chore: Migrating VectorDB to use OpenAI compatible identifier

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

docs/_static/llama-stack-spec.html

@@ -11285,6 +11285,9 @@
                     },
                     "embedding_dimension": {
                         "type": "integer"
+                    },
+                    "provider_vector_db_name": {
+                        "type": "string"
                     }
                 },
                 "additionalProperties": false,
@@ -13543,7 +13546,8 @@
                 },
                 "additionalProperties": false,
                 "required": [
-                    "name"
+                    "name",
+                    "provider_vector_db_id"
                 ],
                 "title": "OpenaiCreateVectorStoreRequest"
             },
@@ -15571,6 +15575,10 @@
                     "provider_vector_db_id": {
                         "type": "string",
                         "description": "The identifier of the vector database in the provider."
+                    },
+                    "provider_vector_db_name": {
+                        "type": "string",
+                        "description": "The name of the vector database."
                     }
                 },
                 "additionalProperties": false,

docs/_static/llama-stack-spec.yaml

@@ -7944,6 +7944,8 @@ components:
           type: string
         embedding_dimension:
           type: integer
+        provider_vector_db_name:
+          type: string
       additionalProperties: false
       required:
         - identifier
@@ -9461,6 +9463,7 @@ components:
       additionalProperties: false
       required:
         - name
+        - provider_vector_db_id
       title: OpenaiCreateVectorStoreRequest
     VectorStoreFileCounts:
       type: object
@@ -10897,6 +10900,9 @@ components:
           type: string
           description: >-
             The identifier of the vector database in the provider.
+        provider_vector_db_name:
+          type: string
+          description: The name of the vector database.
       additionalProperties: false
       required:
         - vector_db_id

llama_stack/apis/vector_dbs/vector_dbs.py

@@ -19,6 +19,7 @@ class VectorDB(Resource):
 
     embedding_model: str
     embedding_dimension: int
+    provider_vector_db_name: str | None = None
 
     @property
     def vector_db_id(self) -> str:
@@ -71,6 +72,7 @@ class VectorDBs(Protocol):
         embedding_dimension: int | None = 384,
         provider_id: str | None = None,
         provider_vector_db_id: str | None = None,
+        provider_vector_db_name: str | None = None,
     ) -> VectorDB:
         """Register a vector database.
 
@@ -79,6 +81,7 @@ class VectorDBs(Protocol):
         :param embedding_dimension: The dimension of the embedding model.
         :param provider_id: The identifier of the provider.
         :param provider_vector_db_id: The identifier of the vector database in the provider.
+        :param provider_vector_db_name: The name of the vector database.
         :returns: A VectorDB.
         """
         ...

llama_stack/apis/vector_io/vector_io.py

@@ -346,7 +346,7 @@ class VectorIO(Protocol):
         embedding_model: str | None = None,
         embedding_dimension: int | None = 384,
         provider_id: str | None = None,
-        provider_vector_db_id: str | None = None,
+        provider_vector_db_id: str = "",
     ) -> VectorStoreObject:
         """Creates a vector store.
 

llama_stack/distribution/routers/vector_io.py

@@ -5,6 +5,7 @@
 # the root directory of this source tree.
 
 import asyncio
+import uuid
 from typing import Any
 
 from llama_stack.apis.common.content_types import (
@@ -82,6 +83,7 @@ class VectorIORouter(VectorIO):
         embedding_dimension: int | None = 384,
         provider_id: str | None = None,
         provider_vector_db_id: str | None = None,
+        provider_vector_db_name: str | None = None,
     ) -> None:
         logger.debug(f"VectorIORouter.register_vector_db: {vector_db_id}, {embedding_model}")
         await self.routing_table.register_vector_db(
@@ -90,6 +92,7 @@ class VectorIORouter(VectorIO):
             embedding_dimension,
             provider_id,
             provider_vector_db_id,
+            provider_vector_db_name,
         )
 
     async def insert_chunks(
@@ -123,7 +126,7 @@ class VectorIORouter(VectorIO):
         embedding_model: str | None = None,
         embedding_dimension: int | None = None,
         provider_id: str | None = None,
-        provider_vector_db_id: str | None = None,
+        provider_vector_db_id: str = "",
     ) -> VectorStoreObject:
         logger.debug(f"VectorIORouter.openai_create_vector_store: name={name}, provider_id={provider_id}")
 
@@ -135,17 +138,17 @@ class VectorIORouter(VectorIO):
             embedding_model, embedding_dimension = embedding_model_info
             logger.info(f"No embedding model specified, using first available: {embedding_model}")
 
-        vector_db_id = name
+        vector_db_id = f"vs_{uuid.uuid4()}"
         registered_vector_db = await self.routing_table.register_vector_db(
             vector_db_id,
             embedding_model,
             embedding_dimension,
             provider_id,
             provider_vector_db_id,
+            name,
         )
-
         return await self.routing_table.get_provider_impl(registered_vector_db.identifier).openai_create_vector_store(
-            vector_db_id,
+            name,
             file_ids=file_ids,
             expires_after=expires_after,
             chunking_strategy=chunking_strategy,

llama_stack/distribution/routing_tables/vector_dbs.py

@@ -35,9 +35,10 @@ class VectorDBsRoutingTable(CommonRoutingTableImpl, VectorDBs):
         embedding_model: str,
         embedding_dimension: int | None = 384,
         provider_id: str | None = None,
-        provider_vector_db_id: str | None = None,
+        provider_vector_db_id: str = "",
+        provider_vector_db_name: str | None = None,
     ) -> VectorDB:
-        if provider_vector_db_id is None:
+        if provider_vector_db_id == "":
             provider_vector_db_id = vector_db_id
         if provider_id is None:
             if len(self.impls_by_provider_id) > 0:
@@ -62,6 +63,7 @@ class VectorDBsRoutingTable(CommonRoutingTableImpl, VectorDBs):
             "provider_resource_id": provider_vector_db_id,
             "embedding_model": embedding_model,
             "embedding_dimension": model.metadata["embedding_dimension"],
+            "vector_db_name": provider_vector_db_name,
         }
         vector_db = TypeAdapter(VectorDBWithOwner).validate_python(vector_db_data)
         await self.register_object(vector_db)

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

@@ -217,7 +217,7 @@ class ChromaVectorIOAdapter(VectorIO, VectorDBsProtocolPrivate):
         embedding_model: str | None = None,
         embedding_dimension: int | None = 384,
         provider_id: str | None = None,
-        provider_vector_db_id: str | None = None,
+        provider_vector_db_id: str = "",
     ) -> VectorStoreObject:
         raise NotImplementedError("OpenAI Vector Stores API is not supported in Chroma")
 

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

@@ -247,7 +247,7 @@ class PGVectorVectorIOAdapter(VectorIO, VectorDBsProtocolPrivate):
         embedding_model: str | None = None,
         embedding_dimension: int | None = 384,
         provider_id: str | None = None,
-        provider_vector_db_id: str | None = None,
+        provider_vector_db_id: str = "",
     ) -> VectorStoreObject:
         raise NotImplementedError("OpenAI Vector Stores API is not supported in PGVector")
 

llama_stack/providers/utils/memory/openai_vector_store_mixin.py

@@ -8,7 +8,6 @@ import asyncio
 import logging
 import mimetypes
 import time
-import uuid
 from abc import ABC, abstractmethod
 from typing import Any
 
@@ -144,11 +143,12 @@ class OpenAIVectorStoreMixin(ABC):
         embedding_model: str | None = None,
         embedding_dimension: int | None = 384,
         provider_id: str | None = None,
-        provider_vector_db_id: str | None = None,
+        provider_vector_db_id: str = "",
     ) -> VectorStoreObject:
         """Creates a vector store."""
-        store_id = name or str(uuid.uuid4())
         created_at = int(time.time())
+        if provider_vector_db_id is None:
+            raise ValueError("Provider vector DB ID is required")
 
         if provider_id is None:
             raise ValueError("Provider ID is required")
@@ -160,13 +160,13 @@ class OpenAIVectorStoreMixin(ABC):
         if embedding_dimension is None:
             raise ValueError("Embedding dimension is required")
 
-        provider_vector_db_id = provider_vector_db_id or store_id
         vector_db = VectorDB(
-            identifier=store_id,
+            identifier=provider_vector_db_id,
             embedding_dimension=embedding_dimension,
             embedding_model=embedding_model,
             provider_id=provider_id,
             provider_resource_id=provider_vector_db_id,
+            provider_vector_db_name=name,
         )
         # Register the vector DB
         await self.register_vector_db(vector_db)
@@ -182,11 +182,11 @@ class OpenAIVectorStoreMixin(ABC):
             in_progress=0,
             total=0,
         )
-        store_info = {
+        store_info: dict[str, Any] = {
-            "id": store_id,
+            "id": provider_vector_db_id,
             "object": "vector_store",
             "created_at": created_at,
-            "name": store_id,
+            "name": name,
             "usage_bytes": 0,
             "file_counts": file_counts.model_dump(),
             "status": status,
@@ -206,18 +206,18 @@ class OpenAIVectorStoreMixin(ABC):
         store_info["metadata"] = metadata
 
         # Save to persistent storage (provider-specific)
-        await self._save_openai_vector_store(store_id, store_info)
+        await self._save_openai_vector_store(provider_vector_db_id, store_info)
 
         # Store in memory cache
-        self.openai_vector_stores[store_id] = store_info
+        self.openai_vector_stores[provider_vector_db_id] = store_info
 
         # Now that our vector store is created, attach any files that were provided
         file_ids = file_ids or []
-        tasks = [self.openai_attach_file_to_vector_store(store_id, file_id) for file_id in file_ids]
+        tasks = [self.openai_attach_file_to_vector_store(provider_vector_db_id, file_id) for file_id in file_ids]
         await asyncio.gather(*tasks)
 
         # Get the updated store info and return it
-        store_info = self.openai_vector_stores[store_id]
+        store_info = self.openai_vector_stores[provider_vector_db_id]
         return VectorStoreObject.model_validate(store_info)
 
     async def openai_list_vector_stores(