app/webservice.py

# MIT License
#
# Copyright (c) 2022 Ahmet Oner & Besim Alibegovic
# Portions Copyright (c) 2024 Team Audio
#
# Permission is hereby granted, free of charge, to any person obtaining a copy
# of this software and associated documentation files (the "Software"), to deal
# in the Software without restriction, including without limitation the rights
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
# copies of the Software, and to permit persons to whom the Software is
# furnished to do so, subject to the following conditions:
#
# The above copyright notice and this permission notice shall be included in all
# copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
# SOFTWARE.

from typing import Union, Annotated
import importlib.metadata
import logging
import os
import tempfile

from celery.result import AsyncResult
from fastapi import FastAPI, File, Query, Request, UploadFile, applications
from fastapi.openapi.docs import get_swagger_ui_html
from fastapi.responses import HTMLResponse, JSONResponse, PlainTextResponse, RedirectResponse, StreamingResponse
from fastapi.staticfiles import StaticFiles
from fastapi.templating import Jinja2Templates
from whisper import tokenizer
import aiofiles

from .util import apierror
from .worker import transcribe

logging.basicConfig(format='[%(asctime)s] [%(name)s] [%(levelname)s] %(message)s', level=logging.INFO, force=True)
logger = logging.getLogger(__name__)

ASR_ENGINE = os.getenv("ASR_ENGINE", "faster_whisper")

ASR_OPTIONS = frozenset([
    "task",
    "language",
    "hotwords",
    "initial_prompt",
    "encode",
    "output",
    "vad_filter",
    "word_timestamps",
    "model_name",
])

DEFAULT_MODEL_NAME = os.getenv("ASR_MODEL", "small")

LANGUAGE_CODES = sorted(list(tokenizer.LANGUAGES.keys()))

TASK_EXPIRATION_SECONDS = 30

projectMetadata = importlib.metadata.metadata('reaspeech')
app = FastAPI(
    # docs_url=None,
    # redoc_url=None,
    title=projectMetadata['Name'].title().replace('-', ' '),
    description=projectMetadata['Summary'],
    version=projectMetadata['Version'],
    contact={
        "url": projectMetadata['Home-page']
    },
    swagger_ui_parameters={"defaultModelsExpandDepth": -1},
    license_info={
        "name": "MIT License",
        "url": projectMetadata['License']
    }
)

assets_path = os.getcwd() + "/swagger-ui-assets"
if os.path.exists(assets_path + "/swagger-ui.css") and os.path.exists(assets_path + "/swagger-ui-bundle.js"):
    app.mount("/assets", StaticFiles(directory=assets_path), name="static")

    def swagger_monkey_patch(*args, **kwargs):
        return get_swagger_ui_html(
            *args,
            **kwargs,
            swagger_favicon_url="",
            swagger_css_url="/assets/swagger-ui.css",
            swagger_js_url="/assets/swagger-ui-bundle.js",
        )

    applications.get_swagger_ui_html = swagger_monkey_patch

static_path = os.getcwd() + "/app/static"
app.mount("/static", StaticFiles(directory=static_path), name="static")

templates_path = os.getcwd() + "/app/templates"
templates = Jinja2Templates(directory=templates_path)

output_directory = os.environ.get("OUTPUT_DIRECTORY", os.getcwd() + "/app/output")
output_url_prefix = os.environ.get("OUTPUT_URL_PREFIX", "/output")
app.mount(output_url_prefix, StaticFiles(directory=output_directory), name="output")

@app.exception_handler(apierror.APIError)
async def api_exception_handler(request: Request, exc: apierror.APIError):
    return exc.to_response()

@app.exception_handler(500)
async def internal_exception_handler(request: Request, exc: Exception):
    return apierror.error_response(exc)

@app.get("/", response_class=RedirectResponse, include_in_schema=False)
async def index():
    return "/reaspeech"

@app.get("/reaspeech", response_class=HTMLResponse, include_in_schema=False)
async def reaspeech(request: Request):
    return templates.TemplateResponse("index.html", {"request": request})

@app.get("/reascript", response_class=PlainTextResponse, include_in_schema=False)
async def reascript(request: Request, name: str, host: str, protocol: str):
    return templates.TemplateResponse("reascript.lua", {
            "request": request,
            "name": name,
            "host": host,
            "protocol": protocol
        },
        media_type='application/x-lua',
        headers={
            'Content-Disposition': f'attachment; filename="{name}.lua"'
        }
    )

@app.post("/asr", tags=["Endpoints"])
async def asr(
    task: Union[str, None] = Query(default="transcribe", enum=["transcribe", "translate"]),
    language: Union[str, None] = Query(default=None, enum=LANGUAGE_CODES),
    hotwords: Union[str, None] = Query(default=None),
    initial_prompt: Union[str, None] = Query(default=None),
    audio_file: UploadFile = File(...),
    encode: bool = Query(default=True, description="Encode audio first through ffmpeg"),
    output: Union[str, None] = Query(default="txt", enum=["txt", "vtt", "srt", "tsv", "json"]),
    vad_filter: Annotated[bool | None, Query(
        description="Enable the voice activity detection (VAD) to filter out parts of the audio without speech",
        include_in_schema=(True if ASR_ENGINE == "faster_whisper" else False)
    )] = False,
    word_timestamps: bool = Query(default=False, description="Word level timestamps"),
    model_name: Union[str, None] = Query(default=None, description="Model name to use for transcription"),
    use_async: bool = Query(default=False, description="Use asynchronous processing")
):
    asr_options = {k: v for k, v in locals().items() if k in ASR_OPTIONS}
    async_str = " (async)" if use_async else ""
    logger.info(f"Transcribing{async_str} {audio_file.filename} with {asr_options}")

    with tempfile.NamedTemporaryFile(delete=False) as temp_file:
        temp_file_path = temp_file.name

    async with aiofiles.open(temp_file_path, 'wb') as out_file:
        while content := await audio_file.read(1024 * 1024):  # Read in chunks of 1MB
            await out_file.write(content)

    transcriber = transcribe.si(temp_file_path, audio_file.filename, asr_options)

    if use_async:
        job = transcriber.apply_async(expires=TASK_EXPIRATION_SECONDS)
        return JSONResponse({"job_id": job.id})

    else:
        result = transcriber.apply().get()

        def reader():
            with open(result['output_path'], "r") as file:
                yield from file

        filename = result['output_filename']
        return StreamingResponse(
            reader(),
            media_type="text/plain",
            headers={
                'Asr-Engine': ASR_ENGINE,
                'Content-Disposition': f'attachment; filename="{filename}"'
            })

@app.get("/jobs/{job_id}", tags=["Endpoints"])
async def job_status(job_id: str):
    job = AsyncResult(job_id)

    result = {
        "job_id": job_id,
        "job_status": job.status,
        "job_result": job.result
    }

    if job.status == "FAILURE":
        result["job_result"] = apierror.error_dict(result["job_result"])

    return JSONResponse(result)

@app.delete("/jobs/{job_id}", tags=["Endpoints"])
async def revoke_job(job_id: str):
    job = AsyncResult(job_id)
    job.revoke(terminate=True)

    result = {
        "job_id": job_id,
        "job_status": job.status
    }
    return JSONResponse(result)