Update app.py

app.py

@@ -1,155 +1,104 @@
+import os
 import gradio as gr
-from huggingface_hub import InferenceClient
-from fastapi import FastAPI, Request
+from fastapi import FastAPI
 from fastapi.middleware.cors import CORSMiddleware
-import os
-
-# ==========================================================
-# 1️⃣ CONFIGURACIÓN DEL MODELO
-# ==========================================================
-
-MODEL_NAME = "meta-llama/Llama-3-8b-instruct"
-
-def generate_llm_response(message, system_message, max_tokens=512, temperature=0.7, top_p=0.95, hf_token=None):
-    """Genera una respuesta real usando el modelo LLM de Hugging Face."""
-    messages = [
-        {"role": "system", "content": system_message},
-        {"role": "user", "content": message}
-    ]
-    try:
-        client = InferenceClient(
-            model=MODEL_NAME,
-            token=hf_token or os.getenv("HF_TOKEN")
-        )
-        response = ""
-        for chunk in client.chat_completion(
-            messages=messages,
-            max_tokens=max_tokens,
-            temperature=temperature,
-            top_p=top_p,
-            stream=True,
-        ):
-            delta = chunk.choices[0].delta.get("content", "")
-            response += delta
-        return response.strip()
-    except Exception as e:
-        return f"⚠️ Error al conectar amb el model: {e}"
-
-
-# ==========================================================
-# 2️⃣ FUNCIÓN PARA GRADIO
-# ==========================================================
-
-def respond(
-    message,
-    history: list[dict[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-    hf_token: gr.OAuthToken,
-):
-    """Función de respuesta para el chat dentro de Gradio."""
-    try:
-        client = InferenceClient(
-            model=MODEL_NAME,
-            token=hf_token.token
-        )
-        messages = [{"role": "system", "content": system_message}]
-        messages.extend(history)
-        messages.append({"role": "user", "content": message})
-
-        for chunk in client.chat_completion(
-            messages=messages,
-            max_tokens=max_tokens,
-            temperature=temperature,
-            top_p=top_p,
-            stream=True,
-        ):
-            delta = chunk.choices[0].delta.get("content", "")
-            yield delta
-    except Exception as e:
-        yield f"⚠️ Error al conectar amb el model: {e}"
-
-
-# ==========================================================
-# 3️⃣ INTERFAZ DE GRADIO (tipo mensajes moderno)
-# ==========================================================
-
+from fastapi.responses import HTMLResponse
+import uvicorn
+
+# ===============================
+# 1️⃣ Función de Respuesta LLM (Mock)
+# ===============================
+def respond(message, history, system_message, max_tokens, temperature, top_p):
+    """
+    Simula una respuesta de un LLM.
+    """
+    messages = [{"role": "system", "content": system_message}]
+    messages.extend(history)
+    messages.append({"role": "user", "content": message})
+
+    if message.lower().strip() in ["hola", "hi"]:
+        response = "Hola Max 000! Soc un chatbot basat en LLM. Com et puc ajudar amb la teva salut avui?"
+    elif "informació" in message.lower():
+        response = "La informació que cerques es pot trobar a la secció d'informes o diagnòstics."
+    else:
+        response = f"He rebut el teu missatge: '{message}'. Prova preguntant sobre cites o historial clínic."
+    
+    return [{"role": "assistant", "content": response}]
+
+# ===============================
+# 2️⃣ Gradio Chat Interface
+# ===============================
 chatbot_llm = gr.ChatInterface(
-    fn=respond,
-    type="messages",  # ✅ formato nuevo sin warning
+    respond,
     textbox=gr.Textbox(placeholder="Escriu la teva pregunta al LLM...", container=False, scale=7),
     theme="soft",
-    title="Asistent LLM (Hugging Face Client)",
+    title="Asistente LLM (Hugging Face Client)",
+    type="messages",  # ✅ Corrección para evitar warning
     additional_inputs=[
-        gr.Textbox(value="Ets l'assistent sanitari de My Health. Respon en català, de manera concisa i útil.", label="Missatge del sistema"),
+        gr.Textbox(value="Ets l'assistent sanitari de My health. Respon en català, de manera concisa i útil.", label="Missatge del sistema"),
         gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Tokens màxims"),
         gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperatura"),
         gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (mostreig nucli)"),
     ],
 )
 
+# ===============================
+# 3️⃣ Función para cargar HTML Dashboard
+# ===============================
+HTML_FILE_PATH = "My_health.html"
 
-# ==========================================================
-# 4️⃣ FASTAPI BACKEND
-# ==========================================================
+def load_html_content():
+    try:
+        if not os.path.exists(HTML_FILE_PATH):
+            return f"<h1>Error: Archivo {HTML_FILE_PATH} no encontrado.</h1>"
+        with open(HTML_FILE_PATH, "r", encoding="utf-8") as f:
+            return f.read()
+    except Exception as e:
+        return f"<h1>Error al cargar el HTML:</h1><p>{e}</p>"
 
-app = FastAPI(title="My Health API")
+# ===============================
+# 4️⃣ FastAPI App para API y HTML
+# ===============================
+app = FastAPI()
 
-# Permitir peticiones CORS
+# Permitir requests desde cualquier frontend
 app.add_middleware(
     CORSMiddleware,
-    allow_origins=["*"],  # en producción, limitar a tu dominio
-    allow_credentials=True,
+    allow_origins=["*"],
     allow_methods=["*"],
     allow_headers=["*"],
 )
 
-@app.post("/chat")
-async def chat_endpoint(request: Request):
-    """Recibe JSON {message: "..."} y responde con el LLM."""
-    data = await request.json()
-    message = data.get("message", "")
-    system_message = data.get("system_message", "Ets l'assistent sanitari de My Health. Respon en català.")
-
-    if not message:
-        return {"response": "⚠️ Missatge buit"}
-
-    response = generate_llm_response(message, system_message)
-    return {"response": response}
-
-
-# ==========================================================
-# 5️⃣ DASHBOARD HTML
-# ==========================================================
-
-HTML_FILE_PATH = "My_health.html"
-
-def load_html_content():
-    if not os.path.exists(HTML_FILE_PATH):
-        return "<h1>Error: archivo HTML no encontrado.</h1>"
-    with open(HTML_FILE_PATH, "r", encoding="utf-8") as f:
-        return f.read()
-
-
-# ==========================================================
-# 6️⃣ GRADIO UI
-# ==========================================================
-
-with gr.Blocks() as demo:
-    gr.Markdown("# 🩺 My Health - Integració Gradio + API + HTML")
-
-    with gr.Sidebar():
-        gr.LoginButton()
+@app.get("/", response_class=HTMLResponse)
+async def home():
+    return load_html_content()
 
+@app.post("/chat")
+async def chat_endpoint(payload: dict):
+    message = payload.get("message", "")
+    system_message = payload.get("system_message", "Ets l'assistent sanitari de My health. Respon en català, de manera concisa i útil.")
+    response = respond(message, history=[], system_message=system_message, max_tokens=512, temperature=0.7, top_p=0.95)
+    return {"response": response[0]["content"]}
+
+# ===============================
+# 5️⃣ Gradio + FastAPI Integration
+# ===============================
+from fastapi.middleware.wsgi import WSGIMiddleware
+
+gradio_app = gr.Blocks()
+with gradio_app:
+    gr.Markdown("# Aplicació My Health - Integració Gradio/HTML")
     with gr.Tabs():
-        with gr.TabItem("🏠 Dashboard HTML"):
+        with gr.TabItem("Dashboard LMS (UI Estàtica amb Chat Flotant)"):
             gr.HTML(value=load_html_content())
-            gr.Markdown("Aquesta pestanya mostra el **dashboard** amb el chat flotant.")
-
-        with gr.TabItem("💬 Chat LLM"):
+        with gr.TabItem("Chat LLM (Accés Directe a Model)"):
             chatbot_llm.render()
 
-# ✅ Montamos Gradio en FastAPI sin reasignar app
-gr.mount_gradio_app(app, demo, path="/")
+app.mount("/gradio", WSGIMiddleware(gradio_app))
+
+# ===============================
+# 6️⃣ Launch Uvicorn (Red / Contenedor)
+# ===============================
+if __name__ == "__main__":
+    port = int(os.environ.get("PORT", 8000))
+    uvicorn.run(app, host="0.0.0.0", port=port)