defaults, options, invalid arguments, generation failures.
  • Web: pages, generation, API-key behavior, all service proxies, streaming, timeouts and invalid upstream responses.
  • Worker: metadata/downloads, audio slices, expiry cleanup, validation, safe downloads, title fitting, real FFmpeg portrait rendering and audio attachment. Audio checks cover song titles independent of video headlines and inline cut fragments after worker files are lost.
  • Browser: font/size/reset controls, exact title PNGs, debounce, responsive preview scaling, map replacement/removal and blob URL cleanup; automatic map transfer between studios, persistent settings and uploaded materials, large-map preparation for video, and graceful handling of unavailable browser storage.
  • Image preparation tools: temporary input/output images only.

Tests mock external AI, HTTP and yt-dlp downloads. An autouse fixture rejects unmocked Requests calls and AI clients. No real credentials are needed. Audio slicing mocks pydub's file backend; actual video/audio encoding uses the FFmpeg binary provided by imageio-ffmpeg (installed with MoviePy), without GPU hardware.

Browser HTTP traffic is routed to Flask test clients. External CDN scripts, styles and fonts are stubbed, so browser tests verify behavior rather than visual appearance of the external Tailwind stylesheet. Live Gemini/YouTube availability and NVIDIA/VideoToolbox encoders are not covered.

Tests use temporary folders and leave repository maps and user outputs intact.