Replace fast api with litestar. Merge pull request #9

* feat: add InvalidSampleRate exception
* feat: replace fastAPI with litestar
* chore: rename server.py to app.py
* chore: update README
* fix: run command in dockerfile
* chore: update README.md

.github/README.md

@@ -1,7 +1,7 @@
 <!-- Created in https://leviarista.github.io/github-profile-header-generator/ -->
 ![Header](./header.png)
 
-## Languages supported
+# Languages supported
 
 > [!NOTE]
 > All models are from the repository: [snakers4/silero-models](https://github.com/snakers4/silero-models)
@@ -46,10 +46,10 @@ All languages support sample rate: 8 000, 24 000, 48 000
 
 # Run API server
 ```bash
-python3 main.py
+litestar run
 ```
 > [!NOTE]
->  The default will be [localhost:8000](http://localhost:8000/docs). All endpoints can be viewed and tested at [localhost:8000/docs](http://localhost:8000/docs)
+>  The default will be [localhost:8000](http://localhost:8000/)
 
 # Run API server via docker
 ```bash
@@ -61,7 +61,7 @@ docker run --rm -p 8000:8000 twirapp/silero-tts-api-server
 
 Build the API server image:
 ```bash
-docker build --rm -f docker/Dockerfile -t silero-tts-api-server .
+docker build -f docker/Dockerfile -t silero-tts-api-server .
 ```
 
 Run the API server container:
@@ -76,6 +76,28 @@ docker-compose -f docker/compose.yml up
 
 </details>
 
+# Documentation
+You can view the automatically generated documentation based on OpenAPI at:
+
+| Provider | Url |
+|--------|--------|
+| [ReDoc](https://redocly.com/redoc) | https://localhost:8000/schema |
+| [Swagger UI](https://swagger.io) | https://localhost:8000/schema/swagger |
+| [Stoplight Elements](https://stoplight-site.webflow.io/open-source/elements) | https://localhost:8000/schema/elements |
+| [RepiDoc](https://rapidocweb.com) | https://localhost:8000/schema/repidoc |
+| OpenAPI schema yaml | https://localhost:8000/schema/openapi.yaml |
+| OpenAPI schema json | https://localhost:8000/schema/openapi.json |
+
+# Endpoints
+
+- `GET` `/generate` - Generate audio in wav format from text
+- `GET` `/speakers` - Get list of speakers
+
+# Environment variables:
+
+- `TEXT_LENGTH_LIMIT` - Maximum length of the text to be processed. Default is 930 characters.
+- `MKL_NUM_THREADS` - Number of threads to use for generating audio. Default number of threads: number of CPU cores.
+
 # Considerations for the future
 This repository is dedicated to twir.app and is designed to meet its requirements.
 

app.py

@@ -0,0 +1,70 @@
+from os import environ
+from typing import Annotated
+
+from dotenv import load_dotenv
+from litestar import Litestar, get, Response
+from litestar.openapi import OpenAPIConfig
+from litestar.config.response_cache import CACHE_FOREVER
+from litestar.params import Parameter
+
+from tts import tts
+from openapi_examples import *
+from http_exceptions import *
+from exceptions import *
+
+
+load_dotenv()
+
+SILERO_MAX_TEXT_LENGTH = 930
+text_length_limit = min(
+    int(environ.get("TEXT_LENGTH_LIMIT", SILERO_MAX_TEXT_LENGTH)),
+    SILERO_MAX_TEXT_LENGTH,
+)
+
+
+@get(
+    "/generate",
+    summary="Generate WAV audio from text",
+    media_type="audio/wav",
+    sync_to_thread=True,
+    raises=genetate_exceptions,
+)
+def generate(
+    text: Annotated[str, Parameter(examples=text_examples)],
+    speaker: Annotated[str, Parameter(examples=speaker_examples)],
+    sample_rate: Annotated[
+        int, Parameter(examples=sample_rate_examples, default=48_000)
+    ],
+) -> Response:
+    if len(text) > text_length_limit:
+        raise TextTooLongHTTPException(
+            {"text": text, "length": len(text), "max_length": text_length_limit}
+        )
+
+    try:
+        audio = tts.generate(text, speaker, sample_rate)
+    except NotFoundModelException:
+        raise NotFoundSpeakerHTTPException({"speaker": speaker})
+    except NotCorrectTextException:
+        raise NotCorrectTextHTTPException({"text": text})
+    except TextTooLongException:
+        raise TextTooLongHTTPException(
+            {"text": text, "length": len(text), "max_length": text_length_limit}
+        )
+    except InvalidSampleRateException:
+        raise InvalidSampleRateHTTPException(
+            {"sample_rate": sample_rate, "valid_sample_rates": tts.VALID_SAMPLE_RATES}
+        )
+    else:
+        return Response(audio, media_type="audio/wav")
+
+
+@get("/speakers", summary="List available speakers", cache=CACHE_FOREVER)
+async def speakers() -> dict[str, list[str]]:
+    return tts.speakers
+
+
+app = Litestar(
+    [generate, speakers],
+    openapi_config=OpenAPIConfig(title="Silero TTS API", version="1.0.0"),
+)

docker/Dockerfile

@@ -27,4 +27,4 @@ COPY --from=models-installer /app/models /app/models
 COPY --from=pip-installer /root/.local /root/.local
 ENV PATH=/root/.local/bin:$PATH
 
-CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8000"]
+CMD ["litestar", "run", "--host", "0.0.0.0", "--port", "8000"]

exceptions.py

@@ -1,11 +1,19 @@
 class NotFoundModelException(Exception):
     def __init__(self, speaker_name: str):
+        self.speaker_name = speaker_name
         super().__init__(f"Model not found for speaker: {speaker_name}")
 
 class NotCorrectTextException(Exception):
     def __init__(self, text: str):
+        self.text = text
         super().__init__(f"Text not correct: {text}")
 
 class TextTooLongException(Exception):
     def __init__(self, text: str):
-        super().__init__(f"Text too long. Length is {len(text)}. Max length is 930 symbols.")
+        self.text = text
+        super().__init__(f"Text too long. Length is {len(text)}. Max length is 930 symbols.")
+
+class InvalidSampleRateException(Exception):
+    def __init__(self, sample_rate: int) -> None:
+        self.sample_rate = sample_rate
+        super().__init__(f"Invalid sample rate {sample_rate}. Supported sample rates are 8 000, 24 000, and 48 000.")

http_exceptions.py

@@ -0,0 +1,44 @@
+from typing import Any
+
+from litestar.exceptions import HTTPException
+from litestar import status_codes as status
+
+
+class BaseHTTPException(HTTPException):
+    headers = {"Content-Type": "application/json"}
+
+    def __init__(self, extra: dict[str, Any] = None) -> None:
+        super().__init__(
+            detail=self.detail,
+            status_code=self.status_code,
+            headers=self.headers,
+            extra=extra,
+        )
+
+
+class NotFoundSpeakerHTTPException(BaseHTTPException):
+    status_code = status.HTTP_404_NOT_FOUND
+    detail = "Speaker not found"
+
+
+class NotCorrectTextHTTPException(BaseHTTPException):
+    status_code = status.HTTP_422_UNPROCESSABLE_ENTITY
+    detail = "Text is not correct"
+
+
+class TextTooLongHTTPException(BaseHTTPException):
+    status_code = status.HTTP_413_REQUEST_ENTITY_TOO_LARGE
+    detail = "Text too long"
+
+
+class InvalidSampleRateHTTPException(BaseHTTPException):
+    status_code = status.HTTP_400_BAD_REQUEST
+    detail = "Invalid sample rate"
+
+
+genetate_exceptions = [
+    NotFoundSpeakerHTTPException,
+    NotCorrectTextHTTPException,
+    TextTooLongHTTPException,
+    InvalidSampleRateHTTPException,
+]

openapi_examples.py

@@ -1,32 +1,19 @@
-from fastapi import Query
+from litestar.openapi.spec import Example
 
-TextExamples = Query(
-    openapi_examples={
-        "ru_1": {
-            "value": "Съешьте ещё этих мягких французских булочек, да выпейте чаю."
-        },
-        "ru_2": {
-            "value": "В недрах тундры выдры в гетрах тырят в вёдра ядра кедров."
-        },
-        "en_1": {
-            "value": "Can you can a canned can into an un-canned can like a canner can can a canned can into an un-canned can?"
-        },
-    }
-)
+text_examples = [
+    Example("ru_1", value="Съешьте ещё этих мягких французских булочек, да выпейте чаю."),
+    Example("ru_2", value="В недрах тундры выдры в гетрах тырят в вёдра ядра кедров."),
+    Example("en_1", value="Can you can a canned can into an un-canned can like a canner can can a canned can into an un-canned can?"),
+]
 
-SpeakerExamples = Query(
-    openapi_examples={
-        "ru_aidar": {"value": "aidar"},
-        "ru_baya": {"value": "baya"},
-        "en_0": {"value": "en_0"},
-    }
-)
+speaker_examples = [
+    Example("ru_aidar", value="aidar"),
+    Example("ru_baya", value="baya"),
+    Example("en_0", value="en_0"),
+]
 
-SampleRateExamples = Query(
-    openapi_examples={
-        "8 000": {"value": 8_000},
-        "24 000": {"value": 24_000},
-        "48 000": {"value": 48_000},
-    },
-    description="Sample rate in Hz",
-)
+sample_rate_examples = [
+    Example("8 000", value=8_000),
+    Example("24 000", value=24_000),
+    Example("48 000", value=48_000),
+]

requirements.txt

@@ -1,5 +1,5 @@
-fastapi==0.109.0
-uvicorn==0.25.0
+litestar==2.5.5
+uvicorn==0.27.1
 
 --extra-index-url https://download.pytorch.org/whl/cpu
 torch==2.2.0