feat: name cut clips by wall-clock time; fix recording filename format

Cut downloads were named by byte offsets (`..._cut_740s-750s.flac`). They are
now named by the actual recording time the slice covers, e.g.
`20260523_22-31-30_22-32-30.flac` for a 22:31:30->22:32:30 cut of a recording
started at 22:00:00.

To make this reliable, the recording filename is now a fixed
`%Y%m%d_%H%M%S` start-time format (`FILENAME_FORMAT`) shared by isr.py and
web.py, replacing the user-configurable `filename_pattern` (web.py never reads
config.ini, so a custom pattern could not be parsed back). web.py parses the
start time out of the filename via `_recording_start()` and builds cut names
with `_cut_filename()`. The DATE column now also comes from the filename
(falling back to mtime only for non-standard names), since mtime is the last
write, not the start.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

web.py

@@ -24,7 +24,7 @@ import subprocess
 import tempfile
 import threading
 import wave
-from datetime import datetime
+from datetime import datetime, timedelta
 from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
 from pathlib import Path
 from urllib.parse import parse_qs, unquote, urlparse
@@ -58,6 +58,12 @@ MIN_RMS             = 0.002  # ≈ −54 dBFS; the floor never drops below this,
 
 CLIP_MAX_SECONDS = 600     # upper bound on /api/clip length
 
+# Recording filenames encode the start time as this strftime format (kept in
+# sync with isr.FILENAME_FORMAT). It is the authoritative recording start and
+# the only reliable clock anchor — mtime drifts to the last write, so cut clip
+# names and the displayed date are both derived from this.
+FILENAME_FORMAT = '%Y%m%d_%H%M%S'
+
 MIME_TYPES = {
     '.wav':  'audio/wav',
     '.mp3':  'audio/mpeg',
@@ -72,6 +78,35 @@ MIME_TYPES = {
 # Audio analysis helpers
 # ---------------------------------------------------------------------------
 
+def _recording_start(stem: str):
+    """Parse the recording start time encoded in a filename stem.
+
+    Returns a datetime, or None if the stem is not in FILENAME_FORMAT
+    (e.g. a manually renamed file). strptime ignores any extension because
+    callers pass Path.stem.
+    """
+    try:
+        return datetime.strptime(stem, FILENAME_FORMAT)
+    except ValueError:
+        return None
+
+
+def _cut_filename(stem: str, ext: str, start: float, end: float) -> str:
+    """Name a cut by the real wall-clock span it covers.
+
+    For a recording that started at 22:00:00, a 22:31:30→22:32:30 slice
+    (start=1890, end=1950) becomes ``20260523_22-31-30_22-32-30.flac``.
+    Falls back to the source stem plus second offsets when the filename is
+    not in FILENAME_FORMAT (e.g. a manually renamed recording).
+    """
+    started = _recording_start(stem)
+    if started is None:
+        return f'{stem}_cut_{int(start)}s-{int(end)}s{ext}'
+    cut_start = started + timedelta(seconds=start)
+    cut_end   = started + timedelta(seconds=end)
+    return f'{cut_start:%Y%m%d}_{cut_start:%H-%M-%S}_{cut_end:%H-%M-%S}{ext}'
+
+
 def _live_wav_header(path: Path, size: int):
     """Return the WAV header (through the 'data' chunk header) with RIFF and
     data sizes rewritten to match the current file size, or None.
@@ -500,6 +535,11 @@ def list_files(recordings_dir: str):
         rel      = str(path.relative_to(base)).replace('\\', '/')
         is_active = rel in active_files
 
+        # The recording start is encoded in the filename and is the true clock
+        # anchor; mtime is only a fallback for files not in FILENAME_FORMAT.
+        started = _recording_start(path.stem)
+        date    = (started or datetime.fromtimestamp(stat.st_mtime)).strftime('%Y-%m-%d %H:%M:%S')
+
         # Skip reading partial headers for in-progress files — the WAV nframes
         # field and FLAC total_samples are both unfinalized while recording,
         # producing wildly incorrect values (e.g. 53375995583:39:01).
@@ -509,7 +549,7 @@ def list_files(recordings_dir: str):
             'name':      rel,
             'size':      stat.st_size,
             'mtime':     stat.st_mtime,
-            'date':      datetime.fromtimestamp(stat.st_mtime).strftime('%Y-%m-%d %H:%M:%S'),
+            'date':      date,
             'duration':  duration,
             'ext':       path.suffix.lower().lstrip('.'),
             'recording': is_active,
@@ -831,7 +871,7 @@ class _Handler(BaseHTTPRequestHandler):
             return
 
         ext      = path.suffix.lower()
-        out_name = f'{path.stem}_cut_{int(start)}s-{int(end)}s{ext}'
+        out_name = _cut_filename(path.stem, ext, start, end)
 
         # For lossless formats, re-encode (not copy) so the container header
         # is rewritten with the correct duration/size. For lossy formats,