From 6097d9329405eeac957511f86f4e4eaacc2bc1b9 Mon Sep 17 00:00:00 2001
From: Eric Huang <erichuang@meta.com>
Date: Wed, 5 Nov 2025 13:52:54 -0800
Subject: [PATCH] fix: actualize chunking strategy in vector store create API

# What does this PR do?


## Test Plan
---
 src/llama_stack/core/routers/vector_io.py                | 9 +++++++++
 .../providers/utils/memory/openai_vector_store_mixin.py  | 2 +-
 tests/integration/vector_io/test_openai_vector_stores.py | 6 +++---
 3 files changed, 13 insertions(+), 4 deletions(-)

diff --git a/src/llama_stack/core/routers/vector_io.py b/src/llama_stack/core/routers/vector_io.py
index 78b38ba95..2e2b8cbca 100644
--- a/src/llama_stack/core/routers/vector_io.py
+++ b/src/llama_stack/core/routers/vector_io.py
@@ -20,6 +20,8 @@ from llama_stack.apis.vector_io import (
     SearchRankingOptions,
     VectorIO,
     VectorStoreChunkingStrategy,
+    VectorStoreChunkingStrategyStatic,
+    VectorStoreChunkingStrategyStaticConfig,
     VectorStoreDeleteResponse,
     VectorStoreFileBatchObject,
     VectorStoreFileContentsResponse,
@@ -167,6 +169,13 @@ class VectorIORouter(VectorIO):
         if embedding_dimension is not None:
             params.model_extra["embedding_dimension"] = embedding_dimension
 
+        # Set chunking strategy explicitly if not provided
+        if params.chunking_strategy is None or params.chunking_strategy.type == "auto":
+            # actualize the chunking strategy to static
+            params.chunking_strategy = VectorStoreChunkingStrategyStatic(
+                static=VectorStoreChunkingStrategyStaticConfig()
+            )
+
         return await provider.openai_create_vector_store(params)
 
     async def openai_list_vector_stores(
diff --git a/src/llama_stack/providers/utils/memory/openai_vector_store_mixin.py b/src/llama_stack/providers/utils/memory/openai_vector_store_mixin.py
index 41d4cb2d7..8a1cb0e66 100644
--- a/src/llama_stack/providers/utils/memory/openai_vector_store_mixin.py
+++ b/src/llama_stack/providers/utils/memory/openai_vector_store_mixin.py
@@ -426,7 +426,7 @@ class OpenAIVectorStoreMixin(ABC):
             "expires_at": None,
             "last_active_at": created_at,
             "file_ids": [],
-            "chunking_strategy": params.chunking_strategy,
+            "chunking_strategy": params.chunking_strategy.model_dump(),
         }
 
         # Add provider information to metadata if provided
diff --git a/tests/integration/vector_io/test_openai_vector_stores.py b/tests/integration/vector_io/test_openai_vector_stores.py
index 9da5dd25b..38a4b8d66 100644
--- a/tests/integration/vector_io/test_openai_vector_stores.py
+++ b/tests/integration/vector_io/test_openai_vector_stores.py
@@ -679,7 +679,7 @@ def test_openai_vector_store_attach_file(
     assert file_attach_response.id == file.id
     assert file_attach_response.vector_store_id == vector_store.id
     assert file_attach_response.status == "completed"
-    assert file_attach_response.chunking_strategy.type == "auto"
+    assert file_attach_response.chunking_strategy.type == "static"
     assert file_attach_response.created_at > 0
     assert not file_attach_response.last_error
 
@@ -815,8 +815,8 @@ def test_openai_vector_store_list_files(
     assert set(file_ids) == {file.id for file in files_list.data}
     assert files_list.data[0].object == "vector_store.file"
     assert files_list.data[0].vector_store_id == vector_store.id
-    assert files_list.data[0].status == "completed"
-    assert files_list.data[0].chunking_strategy.type == "auto"
+    assert files_list.data[0].status in ["completed", "in_progress"]
+    assert files_list.data[0].chunking_strategy.type == "static"
     assert files_list.data[0].created_at > 0
     assert files_list.first_id == files_list.data[0].id
     assert not files_list.data[0].last_error