Use native demo.launch() with FastAPI routes for HuggingFace Spaces lifecycle

app.py

@@ -1,6 +1,6 @@
 """
 Main entry point for Hugging Face Spaces (Gradio SDK).
-Mounts the production FastAPI REST API and provides an interactive Gradio Web UI simultaneously.
+Provides an interactive Gradio Web UI and exposes the FastAPI REST API simultaneously.
 """
 
 import os
@@ -71,7 +71,7 @@ with gr.Blocks(title="Bilingual Arabic-English Text Summarization AI", theme=gr.
     gr.Markdown("""
     **Production NLP Platform** supporting Extractive (TextRank, LSA, Hybrid) and Abstractive (Seq2Seq + Bahdanau Attention) summarization.
     
-    🔗 **REST API Documentation**: Access Swagger UI at `[URL]/docs` | ReDoc at `[URL]/redoc`
+    🔗 **API Endpoints**: `/api/v1/summarize` | `/api/v1/health` | `/api/v1/detect-language` | `/api/v1/evaluate`
     """)
     
     with gr.Row():
@@ -110,9 +110,11 @@ with gr.Blocks(title="Bilingual Arabic-English Text Summarization AI", theme=gr.
         outputs=[output_summary, out_lang, out_reduc, out_lat]
     )
 
-# Mount Gradio onto FastAPI so both work simultaneously!
-app = gr.mount_gradio_app(fastapi_app, demo, path="/")
+# Attach FastAPI endpoints directly into Gradio's internal FastAPI app
+for route in fastapi_app.routes:
+    if hasattr(route, "path") and route.path.startswith("/api/"):
+        demo.app.routes.append(route)
 
-if __name__ == "__main__":
-    import uvicorn
-    uvicorn.run(app, host="0.0.0.0", port=7860)
+# Launch Gradio app directly (required for Hugging Face Spaces lifecycle)
+demo.queue()
+demo.launch(server_name="0.0.0.0", server_port=7860)