Release Notes – v0.3.0

⚠️ Before updating

Port change: WayPointDB is now served on port 8080 instead of 80 (docker-compose.yml). Adjust your reverse proxy or the port mapping if needed.
Photon settings moved to .env: set PHOTON_SERVER_HOST, PHOTON_SERVER_HTTPS and PHOTON_SERVER_API_KEY in the .env file. Values in docker-compose.yml are no longer read from there. Wrap API keys with special characters in single quotes.
Rebuild both containers so the new frontend files are served: docker-compose up -d --build.
The first start creates new database indexes, which can take a moment on large databases.

Redesigned web interface:
- Sidebar navigation with Explore / Data / Settings sections and a trace switcher.
- Light and dark mode, following the system setting with a manual override.
- Consistent tables, forms and buttons, and confirmation messages for actions.
- Full-height map with range presets, a month picker and a collapsible layers panel.
- The per-user background image is sharper, uses the map's speed colors and shows through softly behind translucent panels.
- Works on phones.
Scheduled jobs: every user can schedule jobs to run once a day at a chosen hour. Set TZ in .env (e.g. TZ=Europe/Berlin) to use local time.
Job history: the Jobs page shows running, queued and recently finished jobs, including results and errors.
Geocoding: parallel requests (PHOTON_SERVER_CONCURRENCY, default 4), one request per ~11 m area instead of per point, and clear errors when the Photon server is unreachable or rejects requests.
Imports: status per file, clearer errors for invalid files.

Jobs run as database queries instead of loading all points into memory. On 100k points: statistics ~7× faster, speed data ~45×, delete duplicates ~160×, reduce clusters ~130×.
New indexes on GPS data and daily statistics speed up most pages.
The map no longer rebuilds itself on every refresh and draws tracks on a canvas.

Filter GPS jumps (previously "Filter large speed") only removes the jumped point, not the correct point after it.
Reduce clusters keeps the first and last point of every stay, so time spent there is preserved.
Imports skip invalid entries instead of storing them with the current time, and store timestamps with a timezone offset as UTC.
Choosing an import on the Points page shows all of its points.