Assistant answers general questions and stays silent for overheard talk ("Ignored." is not said); stt prompt setting sent as the transcription prompt
Docs / docs (push) Successful in 39s
Playwright Tests / test-playwright (1, 2) (push) Successful in 4m26s
Playwright Tests / test-playwright (2, 2) (push) Successful in 2m25s
pre-commit / pre-commit (push) Successful in 3m24s
Test Backend / test-backend (3.12) (push) Successful in 6m3s
Test Backend / test-backend (3.14) (push) Successful in 5m41s
Compose Smoke Test / test-compose (push) Successful in 35s
Playwright Tests / merge-reports (push) Successful in 1m20s

This commit is contained in:
Melvin Strobl
2026-11-06 07:27:44 +01:00
parent 4c15f4abe1
commit 261cf1ec2a
6 changed files with 93 additions and 24 deletions
+23 -3
View File
@@ -186,7 +186,18 @@ class SttNode(_Speech):
alone.
"""
__slots__ = ("_heard", "_listening", "_lock", "_session", "_tasks")
class Params(BaseModel):
model_config = ConfigDict(extra="allow")
prompt: str = Field(
default="",
description=(
"Words the speech is likely to hold, such as the names of rooms "
"and devices, given to the server as its transcription prompt."
),
)
__slots__ = ("cfg", "_heard", "_listening", "_lock", "_session", "_tasks")
def __init__(
self,
@@ -195,6 +206,7 @@ class SttNode(_Speech):
params: dict[str, Any] | None = None,
name: str | None = None,
):
self.cfg = self.Params.model_validate(params or {})
self._lock = threading.Lock()
#: The chunks of the utterance being spoken, None between utterances.
self._session: asyncio.Queue[bytes | None] | None = None
@@ -293,7 +305,11 @@ class SttNode(_Speech):
try:
async with asyncio.timeout(UTTERANCE_S):
async for final, text in client.transcribe_stream(
url + "/v1", chunks, rate=RATE, idle_s=IDLE_S
url + "/v1",
chunks,
rate=RATE,
idle_s=IDLE_S,
prompt=self.cfg.prompt,
):
words = text.strip() if final else words + text
if final and words:
@@ -317,7 +333,11 @@ class SttNode(_Speech):
async def _clip(self, url: str, audio: dict[str, Any], data: bytes) -> None:
try:
text = await client.transcribe(
url + "/v1", data, "speech.wav", str(audio.get("media_type"))
url + "/v1",
data,
"speech.wav",
str(audio.get("media_type")),
prompt=self.cfg.prompt,
)
except Exception as exc:
logger.warning("'%s' lost a clip: %s: %s", self.id, type(exc).__name__, exc)