started link transcription

App/Transcription/TranscriptionRoutes.py

@@ -2,9 +2,10 @@ from fastapi import APIRouter, status, Form, UploadFile, File, Query, Background
 from typing_extensions import Annotated
 from .Schemas import UserDetails
 from App import bot
-import aiofiles
+import aiofiles, os
+import tempfile
 from celery.result import AsyncResult
-from App.Worker import transcription_task
+from App.Worker import transcription_task, downloadfile
 from App.Users.Model import User
 from .Model import Transcriptions
 from .Utils.fastapi_tasks import perform_background_task
@@ -16,6 +17,25 @@ from .Utils.fastapi_tasks import perform_background_task
 transcription_router = APIRouter(tags=["User"])
 
 
+@transcription_router.get("/download-audio")
+async def download_audio(url: str):
+    ydl_opts = {
+        "format": "bestaudio/best",
+        "postprocessors": [
+            {
+                "key": "FFmpegExtractAudio",
+                "preferredcodec": "mp3",
+                "preferredquality": "192",
+            }
+        ],
+        "outtmpl": os.path.join(tempfile.gettempdir(), "%(title)s.%(ext)s"),
+    }
+    task = downloadfile.delay(url, ydl_opts)
+    return {
+        "task_id": task.id,
+    }
+
+
 @transcription_router.post("/uploadfile/")
 async def create_file(
     background_tasks: BackgroundTasks,
@@ -42,13 +62,6 @@ async def create_file(
     finally:
         await file.close()
 
-    # telegram
-    # data = await bot.send_file(
-    #     -1001925049183,
-    #     file_size=file.size,
-    #     caption=file.filename,
-    #     file=f"./{file.filename}",
-    # )
     # celery task
     task = transcription_task.delay(file.filename, model)
 

App/Transcription/Utils/audio_transcription.py

@@ -9,7 +9,7 @@ model_size = "tiny"
 def transcribe_file(state, file_path, model_size="tiny"):
     result = {}
     metadata = TranscriptionMetadata()
-    metadata.logs = "STARTING"
+    metadata.logs = "Transcription STARTING"
     state.update_state(
         state="PROGRESS",
         meta=metadata.dict(),

App/Worker.py

@@ -2,12 +2,25 @@ from celery import Celery
 import os
 import time
 from App import celery_config
+import yt_dlp
+import tempfile
 from App.Transcription.Utils.audio_transcription import transcribe_file
 
 celery = Celery()
 celery.config_from_object(celery_config)
 
 
-@celery.task(name="create_task", bind=True)
+@celery.task(name="transcription", bind=True)
 def transcription_task(self, file_path, model_size="tiny"):
     return transcribe_file(state=self, file_path=file_path, model_size=model_size)
+
+
+@celery.task(name="download", bind=True)
+def downloadfile(self, url, ydl_opts):
+    self.update_state(state="Downloading File..", meta={})
+    with yt_dlp.YoutubeDL(ydl_opts) as ydl:
+        info = ydl.extract_info(url, download=True)
+        audio_file = ydl.prepare_filename(info)
+    self.update_state(state="Downloading complete", meta={})
+
+    return transcribe_file(state=self, file_path=audio_file, model_size="tiny")

requirements.txt

@@ -18,6 +18,7 @@ pydantic[email]
 uvicorn==0.21.1
 gunicorn
 ujson
+yt-dlp
 psutil
 orm[mysql]
 celery