Added Transcribe support

.gitignore

@@ -9,4 +9,10 @@ build/
 .vscode/  
 *.log  
 *.swp  
-.DS_Store
+.DS_Store
+/*/*.bin
+*/*.bin
+
+*/*.wav
+*/*.mp3
+*/*.txt

README.md

@@ -84,6 +84,15 @@ The project structure is organized as follows:
 | /api/v1/users/{user_id}/ | delete_user          | DELETE      |
 | /api/v1/transcribe/      | post_audio           | POST        |
 
+## Install Dependecy
+```bash
+# Install ffmpeg for Audio Processing
+sudo apt install ffmpeg
+
+# Install Python Package
+pip install -r requirements.txt
+
+```
 
 ## Run this Project
 

app/api/endpoints/transcribe.py

@@ -1,22 +1,30 @@
-from fastapi import APIRouter, File, UploadFile, Request, Header
+from typing import Annotated, List, Union
+
+from fastapi import APIRouter, File, UploadFile, Request, Header, HTTPException
 from pydantic import BaseModel
 
+from app.utils.utils import save_audio_file, transcribeFile
+
 router = APIRouter()
 
 
-class AudioFile(BaseModel):
+class Transcription(BaseModel):
+    text: str
     filename: str
-    content_type: str
 
 
-@router.post("/")
+@router.post("/", response_model=Transcription)
 async def post_audio(
-    request: Request, file: UploadFile = File(...), Authorization: str = Header(...)
+    request: Request,
+    file: UploadFile = File(...),
+    Authentication: Annotated[Union[str, None], Header()] = None,
 ):
-    """Receive audio file and save it to disk."""
-    print(f"Authorization header: {Authorization}")
-
-    with open(file.filename, "wb") as f:
-        f.write(file.file.read())
+    print(f"Authorization header: {Authentication}")
 
-    return AudioFile(filename=file.filename, content_type=file.content_type)
+    try:
+        """Receive audio file and save it to disk. and then transcribe the audio file"""
+        file_path = save_audio_file(file)
+        data = transcribeFile(file_path)
+        return Transcription(filename=file.filename, text=data)
+    except Exception as e:
+        raise HTTPException(status_code=400, detail=e.__str__())

app/utils/utils.py

@@ -1,4 +1,7 @@
 import json
+import subprocess
+import uuid
+import logging
 
 
 def get_all_routes(app):
@@ -28,3 +31,38 @@ def print_routes(app):
             + f"{', '.join(route['methods'])}"
         )
     print("\n")
+
+
+def transcribeFile(path: str = None, model="ggml-model-whisper-tiny.en-q5_1.bin"):
+    """./binary/whisper -m models/ggml-tiny.en.bin -f Rev.mp3 -nt --output-text out1.txt"""
+    try:
+        if path is None:
+            raise Exception("No path provided")
+        outputFilePath: str = f"transcribe/{uuid.uuid4()}.txt"
+        command: str = f"./binary/whisper -m models/{model} -f {path} -nt --output-text {outputFilePath}"
+        execute_command(command)
+        f = open(outputFilePath, "r")
+        data = f.read()
+        f.close()
+        return data
+    except Exception as e:
+        logging.error(e)
+        raise Exception(e.__str__())
+
+
+def execute_command(command: str) -> str:
+    try:
+        result = subprocess.check_output(command, shell=True, stderr=subprocess.STDOUT)
+        return result.decode("utf-8").strip()
+    except subprocess.CalledProcessError as e:
+        logging.error(e.output.decode("utf-8").strip())
+        raise Exception("Error while transcribing")
+
+
+def save_audio_file(file=None):
+    if file is None:
+        return ""
+    path = f"audio/{uuid.uuid4()}.mp3"
+    with open(path, "wb") as f:
+        f.write(file.file.read())
+    return path