Merge branch 'master' into hard-coded-version

Author: sydney-runkle

docs/docs/concepts/retrieval.mdx

@@ -92,7 +92,7 @@ structured_model = model.with_structured_output(Questions)
 
 # Define the system prompt
 system = """You are a helpful assistant that generates multiple sub-questions related to an input question. \n
-The goal is to break down the input into a set of sub-problems / sub-questions that can be answers in isolation. \n"""
+The goal is to break down the input into a set of sub-problems / sub-questions that can be answered independently. \n"""
 
 # Pass the question to the model
 question = """What are the main components of an LLM-powered autonomous agent system?"""

docs/docs/versions/migrating_chains/map_rerank_docs_chain.ipynb

@@ -13,7 +13,7 @@
     "- Map a process to the set of documents, where the process includes generating a score;\n",
     "- Rank the results by score and return the maximum.\n",
     "\n",
-    "A common process in this scenario is question-answering using pieces of context from a document. Forcing the model to generate score along with its answer helps to select for answers generated only by relevant context.\n",
+    "A common process in this scenario is question-answering using pieces of context from a document. Forcing the model to generate a score along with its answer helps to select for answers generated only by relevant context.\n",
     "\n",
     "An [LangGraph](https://langchain-ai.github.io/langgraph/) implementation allows for the incorporation of [tool calling](/docs/concepts/tool_calling) and other features for this problem. Below we will go through both `MapRerankDocumentsChain` and a corresponding LangGraph implementation on a simple example for illustrative purposes."
    ]

libs/community/langchain_community/retrievers/google_vertex_ai_search.py

@@ -167,6 +167,8 @@ def _convert_website_search_response(
             doc_metadata = document_dict.get("struct_data", {})
             doc_metadata["id"] = document_dict["id"]
             doc_metadata["source"] = derived_struct_data.get("link", "")
+            if derived_struct_data.get("title") is not None:
+                doc_metadata["title"] = derived_struct_data.get("title")
 
             if chunk_type not in derived_struct_data:
                 continue