Update app.py

app.py

@@ -4,43 +4,41 @@ import os
 import re
 import time
 
-# Mistral is reliable and follows interview instructions well
-API_URL = "https://api-inference.huggingface.co/models/mistralai/Mistral-7B-Instruct-v0.3"
+# NEW ROUTER URL
+# Format: https://router.huggingface.co/hf-inference/models/{model_id}
+MODEL_ID = "mistralai/Mistral-7B-Instruct-v0.3"
+API_URL = f"https://router.huggingface.co/hf-inference/models/{MODEL_ID}"
 HF_TOKEN = os.getenv("HF_TOKEN")
 
-headers = {"Authorization": f"Bearer {HF_TOKEN}"}
+headers = {"Authorization": f"Bearer {HF_TOKEN}", "Content-Type": "application/json"}
 
 def query_llm(prompt):
-    """Helper to call API with automatic retry and wait-for-model logic."""
     payload = {
         "inputs": f"<s>[INST] {prompt} [/INST]",
         "parameters": {"max_new_tokens": 250, "temperature": 0.7},
-        "options": {"wait_for_model": True}  # Tells HF to wait for the model to load
+        "options": {"wait_for_model": True}
     }
     
-    for attempt in range(3):  # Try up to 3 times
-        try:
-            response = requests.post(API_URL, headers=headers, json=payload, timeout=60)
-            result = response.json()
-            
-            # If model is still loading despite wait_for_model
-            if isinstance(result, dict) and "error" in result:
-                if "loading" in result["error"].lower():
-                    time.sleep(10) # Wait 10 seconds before retrying
-                    continue
-                return f"AI Error: {result['error']}"
-            
-            if isinstance(result, list) and len(result) > 0:
-                return result[0]["generated_text"].split("[/INST]")[-1].strip()
+    try:
+        # The new router endpoint
+        response = requests.post(API_URL, headers=headers, json=payload, timeout=60)
+        
+        # Check for error messages from the router
+        if response.status_code != 200:
+            error_detail = response.json().get("error", "Unknown Router Error")
+            return f"AI Error ({response.status_code}): {error_detail}"
             
-        except Exception as e:
-            if attempt == 2: return f"Connection Error: {str(e)}"
-            time.sleep(5)
+        result = response.json()
+        if isinstance(result, list) and len(result) > 0:
+            return result[0]["generated_text"].split("[/INST]")[-1].strip()
             
-    return "The model is taking too long to wake up. Please wait 1 minute and refresh."
+        return "Unexpected response format from AI."
+        
+    except Exception as e:
+        return f"Connection Error: {str(e)}"
 
 # ----------------------------
-# AI Functions
+# AI Functions (Logic remains the same)
 # ----------------------------
 
 def generate_question(role, difficulty, resume_text):
@@ -67,7 +65,7 @@ def evaluate_answer(question, answer):
 # ----------------------------
 
 with gr.Blocks(theme=gr.themes.Soft()) as demo:
-    gr.Markdown("# 🤖 AI Interview Simulator")
+    gr.Markdown("# 🤖 AI Interview Simulator (Updated Router)")
     
     with gr.Row():
         role = gr.Textbox(label="Job Role", placeholder="e.g. Data Analyst")
@@ -88,4 +86,5 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
     gen_btn.click(generate_question, [role, diff, resume], q_display)
     eval_btn.click(evaluate_answer, [q_display, ans_input], [fb_display, score_display])
 
-demo.launch()
+if __name__ == "__main__":
+    demo.launch()