refactor(storage): remove core storage api shims and fix imports

src/llama_stack/core/conversations/conversations.py

@@ -11,9 +11,9 @@ from typing import Any, Literal
 from pydantic import BaseModel, TypeAdapter
 
 from llama_stack.core.datatypes import AccessRule, StackRunConfig
-from llama_stack.log import get_logger
 from llama_stack.core.storage.sqlstore.authorized_sqlstore import AuthorizedSqlStore
 from llama_stack.core.storage.sqlstore.sqlstore import sqlstore_impl
+from llama_stack.log import get_logger
 from llama_stack_api import (
     Conversation,
     ConversationDeletedResource,

src/llama_stack/core/server/quota.py

@@ -11,8 +11,8 @@ from datetime import UTC, datetime, timedelta
 from starlette.types import ASGIApp, Receive, Scope, Send
 
 from llama_stack.core.storage.datatypes import KVStoreReference, StorageBackendType
-from llama_stack.log import get_logger
 from llama_stack.core.storage.kvstore.kvstore import _KVSTORE_BACKENDS, kvstore_impl
+from llama_stack.log import get_logger
 from llama_stack_api.internal.kvstore import KVStore
 
 logger = get_logger(name=__name__, category="core::server")

src/llama_stack/core/storage/kvstore/__init__.py

@@ -4,4 +4,6 @@
 # This source code is licensed under the terms described in the LICENSE file in
 # the root directory of this source tree.
 
+from llama_stack_api.internal.kvstore import KVStore as KVStore
+
 from .kvstore import *  # noqa: F401, F403

src/llama_stack/core/storage/kvstore/api.py

@@ -1,9 +0,0 @@
-# Copyright (c) Meta Platforms, Inc. and affiliates.
-# All rights reserved.
-#
-# This source code is licensed under the terms described in the LICENSE file in
-# the root directory of this source tree.
-
-from llama_stack_api.internal.kvstore import KVStore
-
-__all__ = ["KVStore"]

src/llama_stack/core/storage/kvstore/kvstore.py

@@ -13,11 +13,19 @@ from __future__ import annotations
 
 import asyncio
 from collections import defaultdict
+from datetime import datetime
+from typing import cast
 
-from llama_stack.core.storage.datatypes import KVStoreReference, StorageBackendConfig, StorageBackendType
+from llama_stack.core.storage.datatypes import KVStoreReference, StorageBackendConfig
+from llama_stack_api.internal.kvstore import KVStore
 
-from .api import KVStore
-from .config import KVStoreConfig
+from .config import (
+    KVStoreConfig,
+    MongoDBKVStoreConfig,
+    PostgresKVStoreConfig,
+    RedisKVStoreConfig,
+    SqliteKVStoreConfig,
+)
 
 
 def kvstore_dependencies():
@@ -33,7 +41,7 @@ def kvstore_dependencies():
 
 class InmemoryKVStoreImpl(KVStore):
     def __init__(self):
-        self._store = {}
+        self._store: dict[str, str] = {}
 
     async def initialize(self) -> None:
         pass
@@ -41,7 +49,7 @@ class InmemoryKVStoreImpl(KVStore):
     async def get(self, key: str) -> str | None:
         return self._store.get(key)
 
-    async def set(self, key: str, value: str) -> None:
+    async def set(self, key: str, value: str, expiration: datetime | None = None) -> None:
         self._store[key] = value
 
     async def values_in_range(self, start_key: str, end_key: str) -> list[str]:
@@ -70,7 +78,8 @@ def register_kvstore_backends(backends: dict[str, StorageBackendConfig]) -> None
     _KVSTORE_INSTANCES.clear()
     _KVSTORE_LOCKS.clear()
     for name, cfg in backends.items():
-        _KVSTORE_BACKENDS[name] = cfg
+        typed_cfg = cast(KVStoreConfig, cfg)
+        _KVSTORE_BACKENDS[name] = typed_cfg
 
 
 async def kvstore_impl(reference: KVStoreReference) -> KVStore:
@@ -94,19 +103,20 @@ async def kvstore_impl(reference: KVStoreReference) -> KVStore:
         config = backend_config.model_copy()
         config.namespace = reference.namespace
 
-        if config.type == StorageBackendType.KV_REDIS.value:
+        impl: KVStore
+        if isinstance(config, RedisKVStoreConfig):
             from .redis import RedisKVStoreImpl
 
             impl = RedisKVStoreImpl(config)
-        elif config.type == StorageBackendType.KV_SQLITE.value:
+        elif isinstance(config, SqliteKVStoreConfig):
             from .sqlite import SqliteKVStoreImpl
 
             impl = SqliteKVStoreImpl(config)
-        elif config.type == StorageBackendType.KV_POSTGRES.value:
+        elif isinstance(config, PostgresKVStoreConfig):
             from .postgres import PostgresKVStoreImpl
 
             impl = PostgresKVStoreImpl(config)
-        elif config.type == StorageBackendType.KV_MONGODB.value:
+        elif isinstance(config, MongoDBKVStoreConfig):
             from .mongodb import MongoDBKVStoreImpl
 
             impl = MongoDBKVStoreImpl(config)

src/llama_stack/core/storage/kvstore/mongodb/mongodb.py

@@ -9,8 +9,8 @@ from datetime import datetime
 from pymongo import AsyncMongoClient
 from pymongo.asynchronous.collection import AsyncCollection
 
-from llama_stack.log import get_logger
 from llama_stack.core.storage.kvstore import KVStore
+from llama_stack.log import get_logger
 
 from ..config import MongoDBKVStoreConfig
 

src/llama_stack/core/storage/kvstore/postgres/postgres.py

@@ -7,11 +7,12 @@
 from datetime import datetime
 
 import psycopg2
+from psycopg2.extensions import connection as PGConnection
 from psycopg2.extras import DictCursor
 
 from llama_stack.log import get_logger
+from llama_stack_api.internal.kvstore import KVStore
 
-from ..api import KVStore
 from ..config import PostgresKVStoreConfig
 
 log = get_logger(name=__name__, category="providers::utils")
@@ -20,12 +21,12 @@ log = get_logger(name=__name__, category="providers::utils")
 class PostgresKVStoreImpl(KVStore):
     def __init__(self, config: PostgresKVStoreConfig):
         self.config = config
-        self.conn = None
-        self.cursor = None
+        self._conn: PGConnection | None = None
+        self._cursor: DictCursor | None = None
 
     async def initialize(self) -> None:
         try:
-            self.conn = psycopg2.connect(
+            self._conn = psycopg2.connect(
                 host=self.config.host,
                 port=self.config.port,
                 database=self.config.db,
@@ -34,11 +35,11 @@ class PostgresKVStoreImpl(KVStore):
                 sslmode=self.config.ssl_mode,
                 sslrootcert=self.config.ca_cert_path,
             )
-            self.conn.autocommit = True
-            self.cursor = self.conn.cursor(cursor_factory=DictCursor)
+            self._conn.autocommit = True
+            self._cursor = self._conn.cursor(cursor_factory=DictCursor)
 
             # Create table if it doesn't exist
-            self.cursor.execute(
+            self._cursor.execute(
                 f"""
                 CREATE TABLE IF NOT EXISTS {self.config.table_name} (
                     key TEXT PRIMARY KEY,
@@ -51,6 +52,11 @@ class PostgresKVStoreImpl(KVStore):
             log.exception("Could not connect to PostgreSQL database server")
             raise RuntimeError("Could not connect to PostgreSQL database server") from e
 
+    def _cursor_or_raise(self) -> DictCursor:
+        if self._cursor is None:
+            raise RuntimeError("Postgres client not initialized")
+        return self._cursor
+
     def _namespaced_key(self, key: str) -> str:
         if not self.config.namespace:
             return key
@@ -58,7 +64,8 @@ class PostgresKVStoreImpl(KVStore):
 
     async def set(self, key: str, value: str, expiration: datetime | None = None) -> None:
         key = self._namespaced_key(key)
-        self.cursor.execute(
+        cursor = self._cursor_or_raise()
+        cursor.execute(
             f"""
             INSERT INTO {self.config.table_name} (key, value, expiration)
             VALUES (%s, %s, %s)
@@ -70,7 +77,8 @@ class PostgresKVStoreImpl(KVStore):
 
     async def get(self, key: str) -> str | None:
         key = self._namespaced_key(key)
-        self.cursor.execute(
+        cursor = self._cursor_or_raise()
+        cursor.execute(
             f"""
             SELECT value FROM {self.config.table_name}
             WHERE key = %s
@@ -78,12 +86,13 @@ class PostgresKVStoreImpl(KVStore):
             """,
             (key,),
         )
-        result = self.cursor.fetchone()
+        result = cursor.fetchone()
         return result[0] if result else None
 
     async def delete(self, key: str) -> None:
         key = self._namespaced_key(key)
-        self.cursor.execute(
+        cursor = self._cursor_or_raise()
+        cursor.execute(
             f"DELETE FROM {self.config.table_name} WHERE key = %s",
             (key,),
         )
@@ -92,7 +101,8 @@ class PostgresKVStoreImpl(KVStore):
         start_key = self._namespaced_key(start_key)
         end_key = self._namespaced_key(end_key)
 
-        self.cursor.execute(
+        cursor = self._cursor_or_raise()
+        cursor.execute(
             f"""
             SELECT value FROM {self.config.table_name}
             WHERE key >= %s AND key < %s
@@ -101,14 +111,15 @@ class PostgresKVStoreImpl(KVStore):
             """,
             (start_key, end_key),
         )
-        return [row[0] for row in self.cursor.fetchall()]
+        return [row[0] for row in cursor.fetchall()]
 
     async def keys_in_range(self, start_key: str, end_key: str) -> list[str]:
         start_key = self._namespaced_key(start_key)
         end_key = self._namespaced_key(end_key)
 
-        self.cursor.execute(
+        cursor = self._cursor_or_raise()
+        cursor.execute(
             f"SELECT key FROM {self.config.table_name} WHERE key >= %s AND key < %s",
             (start_key, end_key),
         )
-        return [row[0] for row in self.cursor.fetchall()]
+        return [row[0] for row in cursor.fetchall()]

src/llama_stack/core/storage/kvstore/redis/redis.py

@@ -8,16 +8,23 @@ from datetime import datetime
 
 from redis.asyncio import Redis
 
-from ..api import KVStore
+from llama_stack_api.internal.kvstore import KVStore
+
 from ..config import RedisKVStoreConfig
 
 
 class RedisKVStoreImpl(KVStore):
     def __init__(self, config: RedisKVStoreConfig):
         self.config = config
+        self._redis: Redis | None = None
 
     async def initialize(self) -> None:
-        self.redis = Redis.from_url(self.config.url)
+        self._redis = Redis.from_url(self.config.url)
+
+    def _client(self) -> Redis:
+        if self._redis is None:
+            raise RuntimeError("Redis client not initialized")
+        return self._redis
 
     def _namespaced_key(self, key: str) -> str:
         if not self.config.namespace:
@@ -26,30 +33,37 @@ class RedisKVStoreImpl(KVStore):
 
     async def set(self, key: str, value: str, expiration: datetime | None = None) -> None:
         key = self._namespaced_key(key)
-        await self.redis.set(key, value)
+        client = self._client()
+        await client.set(key, value)
         if expiration:
-            await self.redis.expireat(key, expiration)
+            await client.expireat(key, expiration)
 
     async def get(self, key: str) -> str | None:
         key = self._namespaced_key(key)
-        value = await self.redis.get(key)
+        client = self._client()
+        value = await client.get(key)
         if value is None:
             return None
-        await self.redis.ttl(key)
+        await client.ttl(key)
+        if isinstance(value, bytes):
+            return value.decode("utf-8")
+        if isinstance(value, str):
             return value
+        return str(value)
 
     async def delete(self, key: str) -> None:
         key = self._namespaced_key(key)
-        await self.redis.delete(key)
+        await self._client().delete(key)
 
     async def values_in_range(self, start_key: str, end_key: str) -> list[str]:
         start_key = self._namespaced_key(start_key)
         end_key = self._namespaced_key(end_key)
+        client = self._client()
         cursor = 0
         pattern = start_key + "*"  # Match all keys starting with start_key prefix
-        matching_keys = []
+        matching_keys: list[str | bytes] = []
         while True:
-            cursor, keys = await self.redis.scan(cursor, match=pattern, count=1000)
+            cursor, keys = await client.scan(cursor, match=pattern, count=1000)
 
             for key in keys:
                 key_str = key.decode("utf-8") if isinstance(key, bytes) else key
@@ -61,7 +75,7 @@ class RedisKVStoreImpl(KVStore):
 
         # Then fetch all values in a single MGET call
         if matching_keys:
-            values = await self.redis.mget(matching_keys)
+            values = await client.mget(matching_keys)
             return [
                 value.decode("utf-8") if isinstance(value, bytes) else value for value in values if value is not None
             ]
@@ -70,7 +84,18 @@ class RedisKVStoreImpl(KVStore):
 
     async def keys_in_range(self, start_key: str, end_key: str) -> list[str]:
         """Get all keys in the given range."""
-        matching_keys = await self.redis.zrangebylex(self.namespace, f"[{start_key}", f"[{end_key}")
-        if not matching_keys:
-            return []
-        return [k.decode("utf-8") for k in matching_keys]
+        start_key = self._namespaced_key(start_key)
+        end_key = self._namespaced_key(end_key)
+        client = self._client()
+        cursor = 0
+        pattern = start_key + "*"
+        result: list[str] = []
+        while True:
+            cursor, keys = await client.scan(cursor, match=pattern, count=1000)
+            for key in keys:
+                key_str = key.decode("utf-8") if isinstance(key, bytes) else str(key)
+                if start_key <= key_str <= end_key:
+                    result.append(key_str)
+            if cursor == 0:
+                break
+        return result

src/llama_stack/core/storage/kvstore/sqlite/sqlite.py

@@ -10,8 +10,8 @@ from datetime import datetime
 import aiosqlite
 
 from llama_stack.log import get_logger
+from llama_stack_api.internal.kvstore import KVStore
 
-from ..api import KVStore
 from ..config import SqliteKVStoreConfig
 
 logger = get_logger(name=__name__, category="providers::utils")

src/llama_stack/core/storage/sqlstore/__init__.py

@@ -4,4 +4,14 @@
 # This source code is licensed under the terms described in the LICENSE file in
 # the root directory of this source tree.
 
+from llama_stack_api.internal.sqlstore import (
+    ColumnDefinition as ColumnDefinition,
+)
+from llama_stack_api.internal.sqlstore import (
+    ColumnType as ColumnType,
+)
+from llama_stack_api.internal.sqlstore import (
+    SqlStore as SqlStore,
+)
+
 from .sqlstore import *  # noqa: F401,F403

src/llama_stack/core/storage/sqlstore/api.py

@@ -1,10 +0,0 @@
-# Copyright (c) Meta Platforms, Inc. and affiliates.
-# All rights reserved.
-#
-# This source code is licensed under the terms described in the LICENSE file in
-# the root directory of this source tree.
-
-from llama_stack_api import PaginatedResponse
-from llama_stack_api.internal.sqlstore import ColumnDefinition, ColumnType, SqlStore
-
-__all__ = ["ColumnDefinition", "ColumnType", "SqlStore", "PaginatedResponse"]

src/llama_stack/core/storage/sqlstore/authorized_sqlstore.py

@@ -14,8 +14,8 @@ from llama_stack.core.datatypes import User
 from llama_stack.core.request_headers import get_authenticated_user
 from llama_stack.core.storage.datatypes import StorageBackendType
 from llama_stack.log import get_logger
-
-from .api import ColumnDefinition, ColumnType, PaginatedResponse, SqlStore
+from llama_stack_api import PaginatedResponse
+from llama_stack_api.internal.sqlstore import ColumnDefinition, ColumnType, SqlStore
 
 logger = get_logger(name=__name__, category="providers::utils")
 

src/llama_stack/core/storage/sqlstore/sqlalchemy_sqlstore.py

@@ -29,8 +29,7 @@ from sqlalchemy.sql.elements import ColumnElement
 from llama_stack.core.storage.datatypes import SqlAlchemySqlStoreConfig
 from llama_stack.log import get_logger
 from llama_stack_api import PaginatedResponse
-
-from .api import ColumnDefinition, ColumnType, SqlStore
+from llama_stack_api.internal.sqlstore import ColumnDefinition, ColumnType, SqlStore
 
 logger = get_logger(name=__name__, category="providers::utils")
 

src/llama_stack/core/storage/sqlstore/sqlstore.py

@@ -16,8 +16,7 @@ from llama_stack.core.storage.datatypes import (
     StorageBackendConfig,
     StorageBackendType,
 )
-
-from .api import SqlStore
+from llama_stack_api.internal.sqlstore import SqlStore
 
 sql_store_pip_packages = ["sqlalchemy[asyncio]", "aiosqlite", "asyncpg"]
 

src/llama_stack/core/store/registry.py

@@ -12,8 +12,8 @@ import pydantic
 
 from llama_stack.core.datatypes import RoutableObjectWithProvider
 from llama_stack.core.storage.datatypes import KVStoreReference
-from llama_stack.log import get_logger
 from llama_stack.core.storage.kvstore import KVStore, kvstore_impl
+from llama_stack.log import get_logger
 
 logger = get_logger(__name__, category="core::registry")
 

src/llama_stack/distributions/starter/starter.py

@@ -17,6 +17,8 @@ from llama_stack.core.datatypes import (
     ToolGroupInput,
     VectorStoresConfig,
 )
+from llama_stack.core.storage.kvstore.config import PostgresKVStoreConfig
+from llama_stack.core.storage.sqlstore.sqlstore import PostgresSqlStoreConfig
 from llama_stack.core.utils.dynamic import instantiate_class_type
 from llama_stack.distributions.template import DistributionTemplate, RunConfigSettings
 from llama_stack.providers.inline.files.localfs.config import LocalfsFilesImplConfig
@@ -35,8 +37,6 @@ from llama_stack.providers.remote.vector_io.pgvector.config import (
 )
 from llama_stack.providers.remote.vector_io.qdrant.config import QdrantVectorIOConfig
 from llama_stack.providers.remote.vector_io.weaviate.config import WeaviateVectorIOConfig
-from llama_stack.core.storage.kvstore.config import PostgresKVStoreConfig
-from llama_stack.core.storage.sqlstore.sqlstore import PostgresSqlStoreConfig
 from llama_stack_api import RemoteProviderSpec
 
 

src/llama_stack/distributions/template.py

@@ -35,13 +35,13 @@ from llama_stack.core.storage.datatypes import (
     SqlStoreReference,
     StorageBackendType,
 )
-from llama_stack.core.utils.dynamic import instantiate_class_type
-from llama_stack.core.utils.image_types import LlamaStackImageType
-from llama_stack.providers.utils.inference.model_registry import ProviderModelEntry
 from llama_stack.core.storage.kvstore.config import SqliteKVStoreConfig
 from llama_stack.core.storage.kvstore.config import get_pip_packages as get_kv_pip_packages
 from llama_stack.core.storage.sqlstore.sqlstore import SqliteSqlStoreConfig
 from llama_stack.core.storage.sqlstore.sqlstore import get_pip_packages as get_sql_pip_packages
+from llama_stack.core.utils.dynamic import instantiate_class_type
+from llama_stack.core.utils.image_types import LlamaStackImageType
+from llama_stack.providers.utils.inference.model_registry import ProviderModelEntry
 from llama_stack_api import DatasetPurpose, ModelType
 
 

src/llama_stack/providers/inline/agents/meta_reference/agents.py

@@ -6,8 +6,8 @@
 
 
 from llama_stack.core.datatypes import AccessRule
-from llama_stack.log import get_logger
 from llama_stack.core.storage.kvstore import InmemoryKVStoreImpl, kvstore_impl
+from llama_stack.log import get_logger
 from llama_stack.providers.utils.responses.responses_store import ResponsesStore
 from llama_stack_api import (
     Agents,

src/llama_stack/providers/inline/batches/reference/batches.py

@@ -16,8 +16,8 @@ from typing import Any, Literal
 from openai.types.batch import BatchError, Errors
 from pydantic import BaseModel
 
-from llama_stack.log import get_logger
 from llama_stack.core.storage.kvstore import KVStore
+from llama_stack.log import get_logger
 from llama_stack_api import (
     Batches,
     BatchObject,

src/llama_stack/providers/inline/datasetio/localfs/datasetio.py

@@ -5,8 +5,8 @@
 # the root directory of this source tree.
 from typing import Any
 
-from llama_stack.providers.utils.datasetio.url_utils import get_dataframe_from_uri
 from llama_stack.core.storage.kvstore import kvstore_impl
+from llama_stack.providers.utils.datasetio.url_utils import get_dataframe_from_uri
 from llama_stack.providers.utils.pagination import paginate_records
 from llama_stack_api import Dataset, DatasetIO, DatasetsProtocolPrivate, PaginatedResponse
 

src/llama_stack/providers/inline/eval/meta_reference/eval.py

@@ -8,8 +8,8 @@ from typing import Any
 
 from tqdm import tqdm
 
-from llama_stack.providers.utils.common.data_schema_validator import ColumnName
 from llama_stack.core.storage.kvstore import kvstore_impl
+from llama_stack.providers.utils.common.data_schema_validator import ColumnName
 from llama_stack_api import (
     Agents,
     Benchmark,

src/llama_stack/providers/inline/files/localfs/files.py

@@ -13,10 +13,10 @@ from fastapi import Depends, File, Form, Response, UploadFile
 
 from llama_stack.core.datatypes import AccessRule
 from llama_stack.core.id_generation import generate_object_id
-from llama_stack.log import get_logger
-from llama_stack.providers.utils.files.form_data import parse_expires_after
 from llama_stack.core.storage.sqlstore.authorized_sqlstore import AuthorizedSqlStore
 from llama_stack.core.storage.sqlstore.sqlstore import sqlstore_impl
+from llama_stack.log import get_logger
+from llama_stack.providers.utils.files.form_data import parse_expires_after
 from llama_stack_api import (
     ExpiresAfter,
     Files,

src/llama_stack/providers/inline/vector_io/faiss/faiss.py

@@ -14,8 +14,8 @@ import faiss  # type: ignore[import-untyped]
 import numpy as np
 from numpy.typing import NDArray
 
-from llama_stack.log import get_logger
 from llama_stack.core.storage.kvstore import kvstore_impl
+from llama_stack.log import get_logger
 from llama_stack.providers.utils.memory.openai_vector_store_mixin import OpenAIVectorStoreMixin
 from llama_stack.providers.utils.memory.vector_store import ChunkForDeletion, EmbeddingIndex, VectorStoreWithIndex
 from llama_stack_api import (

src/llama_stack/providers/inline/vector_io/sqlite_vec/sqlite_vec.py

@@ -14,8 +14,8 @@ import numpy as np
 import sqlite_vec  # type: ignore[import-untyped]
 from numpy.typing import NDArray
 
-from llama_stack.log import get_logger
 from llama_stack.core.storage.kvstore import kvstore_impl
+from llama_stack.log import get_logger
 from llama_stack.providers.utils.memory.openai_vector_store_mixin import OpenAIVectorStoreMixin
 from llama_stack.providers.utils.memory.vector_store import (
     RERANKER_TYPE_RRF,

src/llama_stack/providers/remote/files/openai/files.py

@@ -10,9 +10,9 @@ from typing import Annotated, Any
 from fastapi import Depends, File, Form, Response, UploadFile
 
 from llama_stack.core.datatypes import AccessRule
-from llama_stack.providers.utils.files.form_data import parse_expires_after
 from llama_stack.core.storage.sqlstore.authorized_sqlstore import AuthorizedSqlStore
 from llama_stack.core.storage.sqlstore.sqlstore import sqlstore_impl
+from llama_stack.providers.utils.files.form_data import parse_expires_after
 from llama_stack_api import (
     ExpiresAfter,
     Files,

src/llama_stack/providers/remote/files/s3/files.py

@@ -19,9 +19,9 @@ if TYPE_CHECKING:
 
 from llama_stack.core.datatypes import AccessRule
 from llama_stack.core.id_generation import generate_object_id
-from llama_stack.providers.utils.files.form_data import parse_expires_after
 from llama_stack.core.storage.sqlstore.authorized_sqlstore import AuthorizedSqlStore
 from llama_stack.core.storage.sqlstore.sqlstore import sqlstore_impl
+from llama_stack.providers.utils.files.form_data import parse_expires_after
 from llama_stack_api import (
     ExpiresAfter,
     Files,

src/llama_stack/providers/remote/vector_io/chroma/chroma.py

@@ -11,9 +11,9 @@ from urllib.parse import urlparse
 import chromadb
 from numpy.typing import NDArray
 
+from llama_stack.core.storage.kvstore import kvstore_impl
 from llama_stack.log import get_logger
 from llama_stack.providers.inline.vector_io.chroma import ChromaVectorIOConfig as InlineChromaVectorIOConfig
-from llama_stack.core.storage.kvstore import kvstore_impl
 from llama_stack.providers.utils.memory.openai_vector_store_mixin import OpenAIVectorStoreMixin
 from llama_stack.providers.utils.memory.vector_store import ChunkForDeletion, EmbeddingIndex, VectorStoreWithIndex
 from llama_stack_api import (

src/llama_stack/providers/remote/vector_io/milvus/milvus.py

@@ -11,9 +11,9 @@ from typing import Any
 from numpy.typing import NDArray
 from pymilvus import AnnSearchRequest, DataType, Function, FunctionType, MilvusClient, RRFRanker, WeightedRanker
 
+from llama_stack.core.storage.kvstore import kvstore_impl
 from llama_stack.log import get_logger
 from llama_stack.providers.inline.vector_io.milvus import MilvusVectorIOConfig as InlineMilvusVectorIOConfig
-from llama_stack.core.storage.kvstore import kvstore_impl
 from llama_stack.providers.utils.memory.openai_vector_store_mixin import OpenAIVectorStoreMixin
 from llama_stack.providers.utils.memory.vector_store import (
     RERANKER_TYPE_WEIGHTED,

src/llama_stack/providers/remote/vector_io/pgvector/pgvector.py

@@ -13,9 +13,9 @@ from psycopg2 import sql
 from psycopg2.extras import Json, execute_values
 from pydantic import BaseModel, TypeAdapter
 
+from llama_stack.core.storage.kvstore import kvstore_impl
 from llama_stack.log import get_logger
 from llama_stack.providers.utils.inference.prompt_adapter import interleaved_content_as_str
-from llama_stack.core.storage.kvstore import kvstore_impl
 from llama_stack.providers.utils.memory.openai_vector_store_mixin import OpenAIVectorStoreMixin
 from llama_stack.providers.utils.memory.vector_store import ChunkForDeletion, EmbeddingIndex, VectorStoreWithIndex
 from llama_stack.providers.utils.vector_io.vector_utils import WeightedInMemoryAggregator, sanitize_collection_name

src/llama_stack/providers/remote/vector_io/qdrant/qdrant.py

@@ -13,9 +13,9 @@ from numpy.typing import NDArray
 from qdrant_client import AsyncQdrantClient, models
 from qdrant_client.models import PointStruct
 
+from llama_stack.core.storage.kvstore import kvstore_impl
 from llama_stack.log import get_logger
 from llama_stack.providers.inline.vector_io.qdrant import QdrantVectorIOConfig as InlineQdrantVectorIOConfig
-from llama_stack.core.storage.kvstore import kvstore_impl
 from llama_stack.providers.utils.memory.openai_vector_store_mixin import OpenAIVectorStoreMixin
 from llama_stack.providers.utils.memory.vector_store import ChunkForDeletion, EmbeddingIndex, VectorStoreWithIndex
 from llama_stack_api import (

src/llama_stack/providers/remote/vector_io/weaviate/weaviate.py

@@ -13,8 +13,8 @@ from weaviate.classes.init import Auth
 from weaviate.classes.query import Filter, HybridFusion
 
 from llama_stack.core.request_headers import NeedsRequestProviderData
-from llama_stack.log import get_logger
 from llama_stack.core.storage.kvstore import kvstore_impl
+from llama_stack.log import get_logger
 from llama_stack.providers.utils.memory.openai_vector_store_mixin import OpenAIVectorStoreMixin
 from llama_stack.providers.utils.memory.vector_store import (
     RERANKER_TYPE_RRF,

src/llama_stack/providers/utils/inference/inference_store.py

@@ -10,6 +10,8 @@ from sqlalchemy.exc import IntegrityError
 
 from llama_stack.core.datatypes import AccessRule
 from llama_stack.core.storage.datatypes import InferenceStoreReference, StorageBackendType
+from llama_stack.core.storage.sqlstore.authorized_sqlstore import AuthorizedSqlStore
+from llama_stack.core.storage.sqlstore.sqlstore import _SQLSTORE_BACKENDS, sqlstore_impl
 from llama_stack.log import get_logger
 from llama_stack_api import (
     ListOpenAIChatCompletionResponse,
@@ -18,10 +20,7 @@ from llama_stack_api import (
     OpenAIMessageParam,
     Order,
 )
-
-from llama_stack.core.storage.sqlstore.api import ColumnDefinition, ColumnType
-from llama_stack.core.storage.sqlstore.authorized_sqlstore import AuthorizedSqlStore
-from llama_stack.core.storage.sqlstore.sqlstore import _SQLSTORE_BACKENDS, sqlstore_impl
+from llama_stack_api.internal.sqlstore import ColumnDefinition, ColumnType
 
 logger = get_logger(name=__name__, category="inference")
 

src/llama_stack/providers/utils/responses/responses_store.py

@@ -6,6 +6,8 @@
 
 from llama_stack.core.datatypes import AccessRule
 from llama_stack.core.storage.datatypes import ResponsesStoreReference, SqlStoreReference
+from llama_stack.core.storage.sqlstore.authorized_sqlstore import AuthorizedSqlStore
+from llama_stack.core.storage.sqlstore.sqlstore import sqlstore_impl
 from llama_stack.log import get_logger
 from llama_stack_api import (
     ListOpenAIResponseInputItem,
@@ -17,10 +19,7 @@ from llama_stack_api import (
     OpenAIResponseObjectWithInput,
     Order,
 )
-
-from llama_stack.core.storage.sqlstore.api import ColumnDefinition, ColumnType
-from llama_stack.core.storage.sqlstore.authorized_sqlstore import AuthorizedSqlStore
-from llama_stack.core.storage.sqlstore.sqlstore import sqlstore_impl
+from llama_stack_api.internal.sqlstore import ColumnDefinition, ColumnType
 
 logger = get_logger(name=__name__, category="openai_responses")
 

src/llama_stack_api/internal/__init__.py

@@ -1,3 +1,9 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
 # Internal subpackage for shared interfaces that are not part of the public API.
 
 __all__: list[str] = []

tests/unit/files/test_files.py

@@ -9,11 +9,11 @@ import pytest
 
 from llama_stack.core.access_control.access_control import default_policy
 from llama_stack.core.storage.datatypes import SqliteSqlStoreConfig, SqlStoreReference
+from llama_stack.core.storage.sqlstore.sqlstore import register_sqlstore_backends
 from llama_stack.providers.inline.files.localfs import (
     LocalfsFilesImpl,
     LocalfsFilesImplConfig,
 )
-from llama_stack.core.storage.sqlstore.sqlstore import register_sqlstore_backends
 from llama_stack_api import OpenAIFilePurpose, Order, ResourceNotFoundError
 
 

tests/unit/fixtures.py

@@ -6,9 +6,9 @@
 
 import pytest
 
-from llama_stack.core.store.registry import CachedDiskDistributionRegistry, DiskDistributionRegistry
 from llama_stack.core.storage.kvstore.config import SqliteKVStoreConfig
 from llama_stack.core.storage.kvstore.sqlite import SqliteKVStoreImpl
+from llama_stack.core.store.registry import CachedDiskDistributionRegistry, DiskDistributionRegistry
 
 
 @pytest.fixture(scope="function")

tests/unit/providers/agents/meta_reference/test_openai_responses.py

@@ -17,6 +17,7 @@ from openai.types.chat.chat_completion_chunk import (
 
 from llama_stack.core.access_control.access_control import default_policy
 from llama_stack.core.storage.datatypes import ResponsesStoreReference, SqliteSqlStoreConfig
+from llama_stack.core.storage.sqlstore.sqlstore import register_sqlstore_backends
 from llama_stack.providers.inline.agents.meta_reference.responses.openai_responses import (
     OpenAIResponsesImpl,
 )
@@ -24,7 +25,6 @@ from llama_stack.providers.utils.responses.responses_store import (
     ResponsesStore,
     _OpenAIResponseObjectWithInputAndMessages,
 )
-from llama_stack.core.storage.sqlstore.sqlstore import register_sqlstore_backends
 from llama_stack_api.agents import Order
 from llama_stack_api.inference import (
     OpenAIAssistantMessageParam,

tests/unit/providers/batches/conftest.py

@@ -13,9 +13,9 @@ from unittest.mock import AsyncMock
 import pytest
 
 from llama_stack.core.storage.datatypes import KVStoreReference, SqliteKVStoreConfig
+from llama_stack.core.storage.kvstore import kvstore_impl, register_kvstore_backends
 from llama_stack.providers.inline.batches.reference.batches import ReferenceBatchesImpl
 from llama_stack.providers.inline.batches.reference.config import ReferenceBatchesImplConfig
-from llama_stack.core.storage.kvstore import kvstore_impl, register_kvstore_backends
 
 
 @pytest.fixture

tests/unit/providers/files/conftest.py

@@ -9,8 +9,8 @@ import pytest
 from moto import mock_aws
 
 from llama_stack.core.storage.datatypes import SqliteSqlStoreConfig, SqlStoreReference
-from llama_stack.providers.remote.files.s3 import S3FilesImplConfig, get_adapter_impl
 from llama_stack.core.storage.sqlstore.sqlstore import register_sqlstore_backends
+from llama_stack.providers.remote.files.s3 import S3FilesImplConfig, get_adapter_impl
 
 
 class MockUploadFile:

tests/unit/providers/vector_io/conftest.py

@@ -11,13 +11,13 @@ import numpy as np
 import pytest
 
 from llama_stack.core.storage.datatypes import KVStoreReference, SqliteKVStoreConfig
+from llama_stack.core.storage.kvstore import register_kvstore_backends
 from llama_stack.providers.inline.vector_io.faiss.config import FaissVectorIOConfig
 from llama_stack.providers.inline.vector_io.faiss.faiss import FaissIndex, FaissVectorIOAdapter
 from llama_stack.providers.inline.vector_io.sqlite_vec import SQLiteVectorIOConfig
 from llama_stack.providers.inline.vector_io.sqlite_vec.sqlite_vec import SQLiteVecIndex, SQLiteVecVectorIOAdapter
 from llama_stack.providers.remote.vector_io.pgvector.config import PGVectorVectorIOConfig
 from llama_stack.providers.remote.vector_io.pgvector.pgvector import PGVectorIndex, PGVectorVectorIOAdapter
-from llama_stack.core.storage.kvstore import register_kvstore_backends
 from llama_stack_api import Chunk, ChunkMetadata, QueryChunksResponse, VectorStore
 
 EMBEDDING_DIMENSION = 768

tests/unit/registry/test_registry.py

@@ -9,12 +9,12 @@ import pytest
 
 from llama_stack.core.datatypes import VectorStoreWithOwner
 from llama_stack.core.storage.datatypes import KVStoreReference, SqliteKVStoreConfig
+from llama_stack.core.storage.kvstore import kvstore_impl, register_kvstore_backends
 from llama_stack.core.store.registry import (
     KEY_FORMAT,
     CachedDiskDistributionRegistry,
     DiskDistributionRegistry,
 )
-from llama_stack.core.storage.kvstore import kvstore_impl, register_kvstore_backends
 from llama_stack_api import Model, VectorStore
 
 

tests/unit/utils/inference/test_inference_store.py

@@ -9,8 +9,8 @@ import time
 import pytest
 
 from llama_stack.core.storage.datatypes import InferenceStoreReference, SqliteSqlStoreConfig
-from llama_stack.providers.utils.inference.inference_store import InferenceStore
 from llama_stack.core.storage.sqlstore.sqlstore import register_sqlstore_backends
+from llama_stack.providers.utils.inference.inference_store import InferenceStore
 from llama_stack_api import (
     OpenAIAssistantMessageParam,
     OpenAIChatCompletion,

tests/unit/utils/responses/test_responses_store.py

@@ -11,8 +11,8 @@ from uuid import uuid4
 import pytest
 
 from llama_stack.core.storage.datatypes import ResponsesStoreReference, SqliteSqlStoreConfig
-from llama_stack.providers.utils.responses.responses_store import ResponsesStore
 from llama_stack.core.storage.sqlstore.sqlstore import register_sqlstore_backends
+from llama_stack.providers.utils.responses.responses_store import ResponsesStore
 from llama_stack_api import OpenAIMessageParam, OpenAIResponseInput, OpenAIResponseObject, OpenAIUserMessageParam, Order