feat: Enable setting a default embedding model in the stack (#3803)

# What does this PR do?

Enables automatic embedding model detection for vector stores and by
using a `default_configured` boolean that can be defined in the
`run.yaml`.

<!-- If resolving an issue, uncomment and update the line below -->
<!-- Closes #[issue-number] -->

## Test Plan
- Unit tests
- Integration tests
- Simple example below:

Spin up the stack:
```bash
uv run llama stack build --distro starter --image-type venv --run
```
Then test with OpenAI's client:
```python
from openai import OpenAI
client = OpenAI(base_url="http://localhost:8321/v1/", api_key="none")
vs = client.vector_stores.create()
```
Previously you needed:

```python
vs = client.vector_stores.create(
    extra_body={
        "embedding_model": "sentence-transformers/all-MiniLM-L6-v2",
        "embedding_dimension": 384,
    }
)
```

The `extra_body` is now unnecessary.

---------

Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>

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

@@ -12,6 +12,11 @@ from .config import ChromaVectorIOConfig
 async def get_adapter_impl(config: ChromaVectorIOConfig, deps: dict[Api, ProviderSpec]):
     from .chroma import ChromaVectorIOAdapter
 
-    impl = ChromaVectorIOAdapter(config, deps[Api.inference], deps.get(Api.files))
+    impl = ChromaVectorIOAdapter(
+        config,
+        deps[Api.inference],
+        deps[Api.models],
+        deps.get(Api.files),
+    )
     await impl.initialize()
     return impl

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

@@ -138,12 +138,14 @@ class ChromaVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtocolP
         self,
         config: RemoteChromaVectorIOConfig | InlineChromaVectorIOConfig,
         inference_api: Api.inference,
+        models_apis: Api.models,
         files_api: Files | None,
     ) -> None:
         super().__init__(files_api=files_api, kvstore=None)
         log.info(f"Initializing ChromaVectorIOAdapter with url: {config}")
         self.config = config
         self.inference_api = inference_api
+        self.models_api = models_apis
         self.client = None
         self.cache = {}
         self.vector_db_store = None

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

@@ -14,6 +14,11 @@ async def get_adapter_impl(config: MilvusVectorIOConfig, deps: dict[Api, Provide
 
     assert isinstance(config, MilvusVectorIOConfig), f"Unexpected config type: {type(config)}"
 
-    impl = MilvusVectorIOAdapter(config, deps[Api.inference], deps.get(Api.files, None))
+    impl = MilvusVectorIOAdapter(
+        config,
+        deps[Api.inference],
+        deps[Api.models],
+        deps.get(Api.files),
+    )
     await impl.initialize()
     return impl

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

@@ -12,8 +12,9 @@ from numpy.typing import NDArray
 from pymilvus import AnnSearchRequest, DataType, Function, FunctionType, MilvusClient, RRFRanker, WeightedRanker
 
 from llama_stack.apis.common.errors import VectorStoreNotFoundError
-from llama_stack.apis.files.files import Files
+from llama_stack.apis.files import Files
 from llama_stack.apis.inference import Inference, InterleavedContent
+from llama_stack.apis.models import Models
 from llama_stack.apis.vector_dbs import VectorDB
 from llama_stack.apis.vector_io import (
     Chunk,
@@ -307,6 +308,7 @@ class MilvusVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtocolP
         self,
         config: RemoteMilvusVectorIOConfig | InlineMilvusVectorIOConfig,
         inference_api: Inference,
+        models_api: Models,
         files_api: Files | None,
     ) -> None:
         super().__init__(files_api=files_api, kvstore=None)
@@ -314,6 +316,7 @@ class MilvusVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtocolP
         self.cache = {}
         self.client = None
         self.inference_api = inference_api
+        self.models_api = models_api
         self.vector_db_store = None
         self.metadata_collection_name = "openai_vector_stores_metadata"
 

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

@@ -12,6 +12,6 @@ from .config import PGVectorVectorIOConfig
 async def get_adapter_impl(config: PGVectorVectorIOConfig, deps: dict[Api, ProviderSpec]):
     from .pgvector import PGVectorVectorIOAdapter
 
-    impl = PGVectorVectorIOAdapter(config, deps[Api.inference], deps.get(Api.files, None))
+    impl = PGVectorVectorIOAdapter(config, deps[Api.inference], deps[Api.models], deps.get(Api.files, None))
     await impl.initialize()
     return impl

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

@@ -14,8 +14,9 @@ from psycopg2.extras import Json, execute_values
 from pydantic import BaseModel, TypeAdapter
 
 from llama_stack.apis.common.errors import VectorStoreNotFoundError
-from llama_stack.apis.files.files import Files
-from llama_stack.apis.inference import InterleavedContent
+from llama_stack.apis.files import Files
+from llama_stack.apis.inference import Inference, InterleavedContent
+from llama_stack.apis.models import Models
 from llama_stack.apis.vector_dbs import VectorDB
 from llama_stack.apis.vector_io import (
     Chunk,
@@ -23,7 +24,7 @@ from llama_stack.apis.vector_io import (
     VectorIO,
 )
 from llama_stack.log import get_logger
-from llama_stack.providers.datatypes import Api, VectorDBsProtocolPrivate
+from llama_stack.providers.datatypes import VectorDBsProtocolPrivate
 from llama_stack.providers.utils.inference.prompt_adapter import (
     interleaved_content_as_str,
 )
@@ -342,12 +343,14 @@ class PGVectorVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtoco
     def __init__(
         self,
         config: PGVectorVectorIOConfig,
-        inference_api: Api.inference,
+        inference_api: Inference,
+        models_api: Models,
         files_api: Files | None = None,
     ) -> None:
         super().__init__(files_api=files_api, kvstore=None)
         self.config = config
         self.inference_api = inference_api
+        self.models_api = models_api
         self.conn = None
         self.cache = {}
         self.vector_db_store = None

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

@@ -12,7 +12,11 @@ from .config import QdrantVectorIOConfig
 async def get_adapter_impl(config: QdrantVectorIOConfig, deps: dict[Api, ProviderSpec]):
     from .qdrant import QdrantVectorIOAdapter
 
-    files_api = deps.get(Api.files)
-    impl = QdrantVectorIOAdapter(config, deps[Api.inference], files_api)
+    impl = QdrantVectorIOAdapter(
+        config,
+        deps[Api.inference],
+        deps[Api.models],
+        deps.get(Api.files),
+    )
     await impl.initialize()
     return impl

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

@@ -15,7 +15,8 @@ from qdrant_client.models import PointStruct
 
 from llama_stack.apis.common.errors import VectorStoreNotFoundError
 from llama_stack.apis.files import Files
-from llama_stack.apis.inference import InterleavedContent
+from llama_stack.apis.inference import Inference, InterleavedContent
+from llama_stack.apis.models import Models
 from llama_stack.apis.vector_dbs import VectorDB
 from llama_stack.apis.vector_io import (
     Chunk,
@@ -25,7 +26,7 @@ from llama_stack.apis.vector_io import (
     VectorStoreFileObject,
 )
 from llama_stack.log import get_logger
-from llama_stack.providers.datatypes import Api, VectorDBsProtocolPrivate
+from llama_stack.providers.datatypes import VectorDBsProtocolPrivate
 from llama_stack.providers.inline.vector_io.qdrant import QdrantVectorIOConfig as InlineQdrantVectorIOConfig
 from llama_stack.providers.utils.kvstore import kvstore_impl
 from llama_stack.providers.utils.memory.openai_vector_store_mixin import OpenAIVectorStoreMixin
@@ -159,7 +160,8 @@ class QdrantVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtocolP
     def __init__(
         self,
         config: RemoteQdrantVectorIOConfig | InlineQdrantVectorIOConfig,
-        inference_api: Api.inference,
+        inference_api: Inference,
+        models_api: Models,
         files_api: Files | None = None,
     ) -> None:
         super().__init__(files_api=files_api, kvstore=None)
@@ -167,6 +169,7 @@ class QdrantVectorIOAdapter(OpenAIVectorStoreMixin, VectorIO, VectorDBsProtocolP
         self.client: AsyncQdrantClient = None
         self.cache = {}
         self.inference_api = inference_api
+        self.models_api = models_api
         self.vector_db_store = None
         self._qdrant_lock = asyncio.Lock()
 

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

@@ -12,6 +12,11 @@ from .config import WeaviateVectorIOConfig
 async def get_adapter_impl(config: WeaviateVectorIOConfig, deps: dict[Api, ProviderSpec]):
     from .weaviate import WeaviateVectorIOAdapter
 
-    impl = WeaviateVectorIOAdapter(config, deps[Api.inference], deps.get(Api.files, None))
+    impl = WeaviateVectorIOAdapter(
+        config,
+        deps[Api.inference],
+        deps[Api.models],
+        deps.get(Api.files),
+    )
     await impl.initialize()
     return impl

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

@@ -14,12 +14,14 @@ from weaviate.classes.query import Filter, HybridFusion
 
 from llama_stack.apis.common.content_types import InterleavedContent
 from llama_stack.apis.common.errors import VectorStoreNotFoundError
-from llama_stack.apis.files.files import Files
+from llama_stack.apis.files import Files
+from llama_stack.apis.inference import Inference
+from llama_stack.apis.models import Models
 from llama_stack.apis.vector_dbs import VectorDB
 from llama_stack.apis.vector_io import Chunk, QueryChunksResponse, VectorIO
 from llama_stack.core.request_headers import NeedsRequestProviderData
 from llama_stack.log import get_logger
-from llama_stack.providers.datatypes import Api, VectorDBsProtocolPrivate
+from llama_stack.providers.datatypes import VectorDBsProtocolPrivate
 from llama_stack.providers.utils.kvstore import kvstore_impl
 from llama_stack.providers.utils.kvstore.api import KVStore
 from llama_stack.providers.utils.memory.openai_vector_store_mixin import (
@@ -281,12 +283,14 @@ class WeaviateVectorIOAdapter(
     def __init__(
         self,
         config: WeaviateVectorIOConfig,
-        inference_api: Api.inference,
+        inference_api: Inference,
+        models_api: Models,
         files_api: Files | None,
     ) -> None:
         super().__init__(files_api=files_api, kvstore=None)
         self.config = config
         self.inference_api = inference_api
+        self.models_api = models_api
         self.client_cache = {}
         self.cache = {}
         self.vector_db_store = None