phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-06-28 02:53:30 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat: Add missing Vector Store Files API surface

Browse source 
This adds the ability to list, retrieve, update, and delete Vector
Store Files. It implements these new APIs for the faiss and sqlite-vec
providers, since those are the two that also have the rest of the
vector store files implementation.

There are a number of new integration tests added, which I ran for
each provider as outlined below.

faiss (from ollama distro):

```
INFERENCE_MODEL="meta-llama/Llama-3.2-3B-Instruct" \
llama stack run llama_stack/templates/ollama/run.yaml

LLAMA_STACK_CONFIG=http://localhost:8321 \
pytest -sv tests/integration/vector_io/test_openai_vector_stores.py \
  --embedding-model=all-MiniLM-L6-v2
```

sqlite-vec (from starter distro):

```
llama stack run llama_stack/templates/starter/run.yaml

LLAMA_STACK_CONFIG=http://localhost:8321 \
pytest -sv tests/integration/vector_io/test_openai_vector_stores.py \
  --embedding-model=all-MiniLM-L6-v2
```

Signed-off-by: Ben Browning <bbrownin@redhat.com>
This commit is contained in:
Ben Browning 2025-06-18 09:14:22 -04:00
parent c20388c424
commit 459d50a365
11 changed files with 1248 additions and 20 deletions
Download patch file Download diff file Expand all files Collapse all files

319
docs/_static/llama-stack-spec.html vendored
View file

@ -3241,6 +3241,47 @@
}
},
"/v1/openai/v1/vector_stores/{vector_store_id}/files": {
"get": {
"responses": {
"200": {
"description": "A VectorStoreListFilesResponse containing the list of files.",
"content": {
"application/json": {
"schema": {
"$ref": "#/components/schemas/VectorStoreListFilesResponse"
}
}
}
},
"400": {
"$ref": "#/components/responses/BadRequest400"
},
"429": {
"$ref": "#/components/responses/TooManyRequests429"
},
"500": {
"$ref": "#/components/responses/InternalServerError500"
},
"default": {
"$ref": "#/components/responses/DefaultError"
}
},
"tags": [
"VectorIO"
],
"description": "List files in a vector store.",
"parameters": [
{
"name": "vector_store_id",
"in": "path",
"description": "The ID of the vector store to list files from.",
"required": true,
"schema": {
"type": "string"
}
}
]
},
"post": {
"responses": {
"200": {
@ -3666,6 +3707,168 @@
]
}
},
"/v1/openai/v1/vector_stores/{vector_store_id}/files/{file_id}": {
"get": {
"responses": {
"200": {
"description": "A VectorStoreFileObject representing the file.",
"content": {
"application/json": {
"schema": {
"$ref": "#/components/schemas/VectorStoreFileObject"
}
}
}
},
"400": {
"$ref": "#/components/responses/BadRequest400"
},
"429": {
"$ref": "#/components/responses/TooManyRequests429"
},
"500": {
"$ref": "#/components/responses/InternalServerError500"
},
"default": {
"$ref": "#/components/responses/DefaultError"
}
},
"tags": [
"VectorIO"
],
"description": "Retrieves a vector store file.",
"parameters": [
{
"name": "vector_store_id",
"in": "path",
"description": "The ID of the vector store containing the file to retrieve.",
"required": true,
"schema": {
"type": "string"
}
},
{
"name": "file_id",
"in": "path",
"description": "The ID of the file to retrieve.",
"required": true,
"schema": {
"type": "string"
}
}
]
},
"post": {
"responses": {
"200": {
"description": "A VectorStoreFileObject representing the updated file.",
"content": {
"application/json": {
"schema": {
"$ref": "#/components/schemas/VectorStoreFileObject"
}
}
}
},
"400": {
"$ref": "#/components/responses/BadRequest400"
},
"429": {
"$ref": "#/components/responses/TooManyRequests429"
},
"500": {
"$ref": "#/components/responses/InternalServerError500"
},
"default": {
"$ref": "#/components/responses/DefaultError"
}
},
"tags": [
"VectorIO"
],
"description": "Updates a vector store file.",
"parameters": [
{
"name": "vector_store_id",
"in": "path",
"description": "The ID of the vector store containing the file to update.",
"required": true,
"schema": {
"type": "string"
}
},
{
"name": "file_id",
"in": "path",
"description": "The ID of the file to update.",
"required": true,
"schema": {
"type": "string"
}
}
],
"requestBody": {
"content": {
"application/json": {
"schema": {
"$ref": "#/components/schemas/OpenaiUpdateVectorStoreFileRequest"
}
}
},
"required": true
}
},
"delete": {
"responses": {
"200": {
"description": "A VectorStoreFileDeleteResponse indicating the deletion status.",
"content": {
"application/json": {
"schema": {
"$ref": "#/components/schemas/VectorStoreFileDeleteResponse"
}
}
}
},
"400": {
"$ref": "#/components/responses/BadRequest400"
},
"429": {
"$ref": "#/components/responses/TooManyRequests429"
},
"500": {
"$ref": "#/components/responses/InternalServerError500"
},
"default": {
"$ref": "#/components/responses/DefaultError"
}
},
"tags": [
"VectorIO"
],
"description": "Delete a vector store file.",
"parameters": [
{
"name": "vector_store_id",
"in": "path",
"description": "The ID of the vector store containing the file to delete.",
"required": true,
"schema": {
"type": "string"
}
},
{
"name": "file_id",
"in": "path",
"description": "The ID of the file to delete.",
"required": true,
"schema": {
"type": "string"
}
}
]
}
},
"/v1/openai/v1/embeddings": {
"post": {
"responses": {
@ -12969,6 +13172,35 @@
],
"title": "OpenaiCreateVectorStoreRequest"
},
"VectorStoreFileCounts": {
"type": "object",
"properties": {
"completed": {
"type": "integer"
},
"cancelled": {
"type": "integer"
},
"failed": {
"type": "integer"
},
"in_progress": {
"type": "integer"
},
"total": {
"type": "integer"
}
},
"additionalProperties": false,
"required": [
"completed",
"cancelled",
"failed",
"in_progress",
"total"
],
"title": "VectorStoreFileCounts"
},
"VectorStoreObject": {
"type": "object",
"properties": {
@ -12990,10 +13222,7 @@
"default": 0
},
"file_counts": {
"type": "object",
"additionalProperties": {
"type": "integer"
}
"$ref": "#/components/schemas/VectorStoreFileCounts"
},
"status": {
"type": "string",
@ -13120,6 +13349,30 @@
"title": "VectorStoreDeleteResponse",
"description": "Response from deleting a vector store."
},
"VectorStoreFileDeleteResponse": {
"type": "object",
"properties": {
"id": {
"type": "string"
},
"object": {
"type": "string",
"default": "vector_store.file.deleted"
},
"deleted": {
"type": "boolean",
"default": true
}
},
"additionalProperties": false,
"required": [
"id",
"object",
"deleted"
],
"title": "VectorStoreFileDeleteResponse",
"description": "Response from deleting a vector store file."
},
"OpenaiEmbeddingsRequest": {
"type": "object",
"properties": {
@ -13348,6 +13601,28 @@
"title": "OpenAIFileObject",
"description": "OpenAI File object as defined in the OpenAI Files API."
},
"VectorStoreListFilesResponse": {
"type": "object",
"properties": {
"object": {
"type": "string",
"default": "list"
},
"data": {
"type": "array",
"items": {
"$ref": "#/components/schemas/VectorStoreFileObject"
}
}
},
"additionalProperties": false,
"required": [
"object",
"data"
],
"title": "VectorStoreListFilesResponse",
"description": "Response from listing vector stores."
},
"OpenAIModel": {
"type": "object",
"properties": {
@ -13661,6 +13936,42 @@
"additionalProperties": false,
"title": "OpenaiUpdateVectorStoreRequest"
},
"OpenaiUpdateVectorStoreFileRequest": {
"type": "object",
"properties": {
"attributes": {
"type": "object",
"additionalProperties": {
"oneOf": [
{
"type": "null"
},
{
"type": "boolean"
},
{
"type": "number"
},
{
"type": "string"
},
{
"type": "array"
},
{
"type": "object"
}
]
},
"description": "The updated key-value attributes to store with the file."
}
},
"additionalProperties": false,
"required": [
"attributes"
],
"title": "OpenaiUpdateVectorStoreFileRequest"
},
"DPOAlignmentConfig": {
"type": "object",
"properties": {

224
docs/_static/llama-stack-spec.yaml vendored
View file

@ -2264,6 +2264,36 @@ paths:
$ref: '#/components/schemas/LogEventRequest'
required: true
/v1/openai/v1/vector_stores/{vector_store_id}/files:
get:
responses:
'200':
description: >-
A VectorStoreListFilesResponse containing the list of files.
content:
application/json:
schema:
$ref: '#/components/schemas/VectorStoreListFilesResponse'
'400':
$ref: '#/components/responses/BadRequest400'
'429':
$ref: >-
#/components/responses/TooManyRequests429
'500':
$ref: >-
#/components/responses/InternalServerError500
default:
$ref: '#/components/responses/DefaultError'
tags:
- VectorIO
description: List files in a vector store.
parameters:
- name: vector_store_id
in: path
description: >-
The ID of the vector store to list files from.
required: true
schema:
type: string
post:
responses:
'200':
@ -2572,6 +2602,121 @@ paths:
required: true
schema:
type: string
/v1/openai/v1/vector_stores/{vector_store_id}/files/{file_id}:
get:
responses:
'200':
description: >-
A VectorStoreFileObject representing the file.
content:
application/json:
schema:
$ref: '#/components/schemas/VectorStoreFileObject'
'400':
$ref: '#/components/responses/BadRequest400'
'429':
$ref: >-
#/components/responses/TooManyRequests429
'500':
$ref: >-
#/components/responses/InternalServerError500
default:
$ref: '#/components/responses/DefaultError'
tags:
- VectorIO
description: Retrieves a vector store file.
parameters:
- name: vector_store_id
in: path
description: >-
The ID of the vector store containing the file to retrieve.
required: true
schema:
type: string
- name: file_id
in: path
description: The ID of the file to retrieve.
required: true
schema:
type: string
post:
responses:
'200':
description: >-
A VectorStoreFileObject representing the updated file.
content:
application/json:
schema:
$ref: '#/components/schemas/VectorStoreFileObject'
'400':
$ref: '#/components/responses/BadRequest400'
'429':
$ref: >-
#/components/responses/TooManyRequests429
'500':
$ref: >-
#/components/responses/InternalServerError500
default:
$ref: '#/components/responses/DefaultError'
tags:
- VectorIO
description: Updates a vector store file.
parameters:
- name: vector_store_id
in: path
description: >-
The ID of the vector store containing the file to update.
required: true
schema:
type: string
- name: file_id
in: path
description: The ID of the file to update.
required: true
schema:
type: string
requestBody:
content:
application/json:
schema:
$ref: '#/components/schemas/OpenaiUpdateVectorStoreFileRequest'
required: true
delete:
responses:
'200':
description: >-
A VectorStoreFileDeleteResponse indicating the deletion status.
content:
application/json:
schema:
$ref: '#/components/schemas/VectorStoreFileDeleteResponse'
'400':
$ref: '#/components/responses/BadRequest400'
'429':
$ref: >-
#/components/responses/TooManyRequests429
'500':
$ref: >-
#/components/responses/InternalServerError500
default:
$ref: '#/components/responses/DefaultError'
tags:
- VectorIO
description: Delete a vector store file.
parameters:
- name: vector_store_id
in: path
description: >-
The ID of the vector store containing the file to delete.
required: true
schema:
type: string
- name: file_id
in: path
description: The ID of the file to delete.
required: true
schema:
type: string
/v1/openai/v1/embeddings:
post:
responses:
@ -9031,6 +9176,27 @@ components:
required:
- name
title: OpenaiCreateVectorStoreRequest
VectorStoreFileCounts:
type: object
properties:
completed:
type: integer
cancelled:
type: integer
failed:
type: integer
in_progress:
type: integer
total:
type: integer
additionalProperties: false
required:
- completed
- cancelled
- failed
- in_progress
- total
title: VectorStoreFileCounts
VectorStoreObject:
type: object
properties:
@ -9047,9 +9213,7 @@ components:
type: integer
default: 0
file_counts:
type: object
additionalProperties:
type: integer
$ref: '#/components/schemas/VectorStoreFileCounts'
status:
type: string
default: completed
@ -9129,6 +9293,25 @@ components:
- deleted
title: VectorStoreDeleteResponse
description: Response from deleting a vector store.
VectorStoreFileDeleteResponse:
type: object
properties:
id:
type: string
object:
type: string
default: vector_store.file.deleted
deleted:
type: boolean
default: true
additionalProperties: false
required:
- id
- object
- deleted
title: VectorStoreFileDeleteResponse
description: >-
Response from deleting a vector store file.
OpenaiEmbeddingsRequest:
type: object
properties:
@ -9320,6 +9503,22 @@ components:
title: OpenAIFileObject
description: >-
OpenAI File object as defined in the OpenAI Files API.
VectorStoreListFilesResponse:
type: object
properties:
object:
type: string
default: list
data:
type: array
items:
$ref: '#/components/schemas/VectorStoreFileObject'
additionalProperties: false
required:
- object
- data
title: VectorStoreListFilesResponse
description: Response from listing vector stores.
OpenAIModel:
type: object
properties:
@ -9524,6 +9723,25 @@ components:
Set of 16 key-value pairs that can be attached to an object.
additionalProperties: false
title: OpenaiUpdateVectorStoreRequest
OpenaiUpdateVectorStoreFileRequest:
type: object
properties:
attributes:
type: object
additionalProperties:
oneOf:
- type: 'null'
- type: boolean
- type: number
- type: string
- type: array
- type: object
description: >-
The updated key-value attributes to store with the file.
additionalProperties: false
required:
- attributes
title: OpenaiUpdateVectorStoreFileRequest
DPOAlignmentConfig:
type: object
properties:

84
llama_stack/apis/vector_io/vector_io.py
View file

@ -38,6 +38,15 @@ class QueryChunksResponse(BaseModel):
scores: list[float]
@json_schema_type
class VectorStoreFileCounts(BaseModel):
completed: int
cancelled: int
failed: int
in_progress: int
total: int
@json_schema_type
class VectorStoreObject(BaseModel):
"""OpenAI Vector Store object."""
@ -47,7 +56,7 @@ class VectorStoreObject(BaseModel):
created_at: int
name: str | None = None
usage_bytes: int = 0
file_counts: dict[str, int] = Field(default_factory=dict)
file_counts: VectorStoreFileCounts
status: str = "completed"
expires_after: dict[str, Any] | None = None
expires_at: int | None = None
@ -183,6 +192,23 @@ class VectorStoreFileObject(BaseModel):
vector_store_id: str
@json_schema_type
class VectorStoreListFilesResponse(BaseModel):
"""Response from listing vector stores."""
object: str = "list"
data: list[VectorStoreFileObject]
@json_schema_type
class VectorStoreFileDeleteResponse(BaseModel):
"""Response from deleting a vector store file."""
id: str
object: str = "vector_store.file.deleted"
deleted: bool = True
class VectorDBStore(Protocol):
def get_vector_db(self, vector_db_id: str) -> VectorDB | None: ...
@ -358,3 +384,59 @@ class VectorIO(Protocol):
:returns: A VectorStoreFileObject representing the attached file.
"""
...
@webmethod(route="/openai/v1/vector_stores/{vector_store_id}/files", method="GET")
async def openai_list_files_in_vector_store(
self,
vector_store_id: str,
) -> VectorStoreListFilesResponse:
"""List files in a vector store.
:param vector_store_id: The ID of the vector store to list files from.
:returns: A VectorStoreListFilesResponse containing the list of files.
"""
...
@webmethod(route="/openai/v1/vector_stores/{vector_store_id}/files/{file_id}", method="GET")
async def openai_retrieve_vector_store_file(
self,
vector_store_id: str,
file_id: str,
) -> VectorStoreFileObject:
"""Retrieves a vector store file.
:param vector_store_id: The ID of the vector store containing the file to retrieve.
:param file_id: The ID of the file to retrieve.
:returns: A VectorStoreFileObject representing the file.
"""
...
@webmethod(route="/openai/v1/vector_stores/{vector_store_id}/files/{file_id}", method="POST")
async def openai_update_vector_store_file(
self,
vector_store_id: str,
file_id: str,
attributes: dict[str, Any],
) -> VectorStoreFileObject:
"""Updates a vector store file.
:param vector_store_id: The ID of the vector store containing the file to update.
:param file_id: The ID of the file to update.
:param attributes: The updated key-value attributes to store with the file.
:returns: A VectorStoreFileObject representing the updated file.
"""
...
@webmethod(route="/openai/v1/vector_stores/{vector_store_id}/files/{file_id}", method="DELETE")
async def openai_delete_vector_store_file(
self,
vector_store_id: str,
file_id: str,
) -> VectorStoreFileDeleteResponse:
"""Delete a vector store file.
:param vector_store_id: The ID of the vector store containing the file to delete.
:param file_id: The ID of the file to delete.
:returns: A VectorStoreFileDeleteResponse indicating the deletion status.
"""
...

58
llama_stack/distribution/routers/vector_io.py
View file

@ -21,7 +21,11 @@ from llama_stack.apis.vector_io import (
VectorStoreObject,
VectorStoreSearchResponsePage,
)
from llama_stack.apis.vector_io.vector_io import VectorStoreChunkingStrategy, VectorStoreFileObject
from llama_stack.apis.vector_io.vector_io import (
VectorStoreChunkingStrategy,
VectorStoreFileDeleteResponse,
VectorStoreFileObject,
)
from llama_stack.log import get_logger
from llama_stack.providers.datatypes import HealthResponse, HealthStatus, RoutingTable
@ -279,6 +283,58 @@ class VectorIORouter(VectorIO):
chunking_strategy=chunking_strategy,
)
async def openai_list_files_in_vector_store(
self,
vector_store_id: str,
) -> list[VectorStoreFileObject]:
logger.debug(f"VectorIORouter.openai_list_files_in_vector_store: {vector_store_id}")
# Route based on vector store ID
provider = self.routing_table.get_provider_impl(vector_store_id)
return await provider.openai_list_files_in_vector_store(
vector_store_id=vector_store_id,
)
async def openai_retrieve_vector_store_file(
self,
vector_store_id: str,
file_id: str,
) -> VectorStoreFileObject:
logger.debug(f"VectorIORouter.openai_retrieve_vector_store_file: {vector_store_id}, {file_id}")
# Route based on vector store ID
provider = self.routing_table.get_provider_impl(vector_store_id)
return await provider.openai_retrieve_vector_store_file(
vector_store_id=vector_store_id,
file_id=file_id,
)
async def openai_update_vector_store_file(
self,
vector_store_id: str,
file_id: str,
attributes: dict[str, Any],
) -> VectorStoreFileObject:
logger.debug(f"VectorIORouter.openai_update_vector_store_file: {vector_store_id}, {file_id}")
# Route based on vector store ID
provider = self.routing_table.get_provider_impl(vector_store_id)
return await provider.openai_update_vector_store_file(
vector_store_id=vector_store_id,
file_id=file_id,
attributes=attributes,
)
async def openai_delete_vector_store_file(
self,
vector_store_id: str,
file_id: str,
) -> VectorStoreFileDeleteResponse:
logger.debug(f"VectorIORouter.openai_delete_vector_store_file: {vector_store_id}, {file_id}")
# Route based on vector store ID
provider = self.routing_table.get_provider_impl(vector_store_id)
return await provider.openai_delete_vector_store_file(
vector_store_id=vector_store_id,
file_id=file_id,
)
async def health(self) -> dict[str, HealthResponse]:
health_statuses = {}
timeout = 1 # increasing the timeout to 1 second for health checks

26
llama_stack/providers/inline/vector_io/faiss/faiss.py
View file

@ -45,6 +45,7 @@ VERSION = "v3"
VECTOR_DBS_PREFIX = f"vector_dbs:{VERSION}::"
FAISS_INDEX_PREFIX = f"faiss_index:{VERSION}::"
OPENAI_VECTOR_STORES_PREFIX = f"openai_vector_stores:{VERSION}::"
OPENAI_VECTOR_STORES_FILES_PREFIX = f"openai_vector_stores_files:{VERSION}::"
class FaissIndex(EmbeddingIndex):
@ -283,3 +284,28 @@ class FaissVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtocolPr
assert self.kvstore is not None
key = f"{OPENAI_VECTOR_STORES_PREFIX}{store_id}"
await self.kvstore.delete(key)
async def _save_openai_vector_store_file(self, store_id: str, file_id: str, file_info: dict[str, Any]) -> None:
"""Save vector store file metadata to kvstore."""
assert self.kvstore is not None
key = f"{OPENAI_VECTOR_STORES_FILES_PREFIX}{store_id}:{file_id}"
await self.kvstore.set(key=key, value=json.dumps(file_info))
async def _load_openai_vector_store_file(self, store_id: str, file_id: str) -> dict[str, Any]:
"""Load vector store file metadata from kvstore."""
assert self.kvstore is not None
key = f"{OPENAI_VECTOR_STORES_FILES_PREFIX}{store_id}:{file_id}"
stored_data = await self.kvstore.get(key)
return json.loads(stored_data) if stored_data else {}
async def _update_openai_vector_store_file(self, store_id: str, file_id: str, file_info: dict[str, Any]) -> None:
"""Update vector store file metadata in kvstore."""
assert self.kvstore is not None
key = f"{OPENAI_VECTOR_STORES_FILES_PREFIX}{store_id}:{file_id}"
await self.kvstore.set(key=key, value=json.dumps(file_info))
async def _delete_openai_vector_store_file_from_storage(self, store_id: str, file_id: str) -> None:
"""Delete vector store file metadata from kvstore."""
assert self.kvstore is not None
key = f"{OPENAI_VECTOR_STORES_FILES_PREFIX}{store_id}:{file_id}"
await self.kvstore.delete(key)

92
llama_stack/providers/inline/vector_io/sqlite_vec/sqlite_vec.py
View file

@ -461,6 +461,14 @@ class SQLiteVecVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtoc
metadata TEXT
);
""")
# Create a table to persist OpenAI vector store files.
cur.execute("""
CREATE TABLE IF NOT EXISTS openai_vector_store_files (
store_id TEXT,
file_id TEXT,
metadata TEXT
);
""")
connection.commit()
# Load any existing vector DB registrations.
cur.execute("SELECT metadata FROM vector_dbs")
@ -615,6 +623,90 @@ class SQLiteVecVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtoc
await asyncio.to_thread(_delete)
async def _save_openai_vector_store_file(self, store_id: str, file_id: str, file_info: dict[str, Any]) -> None:
"""Save vector store file metadata to SQLite database."""
def _store():
connection = _create_sqlite_connection(self.config.db_path)
cur = connection.cursor()
try:
cur.execute(
"INSERT OR REPLACE INTO openai_vector_store_files (store_id, file_id, metadata) VALUES (?, ?, ?)",
(store_id, file_id, json.dumps(file_info)),
)
connection.commit()
except Exception as e:
logger.error(f"Error saving openai vector store file {store_id} {file_id}: {e}")
raise
finally:
cur.close()
connection.close()
try:
await asyncio.to_thread(_store)
except Exception as e:
logger.error(f"Error saving openai vector store file {store_id} {file_id}: {e}")
raise
async def _load_openai_vector_store_file(self, store_id: str, file_id: str) -> dict[str, Any]:
"""Load vector store file metadata from SQLite database."""
def _load():
connection = _create_sqlite_connection(self.config.db_path)
cur = connection.cursor()
try:
cur.execute(
"SELECT metadata FROM openai_vector_store_files WHERE store_id = ? AND file_id = ?",
(store_id, file_id),
)
row = cur.fetchone()
print(f"!!! row is {row}")
if row is None:
return None
(metadata,) = row
return metadata
finally:
cur.close()
connection.close()
stored_data = await asyncio.to_thread(_load)
return json.loads(stored_data) if stored_data else {}
async def _update_openai_vector_store_file(self, store_id: str, file_id: str, file_info: dict[str, Any]) -> None:
"""Update vector store file metadata in SQLite database."""
def _update():
connection = _create_sqlite_connection(self.config.db_path)
cur = connection.cursor()
try:
cur.execute(
"UPDATE openai_vector_store_files SET metadata = ? WHERE store_id = ? AND file_id = ?",
(json.dumps(file_info), store_id, file_id),
)
connection.commit()
finally:
cur.close()
connection.close()
await asyncio.to_thread(_update)
async def _delete_openai_vector_store_file_from_storage(self, store_id: str, file_id: str) -> None:
"""Delete vector store file metadata from SQLite database."""
def _delete():
connection = _create_sqlite_connection(self.config.db_path)
cur = connection.cursor()
try:
cur.execute(
"DELETE FROM openai_vector_store_files WHERE store_id = ? AND file_id = ?", (store_id, file_id)
)
connection.commit()
finally:
cur.close()
connection.close()
await asyncio.to_thread(_delete)
async def insert_chunks(self, vector_db_id: str, chunks: list[Chunk], ttl_seconds: int | None = None) -> None:
if vector_db_id not in self.cache:
raise ValueError(f"Vector DB {vector_db_id} not found. Found: {list(self.cache.keys())}")

34
llama_stack/providers/remote/vector_io/chroma/chroma.py
View file

@ -24,7 +24,11 @@ from llama_stack.apis.vector_io import (
VectorStoreObject,
VectorStoreSearchResponsePage,
)
from llama_stack.apis.vector_io.vector_io import VectorStoreChunkingStrategy, VectorStoreFileObject
from llama_stack.apis.vector_io.vector_io import (
VectorStoreChunkingStrategy,
VectorStoreFileObject,
VectorStoreListFilesResponse,
)
from llama_stack.providers.datatypes import Api, VectorDBsProtocolPrivate
from llama_stack.providers.inline.vector_io.chroma import ChromaVectorIOConfig as InlineChromaVectorIOConfig
from llama_stack.providers.utils.memory.vector_store import (
@ -263,3 +267,31 @@ class ChromaVectorIOAdapter(VectorIO, VectorDBsProtocolPrivate):
chunking_strategy: VectorStoreChunkingStrategy | None = None,
) -> VectorStoreFileObject:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Chroma")
async def openai_list_files_in_vector_store(
self,
vector_store_id: str,
) -> VectorStoreListFilesResponse:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Chroma")
async def openai_retrieve_vector_store_file(
self,
vector_store_id: str,
file_id: str,
) -> VectorStoreFileObject:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Chroma")
async def openai_update_vector_store_file(
self,
vector_store_id: str,
file_id: str,
attributes: dict[str, Any] | None = None,
) -> VectorStoreFileObject:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Chroma")
async def openai_delete_vector_store_file(
self,
vector_store_id: str,
file_id: str,
) -> VectorStoreFileObject:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Chroma")

34
llama_stack/providers/remote/vector_io/milvus/milvus.py
View file

@ -26,7 +26,11 @@ from llama_stack.apis.vector_io import (
VectorStoreObject,
VectorStoreSearchResponsePage,
)
from llama_stack.apis.vector_io.vector_io import VectorStoreChunkingStrategy, VectorStoreFileObject
from llama_stack.apis.vector_io.vector_io import (
VectorStoreChunkingStrategy,
VectorStoreFileObject,
VectorStoreListFilesResponse,
)
from llama_stack.providers.datatypes import Api, VectorDBsProtocolPrivate
from llama_stack.providers.inline.vector_io.milvus import MilvusVectorIOConfig as InlineMilvusVectorIOConfig
from llama_stack.providers.utils.memory.vector_store import (
@ -262,6 +266,34 @@ class MilvusVectorIOAdapter(VectorIO, VectorDBsProtocolPrivate):
) -> VectorStoreFileObject:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Milvus")
async def openai_list_files_in_vector_store(
self,
vector_store_id: str,
) -> VectorStoreListFilesResponse:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Milvus")
async def openai_retrieve_vector_store_file(
self,
vector_store_id: str,
file_id: str,
) -> VectorStoreFileObject:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Milvus")
async def openai_update_vector_store_file(
self,
vector_store_id: str,
file_id: str,
attributes: dict[str, Any] | None = None,
) -> VectorStoreFileObject:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Milvus")
async def openai_delete_vector_store_file(
self,
vector_store_id: str,
file_id: str,
) -> VectorStoreFileObject:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Milvus")
def generate_chunk_id(document_id: str, chunk_text: str) -> str:
"""Generate a unique chunk ID using a hash of document ID and chunk text."""

34
llama_stack/providers/remote/vector_io/qdrant/qdrant.py
View file

@ -24,7 +24,11 @@ from llama_stack.apis.vector_io import (
VectorStoreObject,
VectorStoreSearchResponsePage,
)
from llama_stack.apis.vector_io.vector_io import VectorStoreChunkingStrategy, VectorStoreFileObject
from llama_stack.apis.vector_io.vector_io import (
VectorStoreChunkingStrategy,
VectorStoreFileObject,
VectorStoreListFilesResponse,
)
from llama_stack.providers.datatypes import Api, VectorDBsProtocolPrivate
from llama_stack.providers.inline.vector_io.qdrant import QdrantVectorIOConfig as InlineQdrantVectorIOConfig
from llama_stack.providers.utils.memory.vector_store import (
@ -263,3 +267,31 @@ class QdrantVectorIOAdapter(VectorIO, VectorDBsProtocolPrivate):
chunking_strategy: VectorStoreChunkingStrategy | None = None,
) -> VectorStoreFileObject:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Qdrant")
async def openai_list_files_in_vector_store(
self,
vector_store_id: str,
) -> VectorStoreListFilesResponse:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Qdrant")
async def openai_retrieve_vector_store_file(
self,
vector_store_id: str,
file_id: str,
) -> VectorStoreFileObject:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Qdrant")
async def openai_update_vector_store_file(
self,
vector_store_id: str,
file_id: str,
attributes: dict[str, Any] | None = None,
) -> VectorStoreFileObject:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Qdrant")
async def openai_delete_vector_store_file(
self,
vector_store_id: str,
file_id: str,
) -> VectorStoreFileObject:
raise NotImplementedError("OpenAI Vector Stores API is not supported in Qdrant")

147
llama_stack/providers/utils/memory/openai_vector_store_mixin.py
View file

@ -28,8 +28,11 @@ from llama_stack.apis.vector_io.vector_io import (
VectorStoreChunkingStrategy,
VectorStoreChunkingStrategyAuto,
VectorStoreChunkingStrategyStatic,
VectorStoreFileCounts,
VectorStoreFileDeleteResponse,
VectorStoreFileLastError,
VectorStoreFileObject,
VectorStoreListFilesResponse,
)
from llama_stack.providers.utils.memory.vector_store import content_from_data_and_mime_type, make_overlapped_chunks
@ -70,6 +73,26 @@ class OpenAIVectorStoreMixin(ABC):
"""Delete vector store metadata from persistent storage."""
pass
@abstractmethod
async def _save_openai_vector_store_file(self, store_id: str, file_id: str, file_info: dict[str, Any]) -> None:
"""Save vector store file metadata to persistent storage."""
pass
@abstractmethod
async def _load_openai_vector_store_file(self, store_id: str, file_id: str) -> dict[str, Any]:
"""Load vector store file metadata from persistent storage."""
pass
@abstractmethod
async def _update_openai_vector_store_file(self, store_id: str, file_id: str, file_info: dict[str, Any]) -> None:
"""Update vector store file metadata in persistent storage."""
pass
@abstractmethod
async def _delete_openai_vector_store_file_from_storage(self, store_id: str, file_id: str) -> None:
"""Delete vector store file metadata from persistent storage."""
pass
@abstractmethod
async def register_vector_db(self, vector_db: VectorDB) -> None:
"""Register a vector database (provider-specific implementation)."""
@ -136,18 +159,28 @@ class OpenAIVectorStoreMixin(ABC):
await self.register_vector_db(vector_db)
# Create OpenAI vector store metadata
status = "completed"
file_ids = file_ids or []
file_counts = VectorStoreFileCounts(
cancelled=0,
completed=len(file_ids),
failed=0,
in_progress=0,
total=len(file_ids),
)
# TODO: actually attach these files to the vector store...
store_info = {
"id": store_id,
"object": "vector_store",
"created_at": created_at,
"name": store_id,
"usage_bytes": 0,
"file_counts": {},
"status": "completed",
"file_counts": file_counts.model_dump(),
"status": status,
"expires_after": expires_after,
"expires_at": None,
"last_active_at": created_at,
"file_ids": file_ids or [],
"file_ids": file_ids,
"chunking_strategy": chunking_strategy,
}
@ -170,8 +203,8 @@ class OpenAIVectorStoreMixin(ABC):
created_at=created_at,
name=store_id,
usage_bytes=0,
file_counts={},
status="completed",
file_counts=file_counts,
status=status,
expires_after=expires_after,
expires_at=None,
last_active_at=created_at,
@ -455,14 +488,20 @@ class OpenAIVectorStoreMixin(ABC):
attributes: dict[str, Any] | None = None,
chunking_strategy: VectorStoreChunkingStrategy | None = None,
) -> VectorStoreFileObject:
if vector_store_id not in self.openai_vector_stores:
raise ValueError(f"Vector store {vector_store_id} not found")
store_info = self.openai_vector_stores[vector_store_id].copy()
attributes = attributes or {}
chunking_strategy = chunking_strategy or VectorStoreChunkingStrategyAuto()
created_at = int(time.time())
vector_store_file_object = VectorStoreFileObject(
id=file_id,
attributes=attributes,
chunking_strategy=chunking_strategy,
created_at=int(time.time()),
created_at=created_at,
status="in_progress",
vector_store_id=vector_store_id,
)
@ -510,6 +549,20 @@ class OpenAIVectorStoreMixin(ABC):
vector_db_id=vector_store_id,
chunks=chunks,
)
vector_store_file_object.status = "completed"
# Create OpenAI vector store file metadata
file_info = vector_store_file_object.model_dump(exclude={"last_error"})
# Save to persistent storage (provider-specific)
await self._save_openai_vector_store_file(vector_store_id, file_id, file_info)
# Update in-memory cache
store_info["file_ids"].append(file_id)
store_info["file_counts"]["completed"] += 1
store_info["file_counts"]["total"] += 1
self.openai_vector_stores[vector_store_id] = store_info
except Exception as e:
logger.error(f"Error attaching file to vector store: {e}")
vector_store_file_object.status = "failed"
@ -519,6 +572,84 @@ class OpenAIVectorStoreMixin(ABC):
)
return vector_store_file_object
vector_store_file_object.status = "completed"
return vector_store_file_object
async def openai_list_files_in_vector_store(
self,
vector_store_id: str,
) -> VectorStoreListFilesResponse:
"""List files in a vector store."""
if vector_store_id not in self.openai_vector_stores:
raise ValueError(f"Vector store {vector_store_id} not found")
store_info = self.openai_vector_stores[vector_store_id]
file_objects = []
for file_id in store_info["file_ids"]:
file_info = await self._load_openai_vector_store_file(vector_store_id, file_id)
file_objects.append(VectorStoreFileObject(**file_info))
return VectorStoreListFilesResponse(
data=file_objects,
)
async def openai_retrieve_vector_store_file(
self,
vector_store_id: str,
file_id: str,
) -> VectorStoreFileObject:
"""Retrieves a vector store file."""
if vector_store_id not in self.openai_vector_stores:
raise ValueError(f"Vector store {vector_store_id} not found")
store_info = self.openai_vector_stores[vector_store_id]
if file_id not in store_info["file_ids"]:
raise ValueError(f"File {file_id} not found in vector store {vector_store_id}")
file_info = await self._load_openai_vector_store_file(vector_store_id, file_id)
return VectorStoreFileObject(**file_info)
async def openai_update_vector_store_file(
self,
vector_store_id: str,
file_id: str,
attributes: dict[str, Any],
) -> VectorStoreFileObject:
"""Updates a vector store file."""
if vector_store_id not in self.openai_vector_stores:
raise ValueError(f"Vector store {vector_store_id} not found")
store_info = self.openai_vector_stores[vector_store_id]
if file_id not in store_info["file_ids"]:
raise ValueError(f"File {file_id} not found in vector store {vector_store_id}")
file_info = await self._load_openai_vector_store_file(vector_store_id, file_id)
file_info["attributes"] = attributes
await self._update_openai_vector_store_file(vector_store_id, file_id, file_info)
return VectorStoreFileObject(**file_info)
async def openai_delete_vector_store_file(
self,
vector_store_id: str,
file_id: str,
) -> VectorStoreFileDeleteResponse:
"""Deletes a vector store file."""
if vector_store_id not in self.openai_vector_stores:
raise ValueError(f"Vector store {vector_store_id} not found")
store_info = self.openai_vector_stores[vector_store_id].copy()
file = await self.openai_retrieve_vector_store_file(vector_store_id, file_id)
await self._delete_openai_vector_store_file_from_storage(vector_store_id, file_id)
# Update in-memory cache
store_info["file_ids"].remove(file_id)
store_info["file_counts"][file.status] -= 1
store_info["file_counts"]["total"] -= 1
self.openai_vector_stores[vector_store_id] = store_info
return VectorStoreFileDeleteResponse(
id=file_id,
deleted=True,
)

216
tests/integration/vector_io/test_openai_vector_stores.py
View file

@ -6,8 +6,11 @@
import logging
import time
from io import BytesIO
import pytest
from llama_stack_client import BadRequestError, LlamaStackClient
from openai import BadRequestError as OpenAIBadRequestError
from openai import OpenAI
from llama_stack.apis.vector_io import Chunk
@ -73,11 +76,23 @@ def compat_client_with_empty_stores(compat_client):
logger.warning("Failed to clear vector stores")
pass
def clear_files():
try:
response = compat_client.files.list()
for file in response.data:
compat_client.files.delete(file_id=file.id)
except Exception:
# If the API is not available or fails, just continue
logger.warning("Failed to clear files")
pass
clear_vector_stores()
clear_files()
yield compat_client
# Clean up after the test
clear_vector_stores()
clear_files()
def test_openai_create_vector_store(compat_client_with_empty_stores, client_with_models):
@ -423,3 +438,204 @@ def test_openai_vector_store_search_with_max_num_results(
assert search_response is not None
assert len(search_response.data) == 2
def test_openai_vector_store_attach_file_response_attributes(compat_client_with_empty_stores, client_with_models):
"""Test OpenAI vector store attach file."""
skip_if_provider_doesnt_support_openai_vector_stores(client_with_models)
if isinstance(compat_client_with_empty_stores, LlamaStackClient):
pytest.skip("Vector Store Files attach is not yet supported with LlamaStackClient")
compat_client = compat_client_with_empty_stores
# Create a vector store
vector_store = compat_client.vector_stores.create(name="test_store")
# Create a file
test_content = b"This is a test file"
with BytesIO(test_content) as file_buffer:
file_buffer.name = "openai_test.txt"
file = compat_client.files.create(file=file_buffer, purpose="assistants")
# Attach the file to the vector store
file_attach_response = compat_client.vector_stores.files.create(
vector_store_id=vector_store.id,
file_id=file.id,
)
assert file_attach_response
assert file_attach_response.object == "vector_store.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.created_at > 0
assert not file_attach_response.last_error
updated_vector_store = compat_client.vector_stores.retrieve(vector_store_id=vector_store.id)
assert updated_vector_store.file_counts.completed == 1
assert updated_vector_store.file_counts.total == 1
assert updated_vector_store.file_counts.cancelled == 0
assert updated_vector_store.file_counts.failed == 0
assert updated_vector_store.file_counts.in_progress == 0
def test_openai_vector_store_list_files(compat_client_with_empty_stores, client_with_models):
"""Test OpenAI vector store list files."""
skip_if_provider_doesnt_support_openai_vector_stores(client_with_models)
if isinstance(compat_client_with_empty_stores, LlamaStackClient):
pytest.skip("Vector Store Files list is not yet supported with LlamaStackClient")
compat_client = compat_client_with_empty_stores
# Create a vector store
vector_store = compat_client.vector_stores.create(name="test_store")
# Create some files and attach them to the vector store
file_ids = []
for i in range(3):
with BytesIO(f"This is a test file {i}".encode()) as file_buffer:
file_buffer.name = f"openai_test_{i}.txt"
file = compat_client.files.create(file=file_buffer, purpose="assistants")
compat_client.vector_stores.files.create(
vector_store_id=vector_store.id,
file_id=file.id,
)
file_ids.append(file.id)
files_list = compat_client.vector_stores.files.list(vector_store_id=vector_store.id)
assert files_list
assert files_list.object == "list"
assert files_list.data
assert len(files_list.data) == 3
assert 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].created_at > 0
assert not files_list.data[0].last_error
updated_vector_store = compat_client.vector_stores.retrieve(vector_store_id=vector_store.id)
assert updated_vector_store.file_counts.completed == 3
assert updated_vector_store.file_counts.total == 3
assert updated_vector_store.file_counts.cancelled == 0
assert updated_vector_store.file_counts.failed == 0
assert updated_vector_store.file_counts.in_progress == 0
def test_openai_vector_store_list_files_invalid_vector_store(compat_client_with_empty_stores, client_with_models):
"""Test OpenAI vector store list files with invalid vector store ID."""
skip_if_provider_doesnt_support_openai_vector_stores(client_with_models)
if isinstance(compat_client_with_empty_stores, LlamaStackClient):
pytest.skip("Vector Store Files list is not yet supported with LlamaStackClient")
compat_client = compat_client_with_empty_stores
with pytest.raises((BadRequestError, OpenAIBadRequestError)):
compat_client.vector_stores.files.list(vector_store_id="abc123")
def test_openai_vector_store_delete_file(compat_client_with_empty_stores, client_with_models):
"""Test OpenAI vector store delete file."""
skip_if_provider_doesnt_support_openai_vector_stores(client_with_models)
if isinstance(compat_client_with_empty_stores, LlamaStackClient):
pytest.skip("Vector Store Files list is not yet supported with LlamaStackClient")
compat_client = compat_client_with_empty_stores
# Create a vector store
vector_store = compat_client.vector_stores.create(name="test_store")
# Create some files and attach them to the vector store
file_ids = []
for i in range(3):
with BytesIO(f"This is a test file {i}".encode()) as file_buffer:
file_buffer.name = f"openai_test_{i}.txt"
file = compat_client.files.create(file=file_buffer, purpose="assistants")
compat_client.vector_stores.files.create(
vector_store_id=vector_store.id,
file_id=file.id,
)
file_ids.append(file.id)
files_list = compat_client.vector_stores.files.list(vector_store_id=vector_store.id)
assert len(files_list.data) == 3
# Delete the first file
delete_response = compat_client.vector_stores.files.delete(vector_store_id=vector_store.id, file_id=file_ids[0])
assert delete_response
assert delete_response.id == file_ids[0]
assert delete_response.deleted is True
assert delete_response.object == "vector_store.file.deleted"
updated_vector_store = compat_client.vector_stores.retrieve(vector_store_id=vector_store.id)
assert updated_vector_store.file_counts.completed == 2
assert updated_vector_store.file_counts.total == 2
assert updated_vector_store.file_counts.cancelled == 0
assert updated_vector_store.file_counts.failed == 0
assert updated_vector_store.file_counts.in_progress == 0
# Delete the second file
delete_response = compat_client.vector_stores.files.delete(vector_store_id=vector_store.id, file_id=file_ids[1])
assert delete_response
assert delete_response.id == file_ids[1]
updated_vector_store = compat_client.vector_stores.retrieve(vector_store_id=vector_store.id)
assert updated_vector_store.file_counts.completed == 1
assert updated_vector_store.file_counts.total == 1
assert updated_vector_store.file_counts.cancelled == 0
assert updated_vector_store.file_counts.failed == 0
assert updated_vector_store.file_counts.in_progress == 0
def test_openai_vector_store_update_file(compat_client_with_empty_stores, client_with_models):
"""Test OpenAI vector store update file."""
skip_if_provider_doesnt_support_openai_vector_stores(client_with_models)
if isinstance(compat_client_with_empty_stores, LlamaStackClient):
pytest.skip("Vector Store Files update is not yet supported with LlamaStackClient")
compat_client = compat_client_with_empty_stores
# Create a vector store
vector_store = compat_client.vector_stores.create(name="test_store")
# Create a file
test_content = b"This is a test file"
with BytesIO(test_content) as file_buffer:
file_buffer.name = "openai_test.txt"
file = compat_client.files.create(file=file_buffer, purpose="assistants")
# Attach the file to the vector store
file_attach_response = compat_client.vector_stores.files.create(
vector_store_id=vector_store.id,
file_id=file.id,
attributes={"foo": "bar"},
)
assert file_attach_response.status == "completed"
assert file_attach_response.attributes["foo"] == "bar"
# Update the file's attributes
updated_response = compat_client.vector_stores.files.update(
vector_store_id=vector_store.id,
file_id=file.id,
attributes={"foo": "baz"},
)
assert updated_response.status == "completed"
assert updated_response.attributes["foo"] == "baz"
# Ensure we can retrieve the file and see the updated attributes
retrieved_file = compat_client.vector_stores.files.retrieve(
vector_store_id=vector_store.id,
file_id=file.id,
)
assert retrieved_file.attributes["foo"] == "baz"