feat: add FastAPI webapp for Spotify/Last.FM library management

[jmlrt]

Summary

• Adds a server-rendered web interface (FastAPI + Jinja2 + HTMX + Pico CSS) that imports existing spotfm.* library functions directly — CLI unchanged
• Session-based auth via [web] api_key in spotfm.toml; background jobs tracked in-memory with HTMX polling for status
• Docker deployment via Dockerfile + docker-compose.yml with volume mounts for DB, config, and token cache

Features

Route	Description
GET /	Dashboard: track/playlist/artist/album counts
GET /playlists	Playlists with track counts + per-playlist update buttons
GET /tracks	Track list with filters: playlist, artist, album, year range
GET /duplicates	Duplicate detection by ID (sync) or name/fuzzy (async job)
GET /scrobbles	Recent Last.FM scrobbles
GET /track-counts	track_counts_log CSV visualized as a table
POST /jobs/update-playlists	Background job: update playlists from Spotify API
POST /jobs/discover	Background job: discover new tracks
GET /jobs/{job_id}	HTMX polling partial for job status

Tech stack

• FastAPI 0.115+ — async routes, factory pattern (uvicorn --factory)
• Jinja2 + HTMX 2.x — server-rendered, zero JS build step
• Pico CSS 2.x — classless minimal styling (CDN)
• Starlette SessionMiddleware — signed cookie, api_key as secret
• asyncio.create_task + run_in_executor — background jobs with in-memory registry

Test plan

• uv run pytest tests/web/ — 25 web tests (auth, routes, jobs) pass
• uv run pytest — full suite (340 tests) passes, no regressions
• uv run spfm-web — app starts; login page at http://localhost:8000
• Manual: log in, navigate dashboard, playlists, track filters
• Manual: trigger update-playlists or discover, observe HTMX job polling
• docker compose up --build — container starts, volume mounts work

🤖 Generated with Claude Code