Spaces:

avsolatorio
/

test-data-mcp-server

Running

avsolatorio commited on Sep 17

Commit

ac28a84

1 Parent(s): 3454905

Update tool description

Signed-off-by: Aivin V. Solatorio <[email protected]>

Files changed (1) hide show

wdi_mcp_gradio.py CHANGED Viewed

@@ -7,12 +7,12 @@ import services
 def search_relevant_indicators(query: str, top_k: int):
     """Search for a shortlist of relevant indicators from the World Development Indicators (WDI) Data360 API given the query.
-    This tool is optimized for English language queries, so try to use English for your query. If the user's query is not in English, you may need to translate it to English first. This tool is used to find indicators and does not filter based on geography or time period.
     The search ranking may not be optimal, so the LLM may use this as shortlist and pick the most relevant from the list (if any). You, as an LLM, must always get at least `top_k=20` for better recall.
     Args:
-        query: The search query by the user or one formulated by an LLM based on the user's prompt.
         top_k: The number of shortlisted indicators that will be returned that are semantically related to the query. IMPORTANT: You, as an LLM, must ALWAYS set this argument to at least 20.
     Returns:

 def search_relevant_indicators(query: str, top_k: int):
     """Search for a shortlist of relevant indicators from the World Development Indicators (WDI) Data360 API given the query.
+    This tool is optimized for English language queries, so try to use English for your query. If the user's query is not in English, you may need to translate it to English first. This tool is used to find indicators and does not consider any geography or time period, so you should not include any in your query.
     The search ranking may not be optimal, so the LLM may use this as shortlist and pick the most relevant from the list (if any). You, as an LLM, must always get at least `top_k=20` for better recall.
     Args:
+        query: The search query by the user or one formulated by an LLM based on the user's prompt. This query should be in English. If the user's query is not in English, you may need to translate it to English first. This tool is used to find indicators and does not consider any geography or time period, so you should not include any in your query.
         top_k: The number of shortlisted indicators that will be returned that are semantically related to the query. IMPORTANT: You, as an LLM, must ALWAYS set this argument to at least 20.
     Returns: