add logging config to audio server

add audio client
change delimiter for mcp tools from "::" to ":" to save tokens

audio_server/Containerfile

@@ -16,4 +16,4 @@ RUN apt -y install ffmpeg
 WORKDIR /app
 
 EXPOSE 8000
-CMD ["uvicorn", "audio_server:app", "--host", "0.0.0.0", "--port", "8000"]
+CMD ["uvicorn", "audio_server:app", "--host", "0.0.0.0", "--port", "8000", "--log-config", "logging_config.yaml"]

audio_server/logging_config.yaml

@@ -0,0 +1,42 @@
+version: 1
+disable_existing_logger: false
+formatters:
+  main:
+    format: "\x1b[1;32m%(name)s \x1b[0m- %(levelname)s - %(message)s"
+  default:
+    format: "\x1b[1;93m%(name)s \x1b[0m- %(levelname)s - %(message)s"
+    use_colors: yes
+handlers:
+  main:
+    formatter: main
+    class: logging.StreamHandler
+    stream: ext://sys.stdout
+  default:
+    formatter: default
+    class: logging.StreamHandler
+    stream: ext://sys.stdout
+  access:
+    formatter: default
+    class: logging.StreamHandler
+    stream: ext://sys.stdout
+loggers:
+  audio_server:
+    level: INFO
+    handlers:
+      - main
+    propagate: no
+  uvicorn.error:
+    level: INFO
+    handlers:
+      - default
+    propagate: no
+  uvicorn.access:
+    level: INFO
+    handlers:
+      - default
+    propagate: no
+root:
+  level: INFO
+  handlers:
+    - default
+  propagate: no

audio_server/requirements.txt

@@ -5,4 +5,5 @@ uvicorn~=0.44.0
 starlette~=1.0.0
 python-multipart~=0.0.24
 aiofiles~=25.1.0
-piper-tts~=1.4.2
+piper-tts~=1.4.2
+PyYAML~=6.0.3

audio_server/src/audio_server/__init__.py

@@ -5,13 +5,14 @@ from wave import Wave_write
 from pathlib import Path
 
 import aiofiles
-from piper import PiperVoice, SynthesisConfig
+from piper import PiperVoice
 from piper.download_voices import download_voice
 import psutil
-from fastapi import FastAPI, UploadFile
+from fastapi import FastAPI, UploadFile, HTTPException
 import whisper
-from pydantic import BaseModel
-from fastapi.responses import FileResponse, JSONResponse
+from fastapi.responses import FileResponse
+
+from .models import TranscriptionResponse, VoiceRequest, StatusResponse
 
 app = FastAPI()
 
@@ -28,26 +29,28 @@ download_voice(piper_model_name, piper_path)
 temporary_audio_path = Path("temporary_audio")
 temporary_audio_path.mkdir(parents=True, exist_ok=True)
 
-logger = logging.getLogger(__name__)
+logger = logging.getLogger("audio_server")
+
 
 @app.get("/")
-async def index():
-    return {
-        "memory_usage": psutil.Process(os.getpid()).memory_info().rss,
-        "piper_model": piper_model_name,
-        "whisper_model": whisper_model_name,
-    }
+async def index() -> StatusResponse:
+    return StatusResponse(memory_usage=psutil.Process(os.getpid()).memory_info().rss, piper_model=piper_model_name,
+                          whisper_model=whisper_model_name)
+
 
 @app.get("/transcribe")
-async def transcribe(audio_file: UploadFile):
+async def transcribe(audio_file: UploadFile) -> TranscriptionResponse:
     file_extensions = {
+        "audio/wav": "wav",
         "audio/wave": "wav",
         "audio/mpeg": "mp3",
     }
 
     if audio_file.content_type not in file_extensions.keys():
         logger.error("%s is not supported", audio_file.content_type)
-        return JSONResponse({"detail": "wrong file type"}, status_code=400)
+        print("CONTENT TYPE:")
+        print(audio_file.content_type)
+        raise HTTPException(status_code=400, detail="file type not supported")
 
     file_path = temporary_audio_path / f"{uuid.uuid4()}.{file_extensions[audio_file.content_type]}"
 
@@ -58,14 +61,11 @@ async def transcribe(audio_file: UploadFile):
     # noinspection PyArgumentList
     transcription = whisper_model.transcribe(file_path.as_posix())
 
-    return JSONResponse(transcription)
+    return TranscriptionResponse.model_validate(transcription)
 
-class VoiceRequest(BaseModel):
-    text: str
-    config: SynthesisConfig | None
 
 @app.get("/tts")
-async def tts(voice_request: VoiceRequest):
+async def tts(voice_request: VoiceRequest) -> FileResponse:
     voice = PiperVoice.load(piper_path / f"{piper_model_name}.onnx")
 
     audio_file_path = temporary_audio_path / f"{uuid.uuid4()}.wav"
@@ -73,4 +73,4 @@ async def tts(voice_request: VoiceRequest):
     with Wave_write(audio_file_path.as_posix()) as writer:
         voice.synthesize_wav(voice_request.text, writer, syn_config=voice_request.config)
 
-    return FileResponse(audio_file_path, media_type="audio/wav")
+    return FileResponse(audio_file_path, media_type="audio/wav")

audio_server/src/audio_server/models.py

@@ -0,0 +1,32 @@
+from piper import SynthesisConfig
+from pydantic import BaseModel
+
+
+class StatusResponse(BaseModel):
+    memory_usage: int
+    piper_model: str
+    whisper_model: str
+
+
+# noinspection SpellCheckingInspection
+class TranscriptionSegment(BaseModel):
+    id: int
+    seek: int
+    start: float
+    end: float
+    text: str
+    tokens: list[int]
+    temperature: float
+    avg_logprob: float
+    compression_ratio: float
+    no_speech_prob: float
+
+class TranscriptionResponse(BaseModel):
+    text: str
+    segments: list[TranscriptionSegment]
+    language: str
+
+
+class VoiceRequest(BaseModel):
+    text: str
+    config: SynthesisConfig | None