Update model.py

model.py

@@ -979,7 +979,7 @@ async def query_document_info(niche_cases, query_word, alternative_query_word, s
     2. RAG with semantic search and LLM (general, flexible, cost-optimized).
     """
     print("inside the model.query_doc_info")
-    outputs, links = {}, []
+    outputs, links, accession_found_in_text = {}, [], False
     if model_ai:
       if model_ai == "gemini-1.5-flash-latest":
         genai.configure(api_key=os.getenv("GOOGLE_API_KEY"))
@@ -1245,6 +1245,10 @@ async def query_document_info(niche_cases, query_word, alternative_query_word, s
     f"\nText Snippets:\n{context_for_llm}")
         
         print("this is prompt: ", prompt_for_llm)
+        # check if accession in text or not
+        if alternative_query_word_cleaned.lower() in prompt_for_llm.lower():
+          accession_found_in_text = True
+        
         if model_ai:
           print("back up to ", model_ai)
           #llm_response_text, model_instance = call_llm_api(prompt_for_llm, model=model_ai)
@@ -1378,7 +1382,8 @@ async def query_document_info(niche_cases, query_word, alternative_query_word, s
           f"2. The **second line onward** contains the explanations based on the non-unknown {again_output_format} answer.\n"
           f"\nText Snippets:\n{context_for_llm}")
             print("len of prompt:", len(general_knowledge_prompt))
-
+            if alternative_query_word_cleaned.lower() in general_knowledge_prompt.lower():
+              accession_found_in_text = True
             if general_knowledge_prompt:    
               if model_ai:
                 print("back up to ", model_ai)
@@ -1423,5 +1428,5 @@ async def query_document_info(niche_cases, query_word, alternative_query_word, s
 
     print("all done and method used: ", outputs, method_used)  
     print("total cost: ", total_query_cost)   
-    return outputs, method_used, total_query_cost, links
+    return outputs, method_used, total_query_cost, links, accession_found_in_text