feat: adaptive noise-floor loudness detection with section scoring

Replace the fixed RMS threshold with prominence over a rolling noise
floor (20th percentile per 30s block, min-smoothed so events cannot
raise their own floor, clamped to -54 dBFS). Slow ambience changes such
as rain or daytime traffic hum move the floor instead of flagging
everything; sections now need `margin` dB (default 12) of prominence.

Each section carries a score (peak dB above floor); day-highlight chips
show the top 50 by score when there are too many to list, so the most
striking events are reviewed first.

--threshold is replaced by --margin; analysis caches are now keyed by
margin+min_gap, old threshold-keyed caches never match and are
overwritten on the next analyse. Detector covered by tests/test_web.py.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

CLAUDE.md

@@ -5,7 +5,7 @@ Guidance for Claude Code when working in this repository.
 ## Rules
 
 - **Always update `README.md`** when user-facing behaviour changes (flags, endpoints, Docker setup, features), and **commit it in the same commit** as the code change. README is the external reference; CLAUDE.md documents internals.
-- Run `python -m pytest tests/` after changing `isr.py` (tests cover the recorder only).
+- Run `python -m pytest tests/` after changing `isr.py` or `web.py` (tests cover the recorder and the loud-section detector).
 
 ## Files
 
@@ -21,7 +21,7 @@ Guidance for Claude Code when working in this repository.
 
 ```bash
 python isr.py [config.ini]        # recorder; --list-devices to list ALSA inputs
-python web.py                     # web UI on :8080 (--dir, --port, --threshold, --min-gap, --analyses-dir)
+python web.py                     # web UI on :8080 (--dir, --port, --margin, --min-gap, --analyses-dir)
 python -m pytest tests/           # test suite
 docker compose up -d / down       # web UI mapped to host port 8050
 ```
@@ -35,7 +35,8 @@ Dependencies: `requests` (streams), `numpy` + `soundfile` (FLAC output and FLAC/
 - **Split timing:** files split at clock-aligned boundaries (`get_next_split_time()`), e.g. `split_minutes = 60` → on the hour.
 - **ALSA:** capture spawns `arecord` as a subprocess, raw PCM read in 100 ms chunks by a thread. Device spec resolution: `default` → exact `hw:X,Y` → partial name → fallback to any literal ALSA PCM name (so `shared_mic` from asound.conf works without appearing in `arecord -l`).
 - **Shutdown:** SIGTERM is converted to KeyboardInterrupt in `main()`; `RecorderManager.stop()` joins all threads against a single shared 25 s deadline to stay inside Docker's `stop_grace_period: 30s`.
-- **Analysis cache:** results stored as `<analyses-dir>/<file>.analysis.json` keyed by threshold+min_gap; orphans pruned at web startup. In Docker the recordings mount is **read-only** for the web container, so the cache uses a separate `./analyses` bind mount. The `threshold` and `min_gap` keys MUST stay first in the cache JSON — `_cached_analysis_params()` reads only the first 256 bytes to avoid parsing the large embedded result.
+- **Loud-section detection is adaptive:** per-window dB is compared against a rolling noise floor (`NOISE_PERCENTILE`-th percentile per `NOISE_BLOCK_SECONDS` block, min-smoothed over ±2 blocks so events can't raise their own floor; clamped to ≥ `MIN_RMS`). A section needs `margin` dB of prominence and carries a `score` (peak dB above floor) used for ranking. Tests in `tests/test_web.py`.
+- **Analysis cache:** results stored as `<analyses-dir>/<file>.analysis.json` keyed by margin+min_gap; orphans pruned at web startup. In Docker the recordings mount is **read-only** for the web container, so the cache uses a separate `./analyses` bind mount. The `margin` and `min_gap` keys MUST stay first in the cache JSON — `_cached_analysis_params()` reads only the first 256 bytes to avoid parsing the large embedded result. Old `threshold`-keyed caches never match and get overwritten on the next analyse.
 - **Analyze responses:** `/api/analyze` returns `rms_display` (~800 points), never the full per-window RMS list — the UI doesn't use it and it is ~45x larger.
 - **HTTP/1.1 keep-alive:** `_Handler.protocol_version = 'HTTP/1.1'`; every response path must set an accurate `Content-Length`. `_copy_to_response()` force-closes the connection if it under-delivers (file truncated mid-serve).
 - **Live playback:** for files listed in status.json, `/stream/` patches the header on the fly so the browser sees the duration recorded so far and can seek; responses get `Cache-Control: no-store`. WAV: `_live_wav_header` derives sizes from the byte count. FLAC: `_live_flac_header` parses the sample count out of the last frame header in the file tail (CRC-8-verified to reject false sync matches) and rewrites STREAMINFO total_samples — duration is NOT derivable from byte size for FLAC.