Update app.py

app.py

@@ -1,6 +1,6 @@
 import os
 import torch
-import gradio as gr
+import argparse
 import openai
 from zipfile import ZipFile
 import requests
@@ -10,6 +10,10 @@ import langid
 import traceback
 from dotenv import load_dotenv
 
+from fastapi import FastAPI, File, UploadFile, Form
+from fastapi.responses import JSONResponse
+import uvicorn
+
 # Load environment variables
 load_dotenv()
 
@@ -90,7 +94,7 @@ def predict(audio_file_pth, agree):
     if audio_file_pth is not None:
         speaker_wav = audio_file_pth
     else:
-        text_hint += "[ERROR] Please record your voice using the Microphone.\n"
+        text_hint += "[ERROR] Please provide your voice as an audio file.\n"
         return (text_hint, None)
 
     # Transcribe audio to text using OpenAI Whisper
@@ -142,7 +146,6 @@ def predict(audio_file_pth, agree):
             stop=None,
             temperature=0.7,
         )
-        # Correctly access the response content
         reply_text = response.choices[0].message.content.strip()
         print(f"GPT-4 Reply: {reply_text}")
     except Exception as e:
@@ -177,48 +180,25 @@ def predict(audio_file_pth, agree):
 
     return (text_hint, synthesized_audio_path)
 
-with gr.Blocks(analytics_enabled=False) as demo:
-    gr.Markdown("# Mickey Mouse Voice Assistant")
-
-    with gr.Row():
-        with gr.Column():
-            audio_input = gr.Audio(
-                source="microphone",
-                type="filepath",
-                label="Record Your Voice",
-                info="Click the microphone button to record your voice."
-            )
-            tos_checkbox = gr.Checkbox(
-                label="Agree to Terms & Conditions",
-                value=False,
-                info="I agree to the terms of service."
-            )
-            submit_button = gr.Button("Send")
-
-        with gr.Column():
-            info_output = gr.Textbox(
-                label="Info",
-                interactive=False,
-                lines=4,
-            )
-            audio_output = gr.Audio(
-                label="Mickey's Response",
-                interactive=False,
-                autoplay=True,
-            )
+app = FastAPI()
+
+@app.post("/predict")
+async def predict_endpoint(agree: bool = Form(...), audio_file: UploadFile = File(...)):
+    # Save the uploaded file locally
+    temp_dir = "temp"
+    os.makedirs(temp_dir, exist_ok=True)
+    audio_path = os.path.join(temp_dir, audio_file.filename)
+    with open(audio_path, "wb") as f:
+        f.write(await audio_file.read())
+
+    info, audio_output_path = predict(audio_path, agree)
+    if audio_output_path:
+        # Return a JSON response with info and a path to the audio file.
+        # You could return the file content as base64 if you prefer.
+        return JSONResponse(content={"info": info, "audio_path": audio_output_path})
+    else:
+        return JSONResponse(content={"info": info, "audio_path": None}, status_code=400)
 
-    submit_button.click(
-        predict, 
-        inputs=[audio_input, tos_checkbox], 
-        outputs=[info_output, audio_output]
-    )
 
-# Launch the Gradio app
-demo.queue()
-demo.launch(
-    server_name="0.0.0.0",
-    server_port=int(os.environ.get("PORT", 7860)),
-    debug=True,
-    show_api=True,
-    share=False
-)
+if __name__ == "__main__":
+    uvicorn.run(app, host="0.0.0.0", port=int(os.environ.get("PORT", 7860)), debug=True)