Merge branch 'main' into fix/retry-after-header

Author: debu-sinha

.github/workflows/release.yml

@@ -41,7 +41,7 @@ jobs:
           uv run python -c "from llama_index.core.utils import get_tokenizer; print(get_tokenizer('gpt-5-mini'))"
 
       - name: Generate build provenance attestations
-        uses: actions/attest-build-provenance@v3
+        uses: actions/attest-build-provenance@v4
         id: attest
         with:
           subject-path: "llama-index-core/llama_index/core/_static/**/*"

docs/examples/tools/moss_agent.ipynb …xamples/tools/moss_with_llamaindex.ipynbdocs/examples/tools/moss_agent.ipynb renamed to docs/examples/tools/moss_with_llamaindex.ipynb docs/examples/tools/moss_agent.ipynb renamed to docs/examples/tools/moss_with_llamaindex.ipynb

@@ -5,13 +5,9 @@
    "id": "3e37307f",
    "metadata": {},
    "source": [
-    "# Moss Tool Example\n",
+    "# Moss + LlamaIndex\n",
     "\n",
-    "This notebook demonstrates how to use the `MossToolSpec` with LlamaIndex.\n",
-    "\n",
-    "## Installation\n",
-    "\n",
-    "First, you need to install the package (and LlamaIndex core if not present)."
+    "[Moss](https://moss.dev) is a search runtime for voice agents, copilots, and multimodal apps. Sub-10ms lookups, zero infrastructure. Built in Rust and WebAssembly. Moss runs search inside your agent runtime. Sub-10ms lookups, always-current data, without rebuilding your infra."
    ]
   },
   {
@@ -31,7 +27,7 @@
    "source": [
     "## Setup\n",
     "\n",
-    "Import the necessary classes and initialize the Moss client."
+    "Import the necessary classes and initialize the Moss client with your project credentials. We'll configure query options to tune the search behavior (`top_k` for result count, `alpha` for the semantic/keyword blend)."
    ]
   },
   {
@@ -75,7 +71,7 @@
    "source": [
     "## Indexing Data\n",
     "\n",
-    "Let's add some sample data to our index."
+    "Create sample documents with `DocumentInfo` (each with `id`, `text`, and `metadata`). Then call `index_docs()` to build the Moss index — this creates or replaces the index with your documents and makes them searchable."
    ]
   },
   {
@@ -110,7 +106,7 @@
    "source": [
     "## Using with an Agent\n",
     "\n",
-    "Now we can wrap the tool for use with a LlamaIndex agent."
+    "Now we'll expose the Moss tool's methods (`query`, `list_indexes`, `delete_index`) to a `ReActAgent`. The agent will autonomously decide which tool to call based on the user's question and the search results it retrieves."
    ]
   },
   {

llama-index-core/llama_index/core/ingestion/pipeline.py

@@ -459,21 +459,24 @@ def _node_batcher(
             yield nodes[i : i + batch_size]
 
     def _update_docstore(
-        self, nodes: Sequence[BaseNode], store_doc_text: bool = True
+        self,
+        nodes: Sequence[BaseNode],
+        effective_strategy: DocstoreStrategy,
+        store_doc_text: bool = True,
     ) -> None:
         """Update the document store with the given nodes."""
         assert self.docstore is not None
 
-        if self.docstore_strategy in (
+        if effective_strategy in (
             DocstoreStrategy.UPSERTS,
             DocstoreStrategy.UPSERTS_AND_DELETE,
         ):
             self.docstore.set_document_hashes({n.id_: n.hash for n in nodes})
             self.docstore.add_documents(nodes, store_text=store_doc_text)
-        elif self.docstore_strategy == DocstoreStrategy.DUPLICATES_ONLY:
+        elif effective_strategy == DocstoreStrategy.DUPLICATES_ONLY:
             self.docstore.add_documents(nodes, store_text=store_doc_text)
         else:
-            raise ValueError(f"Invalid docstore strategy: {self.docstore_strategy}")
+            raise ValueError(f"Invalid docstore strategy: {effective_strategy}")
 
     @dispatcher.span
     def run(
@@ -511,30 +514,34 @@ def run(
         """
         input_nodes = self._prepare_inputs(documents, nodes)
 
+        effective_strategy = self.docstore_strategy
+        if (
+            self.docstore is not None
+            and self.vector_store is None
+            and self.docstore_strategy
+            in (DocstoreStrategy.UPSERTS, DocstoreStrategy.UPSERTS_AND_DELETE)
+        ):
+            warnings.warn(
+                f"docstore_strategy='{self.docstore_strategy.value}' requires a vector store "
+                "to apply upsert/delete semantics; falling back to 'duplicates_only' for this run. "
+                "pipeline.docstore_strategy is unchanged.",
+                UserWarning,
+                stacklevel=3,
+            )
+            effective_strategy = DocstoreStrategy.DUPLICATES_ONLY
+
         # check if we need to dedup
         if self.docstore is not None and self.vector_store is not None:
-            if self.docstore_strategy in (
+            if effective_strategy in (
                 DocstoreStrategy.UPSERTS,
                 DocstoreStrategy.UPSERTS_AND_DELETE,
             ):
                 nodes_to_run = self._handle_upserts(input_nodes)
-            elif self.docstore_strategy == DocstoreStrategy.DUPLICATES_ONLY:
+            elif effective_strategy == DocstoreStrategy.DUPLICATES_ONLY:
                 nodes_to_run = self._handle_duplicates(input_nodes)
             else:
-                raise ValueError(f"Invalid docstore strategy: {self.docstore_strategy}")
+                raise ValueError(f"Invalid docstore strategy: {effective_strategy}")
         elif self.docstore is not None and self.vector_store is None:
-            if self.docstore_strategy == DocstoreStrategy.UPSERTS:
-                logger.info(
-                    "Docstore strategy set to upserts, but no vector store. "
-                    "Switching to duplicates_only strategy."
-                )
-                self.docstore_strategy = DocstoreStrategy.DUPLICATES_ONLY
-            elif self.docstore_strategy == DocstoreStrategy.UPSERTS_AND_DELETE:
-                logger.info(
-                    "Docstore strategy set to upserts and delete, but no vector store. "
-                    "Switching to duplicates_only strategy."
-                )
-                self.docstore_strategy = DocstoreStrategy.DUPLICATES_ONLY
             nodes_to_run = self._handle_duplicates(input_nodes)
         else:
             nodes_to_run = input_nodes
@@ -582,27 +589,34 @@ def run(
                 self.vector_store.add(nodes_with_embeddings)
 
         if self.docstore is not None:
-            self._update_docstore(nodes_to_run, store_doc_text=store_doc_text)
+            self._update_docstore(
+                nodes_to_run,
+                effective_strategy=effective_strategy,
+                store_doc_text=store_doc_text,
+            )
 
         return nodes
 
     # ------ async methods ------
     async def _aupdate_docstore(
-        self, nodes: Sequence[BaseNode], store_doc_text: bool = True
+        self,
+        nodes: Sequence[BaseNode],
+        effective_strategy: DocstoreStrategy,
+        store_doc_text: bool = True,
     ) -> None:
         """Update the document store with the given nodes."""
         assert self.docstore is not None
 
-        if self.docstore_strategy in (
+        if effective_strategy in (
             DocstoreStrategy.UPSERTS,
             DocstoreStrategy.UPSERTS_AND_DELETE,
         ):
             await self.docstore.aset_document_hashes({n.id_: n.hash for n in nodes})
             await self.docstore.async_add_documents(nodes, store_text=store_doc_text)
-        elif self.docstore_strategy == DocstoreStrategy.DUPLICATES_ONLY:
+        elif effective_strategy == DocstoreStrategy.DUPLICATES_ONLY:
             await self.docstore.async_add_documents(nodes, store_text=store_doc_text)
         else:
-            raise ValueError(f"Invalid docstore strategy: {self.docstore_strategy}")
+            raise ValueError(f"Invalid docstore strategy: {effective_strategy}")
 
     async def _ahandle_duplicates(
         self,
@@ -700,38 +714,41 @@ async def arun(
         """
         input_nodes = self._prepare_inputs(documents, nodes)
 
+        effective_strategy = self.docstore_strategy
+        if (
+            self.docstore is not None
+            and self.vector_store is None
+            and self.docstore_strategy
+            in (DocstoreStrategy.UPSERTS, DocstoreStrategy.UPSERTS_AND_DELETE)
+        ):
+            warnings.warn(
+                f"docstore_strategy='{self.docstore_strategy.value}' requires a vector store "
+                "to apply upsert/delete semantics; falling back to 'duplicates_only' for this run. "
+                "pipeline.docstore_strategy is unchanged.",
+                UserWarning,
+                stacklevel=3,
+            )
+            effective_strategy = DocstoreStrategy.DUPLICATES_ONLY
+
         # check if we need to dedup
         if self.docstore is not None and self.vector_store is not None:
-            if self.docstore_strategy in (
+            if effective_strategy in (
                 DocstoreStrategy.UPSERTS,
                 DocstoreStrategy.UPSERTS_AND_DELETE,
             ):
                 nodes_to_run = await self._ahandle_upserts(
                     input_nodes, store_doc_text=store_doc_text
                 )
-            elif self.docstore_strategy == DocstoreStrategy.DUPLICATES_ONLY:
+            elif effective_strategy == DocstoreStrategy.DUPLICATES_ONLY:
                 nodes_to_run = await self._ahandle_duplicates(
                     input_nodes, store_doc_text=store_doc_text
                 )
             else:
-                raise ValueError(f"Invalid docstore strategy: {self.docstore_strategy}")
+                raise ValueError(f"Invalid docstore strategy: {effective_strategy}")
         elif self.docstore is not None and self.vector_store is None:
-            if self.docstore_strategy == DocstoreStrategy.UPSERTS:
-                logger.info(
-                    "Docstore strategy set to upserts, but no vector store. "
-                    "Switching to duplicates_only strategy."
-                )
-                self.docstore_strategy = DocstoreStrategy.DUPLICATES_ONLY
-            elif self.docstore_strategy == DocstoreStrategy.UPSERTS_AND_DELETE:
-                logger.info(
-                    "Docstore strategy set to upserts and delete, but no vector store. "
-                    "Switching to duplicates_only strategy."
-                )
-                self.docstore_strategy = DocstoreStrategy.DUPLICATES_ONLY
             nodes_to_run = await self._ahandle_duplicates(
                 input_nodes, store_doc_text=store_doc_text
             )
-
         else:
             nodes_to_run = input_nodes
 
@@ -785,6 +802,10 @@ async def arun(
                 await self.vector_store.async_add(nodes_with_embeddings)
 
         if self.docstore is not None:
-            await self._aupdate_docstore(nodes_to_run, store_doc_text=store_doc_text)
+            await self._aupdate_docstore(
+                nodes_to_run,
+                effective_strategy=effective_strategy,
+                store_doc_text=store_doc_text,
+            )
 
         return nodes

llama-index-core/tests/ingestion/test_pipeline.py

@@ -5,7 +5,7 @@
 import pytest
 from llama_index.core.embeddings.mock_embed_model import MockEmbedding
 from llama_index.core.extractors import KeywordExtractor
-from llama_index.core.ingestion.pipeline import IngestionPipeline
+from llama_index.core.ingestion.pipeline import IngestionPipeline, DocstoreStrategy
 from llama_index.core.llms.mock import MockLLM
 from llama_index.core.node_parser import SentenceSplitter, MarkdownElementNodeParser
 from llama_index.core.readers import ReaderConfig, StringIterableReader
@@ -86,6 +86,7 @@ def test_save_load_pipeline(tmp_path: Path, monkeypatch: pytest.MonkeyPatch) ->
             SentenceSplitter(chunk_size=25, chunk_overlap=0),
         ],
         docstore=SimpleDocumentStore(),
+        docstore_strategy=DocstoreStrategy.DUPLICATES_ONLY,
     )
 
     nodes = pipeline.run(documents=documents)
@@ -169,6 +170,7 @@ def test_pipeline_update_text_content() -> None:
             SentenceSplitter(chunk_size=25, chunk_overlap=0),
         ],
         docstore=SimpleDocumentStore(),
+        docstore_strategy=DocstoreStrategy.DUPLICATES_ONLY,
     )
 
     nodes = pipeline.run(documents=[document1])
@@ -200,6 +202,7 @@ def test_pipeline_update_metadata() -> None:
             SentenceSplitter(chunk_size=25, chunk_overlap=0),
         ],
         docstore=SimpleDocumentStore(),
+        docstore_strategy=DocstoreStrategy.DUPLICATES_ONLY,
     )
 
     nodes = pipeline.run(documents=[document1])
@@ -236,6 +239,7 @@ def test_pipeline_dedup_duplicates_only() -> None:
             SentenceSplitter(chunk_size=25, chunk_overlap=0),
         ],
         docstore=SimpleDocumentStore(),
+        docstore_strategy=DocstoreStrategy.DUPLICATES_ONLY,
     )
 
     nodes = pipeline.run(documents=documents)
@@ -255,6 +259,7 @@ def test_pipeline_parallel() -> None:
             SentenceSplitter(chunk_size=25, chunk_overlap=0),
         ],
         docstore=SimpleDocumentStore(),
+        docstore_strategy=DocstoreStrategy.DUPLICATES_ONLY,
     )
 
     num_workers = min(2, cpu_count())
@@ -280,6 +285,7 @@ def __call__(
             RaisingTransform(),
         ],
         docstore=SimpleDocumentStore(),
+        docstore_strategy=DocstoreStrategy.DUPLICATES_ONLY,
     )
 
     with pytest.raises(RuntimeError):
@@ -340,6 +346,7 @@ async def test_async_pipeline_update_text_content() -> None:
             SentenceSplitter(chunk_size=25, chunk_overlap=0),
         ],
         docstore=SimpleDocumentStore(),
+        docstore_strategy=DocstoreStrategy.DUPLICATES_ONLY,
     )
 
     nodes = await pipeline.arun(documents=[document1])
@@ -372,6 +379,7 @@ async def test_async_pipeline_update_metadata() -> None:
             SentenceSplitter(chunk_size=25, chunk_overlap=0),
         ],
         docstore=SimpleDocumentStore(),
+        docstore_strategy=DocstoreStrategy.DUPLICATES_ONLY,
     )
 
     nodes = await pipeline.arun(documents=[document1])
@@ -409,6 +417,7 @@ async def test_async_pipeline_dedup_duplicates_only() -> None:
             SentenceSplitter(chunk_size=25, chunk_overlap=0),
         ],
         docstore=SimpleDocumentStore(),
+        docstore_strategy=DocstoreStrategy.DUPLICATES_ONLY,
     )
 
     nodes = await pipeline.arun(documents=documents)
@@ -429,6 +438,7 @@ async def test_async_pipeline_parallel() -> None:
             SentenceSplitter(chunk_size=25, chunk_overlap=0),
         ],
         docstore=SimpleDocumentStore(),
+        docstore_strategy=DocstoreStrategy.DUPLICATES_ONLY,
     )
 
     num_workers = min(2, cpu_count())
@@ -457,9 +467,37 @@ def __call__(
             RaisingTransform(),
         ],
         docstore=SimpleDocumentStore(),
+        docstore_strategy=DocstoreStrategy.DUPLICATES_ONLY,
     )
 
     with pytest.raises(RuntimeError):
         await pipeline.arun(documents=[document1])
 
     assert pipeline.docstore.get_node("1", raise_error=False) is None
+
+
+def test_docstore_strategy_not_mutated_on_run_without_vector_store() -> None:
+    for strategy in (DocstoreStrategy.UPSERTS, DocstoreStrategy.UPSERTS_AND_DELETE):
+        pipeline = IngestionPipeline(
+            transformations=[],
+            docstore=SimpleDocumentStore(),
+            docstore_strategy=strategy,
+        )
+        with pytest.warns(UserWarning, match="requires a vector store"):
+            pipeline.run(documents=[Document.example()])
+
+        assert pipeline.docstore_strategy is strategy
+
+
+@pytest.mark.asyncio
+async def test_docstore_strategy_not_mutated_on_arun_without_vector_store() -> None:
+    for strategy in (DocstoreStrategy.UPSERTS, DocstoreStrategy.UPSERTS_AND_DELETE):
+        pipeline = IngestionPipeline(
+            transformations=[],
+            docstore=SimpleDocumentStore(),
+            docstore_strategy=strategy,
+        )
+        with pytest.warns(UserWarning, match="requires a vector store"):
+            await pipeline.arun(documents=[Document.example()])
+
+        assert pipeline.docstore_strategy is strategy

llama-index-integrations/graph_stores/llama-index-graph-stores-neo4j/llama_index/graph_stores/neo4j/base.py

@@ -46,10 +46,13 @@ def __init__(
         node_label: str = "Entity",
         refresh_schema: bool = True,
         timeout: Optional[float] = None,
+        user_agent: str = "LLAMAINDEX-GRAPH",
         **kwargs: Any,
     ) -> None:
         self.node_label = node_label
-        self._driver = neo4j.GraphDatabase.driver(url, auth=(username, password))
+        self._driver = neo4j.GraphDatabase.driver(
+            url, auth=(username, password), user_agent=user_agent
+        )
         self._database = database
         self._timeout = timeout
         self.schema = ""