Update app.py

app.py

@@ -184,11 +184,14 @@ def _build_image_parts_from_indices(indices: List[int]) -> List[Dict[str, Any]]:
 
 SYSTEM = (
     """
-You are a PDF research agent.
+You are a PDF research agent with a single tool: mcp_test_search(query: string, k: int).
+Act iteratively:
+  1) Split the user question into 1–4 focused sub-queries. Subqueries should be asked as natural language questions in the english language, not just keywords.
+  2) For each sub-query, call mcp_test_search (k=5 by default; increase to up to 10 if you need to go deep).
+  3) You will receive the output of mcp_test_search as a list of indices correspondinf to page numbers. Print them out and stop generating. You will be fed the corresponding pages as images in a follow-up message.
+  3) Stop early when confident; otherwise refine and repeat, up to 4 iterations and 20 searches in total. If info is missing, try to continue searching using new keywords and queries.
 
 Workflow:
-  • When you need pages, call the tool: mcp_test_search(query: string, k: int).
-  • The app will attach the images for the LAST search result you produced in this turn in a follow-up message.
   • Use ONLY the provided images for grounding and cite as (p.<page>).
   • If an answer is not present, say “Not found in the provided pages.”
 
@@ -236,7 +239,7 @@ def stream_agent(question: str,
 
     # Optional seeding: attach some likely pages on round 1
     try:
-        seed_indices = search(question, k=5) or []
+        seed_indices = [] # search(question, k=5) or []
     except Exception as e:
         yield f"❌ Search failed: {e}", "", ""
         return
@@ -337,7 +340,7 @@ def stream_agent(question: str,
 
                     # Capture tool RESULT text and try to parse indices
                     elif etype.startswith("response.tool_result"):
-                        print("here")
+                        print("A tool output was detected")
                         delta_text = getattr(event, "delta", "") or getattr(event, "output_text", "")
                         if delta_text:
                             tool_result_buffer += str(delta_text)
@@ -368,7 +371,7 @@ def stream_agent(question: str,
     pending_indices = list(seed_indices)
 
     while round_idx <= max_rounds:
-        print(round_idx, pending_indices)
+        print("Round ", round_idx, ", Indices: ", pending_indices)
         for final_md, summary_md, log_md in run_round(round_idx, pending_indices):
             yield final_md, summary_md, log_md