Update inference.py

inference.py

@@ -3,6 +3,7 @@ from transformers import AutoTokenizer
 from evo_model import EvoTransformerV22 
 from search_utils import web_search
 import openai
+import os
 
 # Load Evo model and tokenizer
 model = EvoTransformerV22()
@@ -12,7 +13,7 @@ model.eval()
 tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
 
 # GPT Setup
-openai.api_key = "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"  # 🔑 Set your actual key securely
+openai.api_key = os.getenv("OPENAI_API_KEY")  # 🔒 Load securely from environment
 
 def get_evo_response(query, options, user_context=""):
     context_texts = web_search(query) + ([user_context] if user_context else [])
@@ -38,10 +39,13 @@ def get_evo_response(query, options, user_context=""):
 def get_gpt_response(query, user_context=""):
     try:
         context_block = f"\n\nContext:\n{user_context}" if user_context else ""
-        completion = openai.ChatCompletion.create(
+        response = openai.chat.completions.create(
             model="gpt-3.5-turbo",
-            messages=[{"role": "user", "content": query + context_block}]
+            messages=[
+                {"role": "user", "content": query + context_block}
+            ],
+            temperature=0.7,
         )
-        return completion.choices[0].message.content.strip()
+        return response.choices[0].message.content.strip()
     except Exception as e:
-        return f"⚠️ GPT error: {str(e)}"
+        return f"⚠️ GPT error:\n\n{str(e)}"