Update app.py

app.py

@@ -1,23 +1,30 @@
 import os
-import csv
-from datetime import datetime
+import time
+import json
+import tempfile
+import threading
+from typing import Optional
+from datetime import datetime, timezone
+
+import httpx
+import whisper
+import nltk
+from nltk.tokenize import sent_tokenize
 from fastapi import FastAPI, UploadFile, File, Form, HTTPException
 from fastapi.middleware.cors import CORSMiddleware
-import whisper
 from transformers import pipeline
 
-# ---------------- Config ----------------
-UPLOAD_DIR = os.getenv("UPLOAD_DIR", "./uploads")
+# ---------------- Fix Whisper cache issue ----------------
+os.environ["XDG_CACHE_HOME"] = "/tmp/.cache"
 
-if not os.path.exists(UPLOAD_DIR):
-    try:
-        os.makedirs(UPLOAD_DIR, exist_ok=True)
-    except PermissionError:
-        UPLOAD_DIR = "."
+# ---------------- Config ----------------
+UPLOAD_DIR = "uploads"
+os.makedirs(UPLOAD_DIR, exist_ok=True)
 
-# ---------------- App ----------------
+# Initialize FastAPI
 app = FastAPI()
 
+# Allow CORS for all origins
 app.add_middleware(
     CORSMiddleware,
     allow_origins=["*"],
@@ -26,45 +33,23 @@ app.add_middleware(
     allow_headers=["*"],
 )
 
-# ---------------- Models ----------------
-asr_model = whisper.load_model("base")  
-summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
-
-# ---------------- Utils ----------------
-def save_to_csv(audio_file: str, transcription: str, summary: str):
-    csv_file = os.path.join(UPLOAD_DIR, "output.csv")
-    file_exists = os.path.isfile(csv_file)
+# Load Whisper model (small for speed)
+model = whisper.load_model("small")
 
-    with open(csv_file, mode="a", newline="", encoding="utf-8") as f:
-        writer = csv.writer(f)
-        if not file_exists:
-            writer.writerow(["Timestamp", "Audio_File", "Transcription", "Summary"])
-        writer.writerow([datetime.now().isoformat(), audio_file, transcription, summary])
 
-# ---------------- Routes ----------------
 @app.post("/transcribe/")
-async def transcribe_audio(file: UploadFile = File(...), summary_length: int = Form(100)):
+async def transcribe_audio(file: UploadFile = File(...)):
     try:
-        # Save uploaded file
+        # Save uploaded file temporarily
         file_path = os.path.join(UPLOAD_DIR, file.filename)
-        with open(file_path, "wb") as buffer:
-            buffer.write(await file.read())
-
-        # Transcribe
-        result = asr_model.transcribe(file_path)
-        transcription = result["text"]
+        with open(file_path, "wb") as f:
+            f.write(await file.read())
 
-        # Summarize
-        summary = summarizer(transcription, max_length=summary_length, min_length=30, do_sample=False)[0]["summary_text"]
+        # Transcribe with Whisper
+        result = model.transcribe(file_path)
 
-        # Save to CSV
-        save_to_csv(file.filename, transcription, summary)
+        # Return text output
+        return {"filename": file.filename, "transcription": result["text"]}
 
-        return {
-            "filename": file.filename,
-            "transcription": transcription,
-            "summary": summary,
-            "csv_saved": True
-        }
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))