Update app.py

app.py

@@ -2,13 +2,20 @@ import gradio as gr
 import os
 from huggingface_hub import InferenceClient
 
-# Use Model
-model_id = "deepseek-ai/deepseek-coder-33b-instruct"
+# --- UPDATED: Best Coding Models on Free Tier (2025/2026) ---
+AVAILABLE_MODELS = [
+    "Qwen/Qwen2.5-Coder-32B-Instruct",    # SOTA Coding Model (Best overall)
+    "Qwen/Qwen2.5-72B-Instruct",          # Larger General Model (Great at Logic)
+    "meta-llama/Llama-3.1-8B-Instruct",   # Fast & Reliable
+    "mistralai/Mistral-Nemo-Instruct-2407", # 12B Model (Very Smart)
+    "microsoft/Phi-3.5-mini-instruct",    # Extremely Fast
+    "bigcode/starcoder2-15b-instruct-v0.1", # Specialized Code Model
+    "HuggingFaceH4/zephyr-7b-beta",       # Backup Model
+]
 
-def respond(message, history, system_message, temperature, request: gr.Request):
+def respond(message, history, system_message, temperature, model_id, request: gr.Request):
     # --- 1. Authentication Logic ---
     token = None
-    # Safely access token (handles both Local run and Spaces)
     if request:
         token = getattr(request, "token", None)
     
@@ -20,12 +27,12 @@ def respond(message, history, system_message, temperature, request: gr.Request):
         return
 
     # --- 2. Setup Client ---
+    # The client is created dynamically based on the selected model_id
     client = InferenceClient(model_id, token=token)
     
-    # --- 3. Build Messages (handling history correctly) ---
+    # --- 3. Build Messages ---
     messages = [{"role": "system", "content": system_message}]
     
-    # We use type="messages" in ChatInterface, so history is already a list of dicts
     for msg in history:
         messages.append(msg)
     
@@ -53,22 +60,43 @@ def respond(message, history, system_message, temperature, request: gr.Request):
                 yield response_text
 
     except Exception as e:
-        yield f"Error: {str(e)}"
+        # Better Error Handling for Model Switching
+        error_msg = str(e)
+        if "404" in error_msg or "model_not_supported" in error_msg:
+            yield f"Error: The model **{model_id}** is currently offline or overloaded. \n\n👉 Please select a different model from the dropdown above."
+        else:
+            yield f"Error: {error_msg}"
 
 # --- 5. Build UI ---
 with gr.Blocks(fill_height=True) as demo:
     with gr.Sidebar():
-        gr.Markdown("# AI Coding Assistant")
-        gr.Markdown(f"Running **{model_id}**")
+        gr.Markdown("# 🤖 Multi-Model Coding Assistant")
+        gr.Markdown("Select a model from the **Additional Inputs** menu below to switch between different AI coding experts.")
         gr.LoginButton("Sign in")
     
     gr.ChatInterface(
         respond,
-        # 'type="messages"' fixes the deprecation warning and makes parsing easier
         type="messages", 
         additional_inputs=[
-            gr.Textbox(value="You are a helpful assistant.", label="System Instruction", lines=2),
-            gr.Slider(minimum=0.1, maximum=2.0, value=0.7, step=0.1, label="Temperature")
+            gr.Textbox(
+                value="You are an expert software engineer. Provide clean, efficient, and well-commented code.", 
+                label="System Instruction", 
+                lines=2
+            ),
+            gr.Slider(
+                minimum=0.1, 
+                maximum=2.0, 
+                value=0.5, 
+                step=0.1, 
+                label="Temperature (Lower = More Precise)"
+            ),
+            # --- Model Selection Dropdown ---
+            gr.Dropdown(
+                choices=AVAILABLE_MODELS,
+                value=AVAILABLE_MODELS[0], # Default to Qwen 2.5 Coder
+                label="Select AI Model",
+                interactive=True
+            )
         ]
     )