mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-07-30 19:09:28 +00:00
f76b2b47aa
Two hand-written fixes in the voice input path: - _handle_voice_channel_input now resolves the bound text channel's channel_prompt via the adapter's _resolve_channel_prompt so voice input gets the same per-channel context as typed messages (fixes #50149). - _enrich_message_with_transcription now guards success=True results whose transcript is empty/whitespace-only (silence, cut-off, inaudible audio): instead of emitting empty quotes the agent gets a clear sentinel note. Reimplemented against the current plain-quoted note wording; original concept and tests by @deacon-botdoctor in PR #41603 (fixes #41603).
249 lines
8.4 KiB
Python
249 lines
8.4 KiB
Python
"""Gateway STT config tests — honor stt.enabled: false from config.yaml."""
|
|
|
|
from pathlib import Path
|
|
from unittest.mock import AsyncMock, patch
|
|
|
|
import pytest
|
|
import yaml
|
|
|
|
from gateway.config import GatewayConfig, Platform, load_gateway_config
|
|
from gateway.platforms.base import MessageEvent, MessageType
|
|
from gateway.session import SessionSource
|
|
|
|
|
|
def test_gateway_config_stt_disabled_from_dict_nested():
|
|
config = GatewayConfig.from_dict({"stt": {"enabled": False}})
|
|
assert config.stt_enabled is False
|
|
|
|
|
|
def test_load_gateway_config_bridges_stt_enabled_from_config_yaml(tmp_path, monkeypatch):
|
|
hermes_home = tmp_path / ".hermes"
|
|
hermes_home.mkdir()
|
|
(hermes_home / "config.yaml").write_text(
|
|
yaml.dump({"stt": {"enabled": False}}),
|
|
encoding="utf-8",
|
|
)
|
|
|
|
monkeypatch.setenv("HERMES_HOME", str(hermes_home))
|
|
monkeypatch.setattr(Path, "home", lambda: tmp_path)
|
|
|
|
config = load_gateway_config()
|
|
|
|
assert config.stt_enabled is False
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_enrich_message_with_transcription_surfaces_path_when_stt_disabled():
|
|
from gateway.run import GatewayRunner
|
|
|
|
runner = GatewayRunner.__new__(GatewayRunner)
|
|
runner.config = GatewayConfig(stt_enabled=False)
|
|
runner._has_setup_skill = lambda: True # Should NOT be consulted in disabled branch.
|
|
|
|
with patch(
|
|
"tools.transcription_tools.transcribe_audio",
|
|
side_effect=AssertionError("transcribe_audio should not be called when STT is disabled"),
|
|
), patch(
|
|
"gateway.run._probe_audio_duration",
|
|
new=AsyncMock(return_value="0:12"),
|
|
):
|
|
result, transcripts = await runner._enrich_message_with_transcription(
|
|
"caption",
|
|
["/tmp/voice.ogg"],
|
|
)
|
|
|
|
assert "/tmp/voice.ogg" in result
|
|
assert "voice message" in result.lower()
|
|
assert "(duration: 0:12)" in result
|
|
assert "caption" in result
|
|
assert transcripts == []
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_enrich_message_with_transcription_omits_duration_on_probe_failure():
|
|
from gateway.run import GatewayRunner
|
|
|
|
runner = GatewayRunner.__new__(GatewayRunner)
|
|
runner.config = GatewayConfig(stt_enabled=False)
|
|
|
|
with patch(
|
|
"gateway.run._probe_audio_duration",
|
|
new=AsyncMock(return_value=None),
|
|
):
|
|
result, transcripts = await runner._enrich_message_with_transcription(
|
|
"",
|
|
["/tmp/voice.ogg"],
|
|
)
|
|
|
|
assert "/tmp/voice.ogg" in result
|
|
assert "duration" not in result.lower()
|
|
assert transcripts == []
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_enrich_message_with_transcription_avoids_bogus_no_provider_message_for_backend_key_errors():
|
|
from gateway.run import GatewayRunner
|
|
|
|
runner = GatewayRunner.__new__(GatewayRunner)
|
|
runner.config = GatewayConfig(stt_enabled=True)
|
|
|
|
with patch(
|
|
"tools.transcription_tools.transcribe_audio",
|
|
return_value={"success": False, "error": "VOICE_TOOLS_OPENAI_KEY not set"},
|
|
):
|
|
result, transcripts = await runner._enrich_message_with_transcription(
|
|
"caption",
|
|
["/tmp/voice.ogg"],
|
|
)
|
|
|
|
assert "No STT provider is configured" not in result
|
|
assert "[voice message could not be transcribed]" in result
|
|
# The opaque backend cause must NOT leak into the LLM-visible prompt.
|
|
assert "VOICE_TOOLS_OPENAI_KEY" not in result
|
|
assert "caption" in result
|
|
assert transcripts == []
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_enrich_message_with_transcription_returns_tuple_for_empty_content_placeholder():
|
|
"""A successful transcription whose caption is the empty-content placeholder
|
|
must still return the ``(text, transcripts)`` tuple.
|
|
|
|
The Discord adapter delivers a captionless voice note as the literal
|
|
``"(The user sent a message with no text content)"`` placeholder. When STT
|
|
succeeds we strip that redundant placeholder and return just the transcript
|
|
prefix — but the method's contract (and every caller, which unpacks the
|
|
result as ``text, transcripts = ...``) requires a 2-tuple. Returning a bare
|
|
string here raised ``ValueError: too many values to unpack`` and dropped the
|
|
whole voice message on the floor.
|
|
"""
|
|
from gateway.run import GatewayRunner
|
|
|
|
runner = GatewayRunner.__new__(GatewayRunner)
|
|
runner.config = GatewayConfig(stt_enabled=True)
|
|
runner._has_setup_skill = lambda: False
|
|
|
|
with patch(
|
|
"tools.transcription_tools.transcribe_audio",
|
|
return_value={
|
|
"success": True,
|
|
"transcript": "hello from a captionless voice note",
|
|
"provider": "local_command",
|
|
},
|
|
):
|
|
result, transcripts = await runner._enrich_message_with_transcription(
|
|
"(The user sent a message with no text content)",
|
|
["/tmp/voice.ogg"],
|
|
)
|
|
|
|
# The redundant placeholder is stripped, leaving only the transcript prefix.
|
|
assert "hello from a captionless voice note" in result
|
|
assert "(The user sent a message with no text content)" not in result
|
|
# Crucially, the transcripts are still surfaced so callers can echo them.
|
|
assert transcripts == ["hello from a captionless voice note"]
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
("user_text", "expected_text"),
|
|
[
|
|
("caption", "[voice message could not be transcribed]\n\ncaption"),
|
|
("", "[voice message could not be transcribed]"),
|
|
(
|
|
"(The user sent a message with no text content)",
|
|
"[voice message could not be transcribed]",
|
|
),
|
|
],
|
|
)
|
|
@pytest.mark.asyncio
|
|
async def test_enrich_message_with_transcription_handles_missing_transcription_module_gracefully(
|
|
user_text,
|
|
expected_text,
|
|
):
|
|
from gateway.run import GatewayRunner
|
|
|
|
runner = GatewayRunner.__new__(GatewayRunner)
|
|
runner.config = GatewayConfig(stt_enabled=True)
|
|
|
|
real_import = __import__
|
|
|
|
def fake_import(name, globals=None, locals=None, fromlist=(), level=0):
|
|
if name == "tools.transcription_tools":
|
|
raise ModuleNotFoundError("No module named 'tools.transcription_tools'")
|
|
return real_import(name, globals, locals, fromlist, level)
|
|
|
|
with patch("builtins.__import__", side_effect=fake_import):
|
|
result, transcripts = await runner._enrich_message_with_transcription(
|
|
user_text,
|
|
["/tmp/voice.ogg"],
|
|
)
|
|
|
|
assert result == expected_text
|
|
assert transcripts == []
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_enrich_message_with_transcription_guards_empty_transcript():
|
|
"""success=True with an empty/whitespace transcript must not emit empty
|
|
quotes — it gets a sentinel note and is excluded from transcripts (#41603)."""
|
|
from gateway.run import GatewayRunner
|
|
|
|
runner = GatewayRunner.__new__(GatewayRunner)
|
|
runner.config = GatewayConfig(stt_enabled=True)
|
|
runner._has_setup_skill = lambda: False
|
|
|
|
with patch(
|
|
"tools.transcription_tools.transcribe_audio",
|
|
return_value={"success": True, "transcript": " \n\t", "provider": "local_command"},
|
|
):
|
|
result, transcripts = await runner._enrich_message_with_transcription(
|
|
"caption",
|
|
["/tmp/voice.ogg"],
|
|
)
|
|
|
|
assert "empty or inaudible" in result
|
|
assert '""' not in result
|
|
assert transcripts == []
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_prepare_inbound_message_text_transcribes_queued_voice_event():
|
|
from gateway.run import GatewayRunner
|
|
|
|
runner = GatewayRunner.__new__(GatewayRunner)
|
|
runner.config = GatewayConfig(stt_enabled=True)
|
|
runner.adapters = {}
|
|
runner._model = "test-model"
|
|
runner._base_url = ""
|
|
runner._has_setup_skill = lambda: False
|
|
|
|
source = SessionSource(
|
|
platform=Platform.TELEGRAM,
|
|
chat_id="123",
|
|
chat_type="dm",
|
|
)
|
|
event = MessageEvent(
|
|
text="",
|
|
message_type=MessageType.VOICE,
|
|
source=source,
|
|
media_urls=["/tmp/queued-voice.ogg"],
|
|
media_types=["audio/ogg"],
|
|
)
|
|
|
|
with patch(
|
|
"tools.transcription_tools.transcribe_audio",
|
|
return_value={
|
|
"success": True,
|
|
"transcript": "queued voice transcript",
|
|
"provider": "local_command",
|
|
},
|
|
):
|
|
result = await runner._prepare_inbound_message_text(
|
|
event=event,
|
|
source=source,
|
|
history=[],
|
|
)
|
|
|
|
assert result is not None
|
|
# Success path: the transcript passes through as a plain quoted line, with
|
|
# no "voice message" meta-commentary that the LLM would echo back.
|
|
assert "queued voice transcript" in result
|