vectara-agentic 0.2.23__tar.gz → 0.2.24__tar.gz

{vectara_agentic-0.2.23/vectara_agentic.egg-info → vectara_agentic-0.2.24}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: vectara_agentic
-Version: 0.2.23
+Version: 0.2.24
 Summary: A Python package for creating AI Assistants and AI Agents with Vectara
 Home-page: https://github.com/vectara/py-vectara-agentic
 Author: Ofer Mendelevitch

{vectara_agentic-0.2.23 → vectara_agentic-0.2.24}/vectara_agentic/_version.py

@@ -1,4 +1,4 @@
 """
 Define the version of the package.
 """
-__version__ = "0.2.23"
+__version__ = "0.2.24"

{vectara_agentic-0.2.23 → vectara_agentic-0.2.24}/vectara_agentic/db_tools.py

@@ -114,7 +114,7 @@ class DatabaseTools:
         if sql_query is None:
             raise ValueError("A query parameter is necessary to filter the data.")
 
-        count_query = f"SELECT COUNT(*) FROM ({sql_query})"
+        count_query = f"SELECT COUNT(*) FROM ({sql_query}) c"
         try:
             count_rows = self._load_data(count_query)
         except Exception as e:

{vectara_agentic-0.2.23 → vectara_agentic-0.2.24}/vectara_agentic/tools.py

@@ -236,10 +236,13 @@ class VectaraToolFactory:
                 )
             unique_ids = set()
             docs = []
+            doc_matches = {}
             for doc in response:
                 if doc.id_ in unique_ids:
+                    doc_matches[doc.id_].append(doc.node.get_content())
                     continue
                 unique_ids.add(doc.id_)
+                doc_matches[doc.id_] = [doc.node.get_content()]
                 docs.append((doc.id_, doc.metadata))
             tool_output = "Matching documents:\n"
             if summarize:
@@ -253,11 +256,24 @@ class VectaraToolFactory:
                 )
                 for doc_id, metadata in docs:
                     summary = summaries_dict.get(doc_id, "")
-                    tool_output += f"document_id: '{doc_id}'\nmetadata: '{metadata}'\nsummary: '{summary}'\n\n"
+                    matching_text = "\n".join(
+                        f"<match {i}>{piece}</match {i}>"
+                        for i, piece in enumerate(doc_matches[doc_id], start=1)
+                    )
+                    tool_output += (
+                        f"document_id: '{doc_id}'\nmetadata: '{metadata}'\n"
+                        f"matching texts: '{matching_text}'\n"
+                        f"summary: '{summary}'\n\n"
+                    )
             else:
                 for doc_id, metadata in docs:
+                    matching_text = "\n".join(
+                        f"<match {i}>{piece}</match {i}>"
+                        for i, piece in enumerate(doc_matches[doc_id], start=1)
+                    )
                     tool_output += (
-                        f"document_id: '{doc_id}'\nmetadata: '{metadata}'\n\n"
+                        f"document_id: '{doc_id}'\nmetadata: '{metadata}'\n"
+                        f"matching texts: '{matching_text}'\n"
                     )
 
             out = ToolOutput(

{vectara_agentic-0.2.23 → vectara_agentic-0.2.24/vectara_agentic.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: vectara_agentic
-Version: 0.2.23
+Version: 0.2.24
 Summary: A Python package for creating AI Assistants and AI Agents with Vectara
 Home-page: https://github.com/vectara/py-vectara-agentic
 Author: Ofer Mendelevitch