fix!: BREAKING CHANGE: vector_store: search API response fix (#4080)

# What does this PR do?
- search_query in the vector store search API should be a list,
according to https://github.com/openai/openai-openapi


## Test Plan
modified tests


---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with
[ReviewStack](https://reviewstack.dev/llamastack/llama-stack/pull/4080).
* #4086
* __->__ #4080

client-sdks/stainless/openapi.yml

@@ -9976,6 +9976,70 @@ components:
         - metadata
       title: VectorStoreObject
       description: OpenAI Vector Store object.
+    VectorStoreChunkingStrategy:
+      oneOf:
+        - $ref: '#/components/schemas/VectorStoreChunkingStrategyAuto'
+        - $ref: '#/components/schemas/VectorStoreChunkingStrategyStatic'
+      discriminator:
+        propertyName: type
+        mapping:
+          auto: '#/components/schemas/VectorStoreChunkingStrategyAuto'
+          static: '#/components/schemas/VectorStoreChunkingStrategyStatic'
+    VectorStoreChunkingStrategyAuto:
+      type: object
+      properties:
+        type:
+          type: string
+          const: auto
+          default: auto
+          description: >-
+            Strategy type, always "auto" for automatic chunking
+      additionalProperties: false
+      required:
+        - type
+      title: VectorStoreChunkingStrategyAuto
+      description: >-
+        Automatic chunking strategy for vector store files.
+    VectorStoreChunkingStrategyStatic:
+      type: object
+      properties:
+        type:
+          type: string
+          const: static
+          default: static
+          description: >-
+            Strategy type, always "static" for static chunking
+        static:
+          $ref: '#/components/schemas/VectorStoreChunkingStrategyStaticConfig'
+          description: >-
+            Configuration parameters for the static chunking strategy
+      additionalProperties: false
+      required:
+        - type
+        - static
+      title: VectorStoreChunkingStrategyStatic
+      description: >-
+        Static chunking strategy with configurable parameters.
+    VectorStoreChunkingStrategyStaticConfig:
+      type: object
+      properties:
+        chunk_overlap_tokens:
+          type: integer
+          default: 400
+          description: >-
+            Number of tokens to overlap between adjacent chunks
+        max_chunk_size_tokens:
+          type: integer
+          default: 800
+          description: >-
+            Maximum number of tokens per chunk, must be between 100 and 4096
+      additionalProperties: false
+      required:
+        - chunk_overlap_tokens
+        - max_chunk_size_tokens
+      title: VectorStoreChunkingStrategyStaticConfig
+      description: >-
+        Configuration for static chunking strategy.
     "OpenAICreateVectorStoreRequestWithExtraBody":
       type: object
       properties:
@@ -10001,15 +10065,7 @@ components:
           description: >-
             (Optional) Expiration policy for the vector store
         chunking_strategy:
-          type: object
-          additionalProperties:
-            oneOf:
-              - type: 'null'
-              - type: boolean
-              - type: number
-              - type: string
-              - type: array
-              - type: object
+          $ref: '#/components/schemas/VectorStoreChunkingStrategy'
           description: >-
             (Optional) Strategy for splitting files into chunks
         metadata:
@@ -10085,70 +10141,6 @@ components:
         - deleted
       title: VectorStoreDeleteResponse
       description: Response from deleting a vector store.
-    VectorStoreChunkingStrategy:
-      oneOf:
-        - $ref: '#/components/schemas/VectorStoreChunkingStrategyAuto'
-        - $ref: '#/components/schemas/VectorStoreChunkingStrategyStatic'
-      discriminator:
-        propertyName: type
-        mapping:
-          auto: '#/components/schemas/VectorStoreChunkingStrategyAuto'
-          static: '#/components/schemas/VectorStoreChunkingStrategyStatic'
-    VectorStoreChunkingStrategyAuto:
-      type: object
-      properties:
-        type:
-          type: string
-          const: auto
-          default: auto
-          description: >-
-            Strategy type, always "auto" for automatic chunking
-      additionalProperties: false
-      required:
-        - type
-      title: VectorStoreChunkingStrategyAuto
-      description: >-
-        Automatic chunking strategy for vector store files.
-    VectorStoreChunkingStrategyStatic:
-      type: object
-      properties:
-        type:
-          type: string
-          const: static
-          default: static
-          description: >-
-            Strategy type, always "static" for static chunking
-        static:
-          $ref: '#/components/schemas/VectorStoreChunkingStrategyStaticConfig'
-          description: >-
-            Configuration parameters for the static chunking strategy
-      additionalProperties: false
-      required:
-        - type
-        - static
-      title: VectorStoreChunkingStrategyStatic
-      description: >-
-        Static chunking strategy with configurable parameters.
-    VectorStoreChunkingStrategyStaticConfig:
-      type: object
-      properties:
-        chunk_overlap_tokens:
-          type: integer
-          default: 400
-          description: >-
-            Number of tokens to overlap between adjacent chunks
-        max_chunk_size_tokens:
-          type: integer
-          default: 800
-          description: >-
-            Maximum number of tokens per chunk, must be between 100 and 4096
-      additionalProperties: false
-      required:
-        - chunk_overlap_tokens
-        - max_chunk_size_tokens
-      title: VectorStoreChunkingStrategyStaticConfig
-      description: >-
-        Configuration for static chunking strategy.
     "OpenAICreateVectorStoreFileBatchRequestWithExtraBody":
       type: object
       properties:
@@ -10606,7 +10598,9 @@ components:
           description: >-
             Object type identifier for the search results page
         search_query:
-          type: string
+          type: array
+          items:
+            type: string
           description: >-
             The original search query that was executed
         data:

docs/static/llama-stack-spec.yaml

@@ -9260,6 +9260,70 @@ components:
         - metadata
       title: VectorStoreObject
       description: OpenAI Vector Store object.
+    VectorStoreChunkingStrategy:
+      oneOf:
+        - $ref: '#/components/schemas/VectorStoreChunkingStrategyAuto'
+        - $ref: '#/components/schemas/VectorStoreChunkingStrategyStatic'
+      discriminator:
+        propertyName: type
+        mapping:
+          auto: '#/components/schemas/VectorStoreChunkingStrategyAuto'
+          static: '#/components/schemas/VectorStoreChunkingStrategyStatic'
+    VectorStoreChunkingStrategyAuto:
+      type: object
+      properties:
+        type:
+          type: string
+          const: auto
+          default: auto
+          description: >-
+            Strategy type, always "auto" for automatic chunking
+      additionalProperties: false
+      required:
+        - type
+      title: VectorStoreChunkingStrategyAuto
+      description: >-
+        Automatic chunking strategy for vector store files.
+    VectorStoreChunkingStrategyStatic:
+      type: object
+      properties:
+        type:
+          type: string
+          const: static
+          default: static
+          description: >-
+            Strategy type, always "static" for static chunking
+        static:
+          $ref: '#/components/schemas/VectorStoreChunkingStrategyStaticConfig'
+          description: >-
+            Configuration parameters for the static chunking strategy
+      additionalProperties: false
+      required:
+        - type
+        - static
+      title: VectorStoreChunkingStrategyStatic
+      description: >-
+        Static chunking strategy with configurable parameters.
+    VectorStoreChunkingStrategyStaticConfig:
+      type: object
+      properties:
+        chunk_overlap_tokens:
+          type: integer
+          default: 400
+          description: >-
+            Number of tokens to overlap between adjacent chunks
+        max_chunk_size_tokens:
+          type: integer
+          default: 800
+          description: >-
+            Maximum number of tokens per chunk, must be between 100 and 4096
+      additionalProperties: false
+      required:
+        - chunk_overlap_tokens
+        - max_chunk_size_tokens
+      title: VectorStoreChunkingStrategyStaticConfig
+      description: >-
+        Configuration for static chunking strategy.
     "OpenAICreateVectorStoreRequestWithExtraBody":
       type: object
       properties:
@@ -9285,15 +9349,7 @@ components:
           description: >-
             (Optional) Expiration policy for the vector store
         chunking_strategy:
-          type: object
-          additionalProperties:
-            oneOf:
-              - type: 'null'
-              - type: boolean
-              - type: number
-              - type: string
-              - type: array
-              - type: object
+          $ref: '#/components/schemas/VectorStoreChunkingStrategy'
           description: >-
             (Optional) Strategy for splitting files into chunks
         metadata:
@@ -9369,70 +9425,6 @@ components:
         - deleted
       title: VectorStoreDeleteResponse
       description: Response from deleting a vector store.
-    VectorStoreChunkingStrategy:
-      oneOf:
-        - $ref: '#/components/schemas/VectorStoreChunkingStrategyAuto'
-        - $ref: '#/components/schemas/VectorStoreChunkingStrategyStatic'
-      discriminator:
-        propertyName: type
-        mapping:
-          auto: '#/components/schemas/VectorStoreChunkingStrategyAuto'
-          static: '#/components/schemas/VectorStoreChunkingStrategyStatic'
-    VectorStoreChunkingStrategyAuto:
-      type: object
-      properties:
-        type:
-          type: string
-          const: auto
-          default: auto
-          description: >-
-            Strategy type, always "auto" for automatic chunking
-      additionalProperties: false
-      required:
-        - type
-      title: VectorStoreChunkingStrategyAuto
-      description: >-
-        Automatic chunking strategy for vector store files.
-    VectorStoreChunkingStrategyStatic:
-      type: object
-      properties:
-        type:
-          type: string
-          const: static
-          default: static
-          description: >-
-            Strategy type, always "static" for static chunking
-        static:
-          $ref: '#/components/schemas/VectorStoreChunkingStrategyStaticConfig'
-          description: >-
-            Configuration parameters for the static chunking strategy
-      additionalProperties: false
-      required:
-        - type
-        - static
-      title: VectorStoreChunkingStrategyStatic
-      description: >-
-        Static chunking strategy with configurable parameters.
-    VectorStoreChunkingStrategyStaticConfig:
-      type: object
-      properties:
-        chunk_overlap_tokens:
-          type: integer
-          default: 400
-          description: >-
-            Number of tokens to overlap between adjacent chunks
-        max_chunk_size_tokens:
-          type: integer
-          default: 800
-          description: >-
-            Maximum number of tokens per chunk, must be between 100 and 4096
-      additionalProperties: false
-      required:
-        - chunk_overlap_tokens
-        - max_chunk_size_tokens
-      title: VectorStoreChunkingStrategyStaticConfig
-      description: >-
-        Configuration for static chunking strategy.
     "OpenAICreateVectorStoreFileBatchRequestWithExtraBody":
       type: object
       properties:
@@ -9890,7 +9882,9 @@ components:
           description: >-
             Object type identifier for the search results page
         search_query:
-          type: string
+          type: array
+          items:
+            type: string
           description: >-
             The original search query that was executed
         data:

docs/static/stainless-llama-stack-spec.yaml

@@ -9976,6 +9976,70 @@ components:
         - metadata
       title: VectorStoreObject
       description: OpenAI Vector Store object.
+    VectorStoreChunkingStrategy:
+      oneOf:
+        - $ref: '#/components/schemas/VectorStoreChunkingStrategyAuto'
+        - $ref: '#/components/schemas/VectorStoreChunkingStrategyStatic'
+      discriminator:
+        propertyName: type
+        mapping:
+          auto: '#/components/schemas/VectorStoreChunkingStrategyAuto'
+          static: '#/components/schemas/VectorStoreChunkingStrategyStatic'
+    VectorStoreChunkingStrategyAuto:
+      type: object
+      properties:
+        type:
+          type: string
+          const: auto
+          default: auto
+          description: >-
+            Strategy type, always "auto" for automatic chunking
+      additionalProperties: false
+      required:
+        - type
+      title: VectorStoreChunkingStrategyAuto
+      description: >-
+        Automatic chunking strategy for vector store files.
+    VectorStoreChunkingStrategyStatic:
+      type: object
+      properties:
+        type:
+          type: string
+          const: static
+          default: static
+          description: >-
+            Strategy type, always "static" for static chunking
+        static:
+          $ref: '#/components/schemas/VectorStoreChunkingStrategyStaticConfig'
+          description: >-
+            Configuration parameters for the static chunking strategy
+      additionalProperties: false
+      required:
+        - type
+        - static
+      title: VectorStoreChunkingStrategyStatic
+      description: >-
+        Static chunking strategy with configurable parameters.
+    VectorStoreChunkingStrategyStaticConfig:
+      type: object
+      properties:
+        chunk_overlap_tokens:
+          type: integer
+          default: 400
+          description: >-
+            Number of tokens to overlap between adjacent chunks
+        max_chunk_size_tokens:
+          type: integer
+          default: 800
+          description: >-
+            Maximum number of tokens per chunk, must be between 100 and 4096
+      additionalProperties: false
+      required:
+        - chunk_overlap_tokens
+        - max_chunk_size_tokens
+      title: VectorStoreChunkingStrategyStaticConfig
+      description: >-
+        Configuration for static chunking strategy.
     "OpenAICreateVectorStoreRequestWithExtraBody":
       type: object
       properties:
@@ -10001,15 +10065,7 @@ components:
           description: >-
             (Optional) Expiration policy for the vector store
         chunking_strategy:
-          type: object
-          additionalProperties:
-            oneOf:
-              - type: 'null'
-              - type: boolean
-              - type: number
-              - type: string
-              - type: array
-              - type: object
+          $ref: '#/components/schemas/VectorStoreChunkingStrategy'
           description: >-
             (Optional) Strategy for splitting files into chunks
         metadata:
@@ -10085,70 +10141,6 @@ components:
         - deleted
       title: VectorStoreDeleteResponse
       description: Response from deleting a vector store.
-    VectorStoreChunkingStrategy:
-      oneOf:
-        - $ref: '#/components/schemas/VectorStoreChunkingStrategyAuto'
-        - $ref: '#/components/schemas/VectorStoreChunkingStrategyStatic'
-      discriminator:
-        propertyName: type
-        mapping:
-          auto: '#/components/schemas/VectorStoreChunkingStrategyAuto'
-          static: '#/components/schemas/VectorStoreChunkingStrategyStatic'
-    VectorStoreChunkingStrategyAuto:
-      type: object
-      properties:
-        type:
-          type: string
-          const: auto
-          default: auto
-          description: >-
-            Strategy type, always "auto" for automatic chunking
-      additionalProperties: false
-      required:
-        - type
-      title: VectorStoreChunkingStrategyAuto
-      description: >-
-        Automatic chunking strategy for vector store files.
-    VectorStoreChunkingStrategyStatic:
-      type: object
-      properties:
-        type:
-          type: string
-          const: static
-          default: static
-          description: >-
-            Strategy type, always "static" for static chunking
-        static:
-          $ref: '#/components/schemas/VectorStoreChunkingStrategyStaticConfig'
-          description: >-
-            Configuration parameters for the static chunking strategy
-      additionalProperties: false
-      required:
-        - type
-        - static
-      title: VectorStoreChunkingStrategyStatic
-      description: >-
-        Static chunking strategy with configurable parameters.
-    VectorStoreChunkingStrategyStaticConfig:
-      type: object
-      properties:
-        chunk_overlap_tokens:
-          type: integer
-          default: 400
-          description: >-
-            Number of tokens to overlap between adjacent chunks
-        max_chunk_size_tokens:
-          type: integer
-          default: 800
-          description: >-
-            Maximum number of tokens per chunk, must be between 100 and 4096
-      additionalProperties: false
-      required:
-        - chunk_overlap_tokens
-        - max_chunk_size_tokens
-      title: VectorStoreChunkingStrategyStaticConfig
-      description: >-
-        Configuration for static chunking strategy.
     "OpenAICreateVectorStoreFileBatchRequestWithExtraBody":
       type: object
       properties:
@@ -10606,7 +10598,9 @@ components:
           description: >-
             Object type identifier for the search results page
         search_query:
-          type: string
+          type: array
+          items:
+            type: string
           description: >-
             The original search query that was executed
         data:

src/llama_stack/apis/vector_io/vector_io.py

@@ -260,7 +260,7 @@ class VectorStoreSearchResponsePage(BaseModel):
     """
 
     object: str = "vector_store.search_results.page"
-    search_query: str
+    search_query: list[str]
     data: list[VectorStoreSearchResponse]
     has_more: bool = False
     next_page: str | None = None
@@ -478,7 +478,7 @@ class OpenAICreateVectorStoreRequestWithExtraBody(BaseModel, extra="allow"):
     name: str | None = None
     file_ids: list[str] | None = None
     expires_after: dict[str, Any] | None = None
-    chunking_strategy: dict[str, Any] | None = None
+    chunking_strategy: VectorStoreChunkingStrategy | None = None
     metadata: dict[str, Any] | None = None
 
 

src/llama_stack/providers/utils/memory/openai_vector_store_mixin.py

@@ -637,7 +637,7 @@ class OpenAIVectorStoreMixin(ABC):
                     break
 
             return VectorStoreSearchResponsePage(
-                search_query=search_query,
+                search_query=query if isinstance(query, list) else [query],
                 data=data,
                 has_more=False,  # For simplicity, we don't implement pagination here
                 next_page=None,
@@ -647,7 +647,7 @@ class OpenAIVectorStoreMixin(ABC):
             logger.error(f"Error searching vector store {vector_store_id}: {e}")
             # Return empty results on error
             return VectorStoreSearchResponsePage(
-                search_query=search_query,
+                search_query=query if isinstance(query, list) else [query],
                 data=[],
                 has_more=False,
                 next_page=None,

tests/integration/vector_io/test_openai_vector_stores.py

@@ -350,7 +350,7 @@ def test_openai_vector_store_search_empty(
     assert search_response is not None
     assert hasattr(search_response, "data")
     assert len(search_response.data) == 0  # Empty store should return no results
-    assert search_response.search_query == "test query"
+    assert search_response.search_query == ["test query"]
     assert search_response.has_more is False