Initial commit: RehearsalHub POC

Full-stack self-hosted band rehearsal platform:

Backend (FastAPI + SQLAlchemy 2.0 async):
- Auth with JWT (register, login, /me, settings)
- Band management with Nextcloud folder integration
- Song management with audio version tracking
- Nextcloud scan to auto-import audio files
- Band membership with link-based invite system
- Song comments
- Audio analysis worker (BPM, key, loudness, waveform)
- Nextcloud activity watcher for auto-import
- WebSocket support for real-time annotation updates
- Alembic migrations (0001–0003)
- Repository pattern, Ruff + mypy configured

Frontend (React 18 + Vite + TypeScript strict):
- Login/register page with post-login redirect
- Home page with band list and creation form
- Band page with member panel, invite link, song list, NC scan
- Song page with waveform player, annotations, comment thread
- Settings page for per-user Nextcloud credentials
- Invite acceptance page (/invite/:token)
- ESLint v9 flat config + TypeScript strict mode

Infrastructure:
- Docker Compose: PostgreSQL, Redis, API, worker, watcher, nginx
- nginx reverse proxy for static files + /api/ proxy
- make check runs all linters before docker compose build

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

worker/tests/test_waveform.py

@@ -0,0 +1,49 @@
+"""Tests for waveform peak extraction (no Essentia required)."""
+
+import json
+
+import numpy as np
+import pytest
+
+from worker.pipeline.waveform import extract_peaks, generate_waveform_file, peaks_to_json
+
+
+def test_extract_peaks_returns_correct_length(sine_440hz):
+    audio, sr = sine_440hz
+    peaks = extract_peaks(audio, num_points=500)
+    assert len(peaks) == 500
+
+
+def test_extract_peaks_normalized_between_0_and_1(sine_440hz):
+    audio, sr = sine_440hz
+    peaks = extract_peaks(audio, num_points=200)
+    assert all(0.0 <= p <= 1.0 for p in peaks)
+    assert max(peaks) == pytest.approx(1.0, abs=0.01)
+
+
+def test_extract_peaks_empty_audio():
+    audio = np.array([], dtype=np.float32)
+    peaks = extract_peaks(audio, num_points=100)
+    assert len(peaks) == 100
+    assert all(p == 0.0 for p in peaks)
+
+
+def test_peaks_to_json_valid_structure(sine_440hz):
+    audio, _ = sine_440hz
+    peaks = extract_peaks(audio)
+    json_str = peaks_to_json(peaks)
+    data = json.loads(json_str)
+    assert data["version"] == 2
+    assert data["channels"] == 1
+    assert len(data["data"]) == len(peaks)
+
+
+@pytest.mark.asyncio
+async def test_generate_waveform_file_writes_json(tmp_path, sine_440hz):
+    audio, _ = sine_440hz
+    output = str(tmp_path / "waveform.json")
+    await generate_waveform_file(audio, output)
+    with open(output) as f:
+        data = json.load(f)
+    assert data["version"] == 2
+    assert len(data["data"]) == 1000