Upload app.py with huggingface_hub

app.py

@@ -1,4 +1,10 @@
-print("--- [v137-clean] 🚀 BOOTING APP.PY ---")
+import os
+import sys
+
+# --- [v137-clean-v3] 🚀 INITIALIZING SYSTEM ---
+print(f"PYTHON VERSION: {sys.version}")
+print(f"WORKING DIR: {os.getcwd()}")
+
 try:
     import spaces
 except ImportError:
@@ -11,16 +17,13 @@ except ImportError:
 import gradio as gr
 from fastapi import FastAPI, Request
 from fastapi.middleware.cors import CORSMiddleware
-import uvicorn
 import base64
 import torch
-import os
 import tempfile
 import json
 import time
 import gc
 import traceback
-import soundfile as sf
 import numpy as np
 from transformers import pipeline, AutoModelForSpeechSeq2Seq, AutoProcessor
 from TTS.api import TTS
@@ -35,9 +38,6 @@ os.environ["CUBLAS_WORKSPACE_CONFIG"] = ":4096:8"
 torch.backends.cuda.matmul.allow_tf32 = False
 torch.backends.cudnn.allow_tf32 = False
 
-app = FastAPI()
-app.add_middleware(CORSMiddleware, allow_origins=["*"], allow_methods=["*"], allow_headers=["*"])
-
 MODELS = {"stt": None, "tts": None}
 
 def load_gpu_models():
@@ -48,7 +48,6 @@ def load_gpu_models():
     if MODELS.get("stt") is None:
         print("--- [v137] 📥 LOADING NATIVE WHISPER (Large-v3-Turbo) ---")
         model_id = "openai/whisper-large-v3-turbo"
-        # Load model with SDPA (Flash Attention) for H200
         model = AutoModelForSpeechSeq2Seq.from_pretrained(
             model_id, torch_dtype=torch.float16, low_cpu_mem_usage=True, use_safetensors=True
         ).to(device)
@@ -147,7 +146,33 @@ def core_process(request_dict):
         print(f"--- [v137] ✨ DONE ({time.time()-t1:.1f}s) ---")
         torch.cuda.empty_cache() 
 
-@app.post("/process")
+# --- Gradio UI Logic ---
+def gradio_stt_fn(audio_path) -> str:
+    if not audio_path: return "No audio provided."
+    try:
+        with open(audio_path, "rb") as f:
+            b64 = base64.b64encode(f.read()).decode()
+        res = core_process({"action": "stt", "file": b64})
+        return res.get("text", f"Error: {res.get('error')}")
+    except Exception as e:
+        return f"UI Error: {str(e)}"
+
+# --- Interface Definition ---
+with gr.Blocks(title="S2ST H200 v137") as demo:
+    gr.Markdown("# 🚀 S2ST AI Engine v137 (HOPPER NATIVE)")
+    gr.Markdown("**H200 Stable | Transformers Whisper | XTTS-v2 VRAM Singleton**")
+    with gr.Row():
+        audio_in = gr.Audio(type="filepath", label="Input Audio")
+        stt_btn = gr.Button("Transcribe (STT)")
+        txt_out = gr.Textbox(label="Result")
+    stt_btn.click(fn=gradio_stt_fn, inputs=audio_in, outputs=txt_out)
+
+# --- FastAPI Route Integration ---
+print("--- [v137-clean-v3] 🔧 INTEGRATING FASTAPI ROUTES ---")
+fastapi_app = demo.app # Access Gradio's internal FastAPI app
+fastapi_app.add_middleware(CORSMiddleware, allow_origins=["*"], allow_methods=["*"], allow_headers=["*"])
+
+@fastapi_app.post("/process")
 async def api_process(request: Request):
     try:
         data = await request.json()
@@ -155,29 +180,11 @@ async def api_process(request: Request):
         return core_process(data)
     except Exception as e: return {"error": str(e)}
 
-@app.get("/health")
-def health(): return {"status": "ok", "v": "137"}
-
-# Named function for Gradio to avoid lambda schema issues
-def gradio_stt(audio_path):
-    if not audio_path: return ""
-    with open(audio_path, "rb") as f:
-        b64 = base64.b64encode(f.read()).decode()
-    res = core_process({"action": "stt", "file": b64})
-    return res.get("text", f"Error: {res.get('error')}")
-
-with gr.Blocks() as demo:
-    gr.Markdown("# 🚀 v137 HOPPER NATIVE (H200 Stable)")
-    gr.Markdown("Direct GPU path | Transformers Whisper | XTTS-v2 Singleton")
-    with gr.Row():
-        audio_in = gr.Audio(type="filepath", label="Input Audio")
-        stt_btn = gr.Button("STT")
-        txt_out = gr.Textbox(label="STT Result")
-    stt_btn.click(fn=gradio_stt, inputs=audio_in, outputs=txt_out)
-
-print("--- [v137-clean] 🔧 MOUNTING GRADIO ---")
-app = gr.mount_gradio_app(app, demo, path="/")
+@fastapi_app.get("/api/v137_health")
+def api_v137_health():
+    return {"status": "ok", "v": "137", "details": "NATIVE_INTEGRATION"}
 
+# --- Start System ---
 if __name__ == "__main__":
-    print("--- [v137-clean] 📡 STARTING UVICORN ---")
-    uvicorn.run(app, host="0.0.0.0", port=7860, log_level="info")
+    print("--- [v137-clean-v3] 📡 LAUNCHING SYSTEM ---")
+    demo.launch(server_name="0.0.0.0", server_port=7860, show_api=False, quiet=True)