Update app.py

app.py

@@ -17,6 +17,7 @@ model = AutoModelForCausalLM.from_pretrained(
     trust_remote_code=True,
 )
 
+# v0
 @spaces.GPU(duration=300)
 def generate_response(query):
     # 🚀 generate_response 함수 안에서 매번 로드
@@ -58,9 +59,23 @@ def generate_response(query):
 
     return tokenizer.decode(outputs[0], skip_special_tokens=True)
 
+# v1
 @spaces.GPU(duration=300)
 def generate_response_with_pipeline(query):
-    return rag_pipeline(query)    
+    return rag_pipeline(query)
+    
+# v2
+search_cache = {}
+@spaces.GPU(duration=300)
+def search_documents_with_cache(query: str):
+    if query in search_cache:
+        print(f"⚡ 캐시 사용: '{query}'")
+        return search_cache[query]
+
+    results = rag_pipeline(query)
+    search_cache[query] = results
+    return results
+
 
-demo = gr.Interface(fn=generate_response_with_pipeline, inputs="text", outputs="text")
+demo = gr.Interface(fn=search_documents_with_cache, inputs="text", outputs="text")
 demo.launch()