Update inference.py

inference.py

@@ -9,7 +9,7 @@ import time
 import psutil
 import platform
 
-# 🔐 Secure API Key
+# 🔐 Load OpenAI API Key
 openai.api_key = os.getenv("OPENAI_API_KEY")
 
 # 📦 Constants
@@ -18,7 +18,7 @@ tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
 model = None
 last_mod_time = 0
 
-# 🔁 Reload Evo model if updated
+# 🔁 Load Evo model
 def load_model():
     global model, last_mod_time
     try:
@@ -34,7 +34,7 @@ def load_model():
         model = None
     return model
 
-# 🔮 Evo prediction core
+# 🔮 Evo inference
 def evo_infer(query, options, user_context=""):
     model = load_model()
     if model is None:
@@ -59,8 +59,8 @@ def evo_infer(query, options, user_context=""):
         context_str
     )
 
-# 💬 GPT-3.5 fallback
-def gpt_infer(query, user_context=""):
+# 💬 GPT fallback (renamed properly to fix import)
+def get_gpt_response(query, user_context=""):
     try:
         context_block = f"\n\nContext:\n{user_context}" if user_context else ""
         response = openai.chat.completions.create(
@@ -72,7 +72,7 @@ def gpt_infer(query, user_context=""):
     except Exception as e:
         return f"⚠️ GPT error:\n{str(e)}"
 
-# 🧠 Used by app.py
+# 🔄 UI integration
 def evo_chat_predict(history, query, options):
     context = "\n".join(history[-6:]) if history else ""
     evo_ans, evo_score, evo_reason, evo_ctx = evo_infer(query, options, context)
@@ -83,7 +83,7 @@ def evo_chat_predict(history, query, options):
         "context_used": evo_ctx
     }
 
-# 📊 Architecture stats
+# 📊 Evo config
 def get_model_config():
     return {
         "num_layers": 6,
@@ -94,7 +94,7 @@ def get_model_config():
         "accuracy": "~64.5%"
     }
 
-# 🖥️ System diagnostics
+# 💻 System info
 def get_system_stats():
     gpu_info = torch.cuda.get_device_properties(0) if torch.cuda.is_available() else None
     memory = psutil.virtual_memory()