feat(voice_worker): free a speaker's place when they leave the voice channel

- New speaker_leave message: ends the speaker's open utterance (its final is still sent) and drops
  their state, so a long session never hits the 32-speaker limit as people come and go.
- Spec and test cover more than 32 speakers in one session, each leaving mid-utterance.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
javis-bot
2026-11-11 11:14:25 +09:00
parent 1309525aa8
commit 5838f9edd8
3 changed files with 40 additions and 3 deletions

View File

@@ -246,8 +246,8 @@ class Session:
self.error("bad_message", "not an object")
return
kind = msg.get("type")
handler = {"speaking_end": self._on_speaking_end, "config": self._on_config, "tts": self._on_tts,
"tts_cancel": self._on_tts_cancel}.get(kind)
handler = {"speaking_end": self._on_speaking_end, "speaker_leave": self._on_speaker_leave,
"config": self._on_config, "tts": self._on_tts, "tts_cancel": self._on_tts_cancel}.get(kind)
if handler is None:
self.error("bad_message", f"unknown type: {kind!r}")
return
@@ -258,6 +258,16 @@ class Session:
if st is not None:
self._handle_events(msg["speaker"], st, st.segmenter.speaking_end(time.monotonic()))
def _on_speaker_leave(self, msg: dict) -> None:
# The speaker left the voice channel: finish any open utterance, then free their place.
speaker = msg.get("speaker")
st = self.speakers.pop(speaker, None) if isinstance(speaker, str) else None
if st is not None:
self._handle_events(speaker, st, st.segmenter.speaking_end(time.monotonic()))
if self.worker.scheduler:
self.worker.scheduler.drop_partials(lambda key: key == (self.sid, speaker))
debug_log(f"speaker {speaker} left session {self.sid}", "voice_worker")
def _on_config(self, msg: dict) -> None:
changes = {k: v for k, v in msg.items() if k != "type"}
try:

View File

@@ -54,6 +54,7 @@ token closes with 4401.
```
-> binary 0x01 | u8 n | speaker id (n bytes UTF-8) | PCM16LE samples in the hello "audio" format
-> {"type": "speaking_end", "speaker": "..."} // Discord speaking stopped for this speaker
-> {"type": "speaker_leave", "speaker": "..."} // the speaker left the voice channel
-> {"type": "config", ...any STT settings...} // takes effect immediately
<- {"type": "speech_start", "speaker": "...", "utt": 3}
<- {"type": "partial", "speaker": "...", "utt": 3, "text": "..."}
@@ -61,6 +62,10 @@ token closes with 4401.
"audio_ms": 2310, "wait_ms": 160, "stt_ms": 120}
```
`speaker_leave` ends an open utterance as `speaking_end` does (its final is still sent) and frees the
speaker's state. A session holds at most 32 speakers at once, so the app sends it whenever someone
leaves; a speaker who comes back starts again at `utt` 1.
`utt` counts utterances per speaker. `final` always closes an utterance that sent `speech_start`, even
when the text is empty (noise that Whisper rejected). `wait_ms` is from the moment the end was decided
to the moment the final was sent; `stt_ms` is the model time of the batch it went through.
@@ -179,6 +184,6 @@ never close the connection after the handshake.
## Debug logging
`debug_log(..., "voice_worker")` at: connection accepted or refused, settings changed, utterance end
`debug_log(..., "voice_worker")` at: connection accepted or refused, settings changed, speaker left, utterance end
(speaker, reason, length), each STT batch (kind, size, model time), each TTS sentence (length, time),
cancel, and engine failures.

View File

@@ -328,3 +328,25 @@ def test_48k_stereo_input_reaches_the_model_at_16k():
final = asyncio.run(with_worker(body, transcriber=t))[-1]
assert final["text"] == "level 0.9"
assert 0.5 * 16000 <= lengths[0] <= 0.9 * 16000
@pytest.mark.unit
def test_speakers_who_leave_free_their_place_and_finish_their_utterance():
from voice_worker.server import MAX_SPEAKERS
async def body(url):
async with connect(url) as ws:
await hello(ws, stt={"partials": False})
seen = []
for i in range(MAX_SPEAKERS + 8):
await stream(ws, f"user{i}", 0.9, 400)
# Leaving mid-utterance still delivers that utterance's final.
await ws.send(json.dumps({"type": "speaker_leave", "speaker": f"user{i}"}))
seen += await recv_until(ws, lambda m: m.get("type") in ("final", "error"))
return seen
seen = asyncio.run(with_worker(body))
assert not [m for m in seen if m.get("type") == "error"]
finals = [m for m in seen if m.get("type") == "final"]
assert [m["speaker"] for m in finals] == [f"user{i}" for i in range(len(finals))]
assert len(finals) > MAX_SPEAKERS and all(m["text"] == "level 0.9" for m in finals)