fix speed parameter type; other minor tweaks

.github/workflows/docker-publish.yml

@@ -45,7 +45,9 @@ jobs:
           username: ${{ github.actor }}
           password: ${{ secrets.GITHUB_TOKEN }}
 
-      # Extract metadata (tags, labels) for Docker
+      # Extra
+            type=semver,pattern={{major}}.{{minor}}
+            type=semver,pattern={{major}}ct metadata (tags, labels) for Docker
       # https://github.com/docker/metadata-action
       - name: Extract Docker metadata
         id: meta
@@ -55,8 +57,6 @@ jobs:
             latest=${{ !github.event.release.prerelease }}
           tags: |
             type=semver,pattern={{version}}
-            type=semver,pattern={{major}}.{{minor}}
-            type=semver,pattern={{major}}
           images: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME_LOWER }}
 
       - name: Set up Docker Buildx

README.md

@@ -3,7 +3,7 @@
 An API based on FastAPI for using neural text-to-speech synthesis. The API is designed to be used together with our
 [text-to-speech workers](https://github.com/TartuNLP/text-to-speech-worker).
 
-The project is developed by the [NLP research group](https://tartunlp.ai) at the [Universty of Tartu](https://ut.ee).
+The project is developed by the [NLP research group](https://tartunlp.ai) at the [University of Tartu](https://ut.ee).
 Speech synthesis can also be tested in our [web demo](https://www.neurokone.ee/).
 
 ## API usage
@@ -32,7 +32,7 @@ in `.wav` format.
 
 The API forwards requests to various TTS engines using the RabbitMQ message broker. The communication uses a direct
 exchange named `text-to-speech` and a routing key of the request parameters using the format `text-to-speech.$speaker`.
-TTS workers estabilish queues within the exchange that are bound routing keys that illustrate which requests that the
+TTS workers establish queues within the exchange that are bound routing keys that illustrate which requests that the
 particular model can handle.
 
 ## Setup
@@ -60,7 +60,7 @@ The following environment variables should be specified when running the contain
 The entrypoint of the container
 is `["uvicorn", "app:app", "--host", "0.0.0.0", "--proxy-headers", "--log-config", "config/logging.ini"]`.
 
-The `CMD` option can be used to override flags in the entrypoint or to overridde
+The `CMD` option can be used to override flags in the entrypoint or to override
 different [Uvicorn parameters](https://www.uvicorn.org/deployment/). For example,
 `["--log-config", "config/logging.debug.ini", "--root-path", "/text-to-speech"]` enables debug logging and allows the API
 to be mounted under to the non-root path `/text-to-speech` when using a proxy server such as Nginx.

app/app.py

@@ -9,7 +9,6 @@
 from .tts import v2_router
 
 logger = logging.getLogger(__name__)
-logger.warning(api_settings)
 
 app = FastAPI(title="Text-to-Speech API",
               version=api_settings.version if api_settings.version else "dev",

app/mq_connector.py

@@ -15,7 +15,7 @@
 
 
 def uuid4():
-    """Cryptographycally secure UUID generator."""
+    """Cryptographically secure UUID generator."""
     return uuid.UUID(bytes=os.urandom(16), version=4)
 
 
@@ -51,7 +51,7 @@ async def on_response(self, message: AbstractIncomingMessage):
         if message.correlation_id in self.futures:
             body = json.loads(message.body.decode())
             if body['status_code'] != 200:
-                LOGGER.info(f"Received erronous response for request: {{"
+                LOGGER.info(f"Received erroneous response for request: {{"
                             f"id: {message.correlation_id}, "
                             f"code: {body['status_code']}}}")
             future = self.futures.pop(message.correlation_id, None)
@@ -102,7 +102,7 @@ async def publish_request(self, body: BaseModel, speaker: str) -> Tuple[dict, st
             raise HTTPException(408)
 
         if worker_response['status_code'] != 200:
-            LOGGER.info(f"Received erronous response for request: {{"
+            LOGGER.info(f"Received erroneous response for request: {{"
                         f"id: {correlation_id}, "
                         f"code: {worker_response['status_code']}}}")
             raise HTTPException(worker_response['status_code'], detail=worker_response['status'])

app/tts/v2_router.py

@@ -28,8 +28,10 @@ async def get_config():
 @v2_router.post('', response_class=Response,
                 description="Submit a text-to-speech request.",
                 responses={
+                    200: {"content": {"audio/wav": {}}, "description": "Returns the synthesized audio."},
                     422: {"model": ErrorMessage},
-                    200: {"content": {"audio/wav": {}}, "description": "Returns the synthesized audio."}
+                    408: {"model": ErrorMessage},
+                    500: {"model": ErrorMessage}
                 })
 async def synthesis(body: Request):
     content, correlation_id = await mq_connector.publish_request(body, body.speaker)
@@ -44,7 +46,12 @@ async def synthesis(body: Request):
 
 @v2_router.post('/verbose', response_model=ResponseContent,
                 description="Submit a text-to-speech request and return only some information about the output.",
-                responses={422: {"model": ErrorMessage}, 200: {"model": ResponseContent}})
+                responses={
+                    200: {"content": {"audio/wav": {}}, "description": "Returns the synthesized audio."},
+                    422: {"model": ErrorMessage},
+                    408: {"model": ErrorMessage},
+                    500: {"model": ErrorMessage}
+                })
 async def synthesis_info(body: Request):
     content, correlation_id = await mq_connector.publish_request(body, body.speaker)
     return content

app/tts/v2_schemas.py

@@ -17,7 +17,7 @@ class Request(BaseModel):
     speaker: str = Field(...,
                          example="mari",
                          description="Name of speaker (not case-sensitive).")
-    speed: int = Field(default=1,
+    speed: float = Field(default=1,
                        description="Output audio speed multiplier.",
                        ge=0.5,
                        le=2)

requirements.txt

@@ -1,5 +1,5 @@
 aio-pika==8.2.5
-fastapi==0.87.0
+fastapi==0.88.0
 uvicorn==0.20.0
 pydantic~=1.10.2
 pyyaml==6.0