feat: Update query processing guidelines for fund ID retrieval and clarify terminology

app/services/querying.py

@@ -38,10 +38,21 @@ class QueryProcessor:
         # Update system message to specifically request only fund IDs
         system_message_updated = (
             prompt_template.format(dialect="SQLite", top_k=5)
-            + "\n\nIMPORTANT: You must ONLY return the fund IDs (id field from the funds table) that match the user's criteria. "
-            + "Do NOT return any other information, explanations, or data. "
-            + "Your response should be ONLY a comma-separated list of numbers representing the fund IDs. "
-            + "Example format: 1, 5, 12, 23"
+            + "\n\n=== IMPORTANT TERMINOLOGY ==="
+            + "\n- When users say 'investors' or 'find me investors', they mean FUNDS"
+            + "\n- Always query the 'funds' table for investment opportunities"
+            + "\n- The 'investors' table is for parent company information only"
+            + "\n- Relationship: investors (1) -> (many) funds"
+            + "\n\n=== YOUR TASK ==="
+            + "\nReturn ONLY fund IDs (funds.id) that match the user's criteria."
+            + "\nFormat: comma-separated numbers only (e.g., 1, 5, 12, 23)"
+            + "\nNo explanations, no other data."
+            + "\n\n=== QUERY GUIDELINES ==="
+            + "\n1. For geographic searches: use funds.geographic_focus"
+            + "\n2. For sector searches: JOIN with fund_sectors table"
+            + "\n3. For stage searches: JOIN with fund_investment_stages table"
+            + "\n4. If no results: respond with 'NO_RESULTS'"
+            + "\n5. Never repeat the same failed query"
         )
         self.agent = create_react_agent(
             model=self.llm,