Update inference.py

inference.py

@@ -70,3 +70,18 @@ def get_gpt_response(query, user_context=""):
 # ✅ Final callable interface
 def infer(query, options, user_context=""):
     return get_evo_response(query, options, user_context)
+
+# 🧠 Unified chat-style interface for EvoRAG
+def evo_chat_predict(history, query, options):
+    # Use the last few exchanges as context (up to 3 pairs)
+    context = "\n".join(history[-6:]) if history else ""
+
+    evo_ans, evo_score, evo_reason, evo_ctx = get_evo_response(query, options, context)
+
+    return {
+        "answer": evo_ans,
+        "confidence": round(evo_score, 3),
+        "reasoning": evo_reason,
+        "context_used": evo_ctx
+    }
+