Update app.py

app.py

@@ -1,11 +1,17 @@
 import os
-import pandas as pd
+import csv
 import whisper
 import gradio as gr
 from transformers import pipeline
 from fastapi import FastAPI, UploadFile, File
 from fastapi.middleware.wsgi import WSGIMiddleware
 
+# ================== DIRECTORIES =====================
+UPLOAD_DIR = "./uploads"
+CSV_DIR = "./csv_output"
+os.makedirs(UPLOAD_DIR, exist_ok=True)
+os.makedirs(CSV_DIR, exist_ok=True)
+
 # ================== LOAD MODELS =====================
 # Load Whisper (Speech-to-Text)
 whisper_model = whisper.load_model("base")
@@ -36,10 +42,14 @@ def transcribe_and_summarize(audio_path):
     except Exception as e:
         summary = f"Summarization failed: {e}"
 
-    # Save transcription + summary to CSV
-    os.makedirs("./outputs", exist_ok=True)
-    df = pd.DataFrame([{"transcription": transcription, "summary": summary}])
-    df.to_csv("./outputs/transcriptions.csv", mode='a', index=False, header=not os.path.exists("./outputs/transcriptions.csv"))
+    # Save to CSV
+    csv_file = os.path.join(CSV_DIR, "transcription_summary.csv")
+    file_exists = os.path.isfile(csv_file)
+    with open(csv_file, mode="a", newline="", encoding="utf-8") as f:
+        writer = csv.writer(f)
+        if not file_exists:
+            writer.writerow(["Audio File", "Transcription", "Summary"])
+        writer.writerow([os.path.basename(audio_path), transcription, summary])
 
     # API URL (replace with your deployed Hugging Face Space URL)
     api_link = "https://your-username-your-space-name.hf.space/run/predict"
@@ -64,12 +74,10 @@ api_app = FastAPI(title="Audio Transcriber + Summarizer API")
 
 @api_app.post("/api/transcribe/")
 async def transcribe_api(file: UploadFile = File(...)):
-    # Save uploaded file temporarily
-    temp_path = f"/tmp/{file.filename}"
+    temp_path = os.path.join(UPLOAD_DIR, file.filename)
     with open(temp_path, "wb") as f:
         f.write(await file.read())
 
-    # Call the same function as Gradio
     transcription, summary, api_link = transcribe_and_summarize(temp_path)
     return {"transcription": transcription, "summary": summary, "api_url": api_link}