feat: rebase and implement file API methods

Signed-off-by: Varsha Prasad Narsing <varshaprasad96@gmail.com>

tests/integration/vector_io/test_openai_vector_stores.py

@@ -22,7 +22,14 @@ logger = logging.getLogger(__name__)
 def skip_if_provider_doesnt_support_openai_vector_stores(client_with_models):
     vector_io_providers = [p for p in client_with_models.providers.list() if p.api == "vector_io"]
     for p in vector_io_providers:
-        if p.provider_type in ["inline::faiss", "inline::sqlite-vec", "inline::milvus", "inline::qdrant", "inline::chromadb"]:
+        if p.provider_type in [
+            "inline::faiss",
+            "inline::sqlite-vec",
+            "inline::milvus",
+            "inline::qdrant",
+            "inline::chromadb",
+            "remote::qdrant",
+        ]:
             return
 
     pytest.skip("OpenAI vector stores are not supported by any provider")
@@ -31,7 +38,14 @@ def skip_if_provider_doesnt_support_openai_vector_stores(client_with_models):
 def skip_if_provider_doesnt_support_openai_vector_store_files_api(client_with_models):
     vector_io_providers = [p for p in client_with_models.providers.list() if p.api == "vector_io"]
     for p in vector_io_providers:
-        if p.provider_type in ["inline::faiss", "inline::sqlite-vec", "inline::milvus", "inline::qdrant", "remote::pgvector"]:
+        if p.provider_type in [
+            "inline::faiss",
+            "inline::sqlite-vec",
+            "inline::milvus",
+            "inline::qdrant",
+            "remote::pgvector",
+            "remote::qdrant",
+        ]:
             return
 
     pytest.skip("OpenAI vector stores are not supported by any provider")

tests/integration/vector_io/test_vector_io.py

@@ -125,6 +125,8 @@ def test_insert_chunks(client_with_empty_registry, embedding_model_id, embedding
 def test_insert_chunks_with_precomputed_embeddings(client_with_empty_registry, embedding_model_id, embedding_dimension):
     vector_io_provider_params_dict = {
         "inline::milvus": {"score_threshold": -1.0},
+        "remote::qdrant": {"score_threshold": -1.0},
+        "inline::qdrant": {"score_threshold": -1.0},
     }
     vector_db_id = "test_precomputed_embeddings_db"
     client_with_empty_registry.vector_dbs.register(
@@ -168,6 +170,8 @@ def test_query_returns_valid_object_when_identical_to_embedding_in_vdb(
 ):
     vector_io_provider_params_dict = {
         "inline::milvus": {"score_threshold": 0.0},
+        "remote::qdrant": {"score_threshold": 0.0},
+        "inline::qdrant": {"score_threshold": 0.0},
     }
     vector_db_id = "test_precomputed_embeddings_db"
     client_with_empty_registry.vector_dbs.register(

tests/unit/providers/vector_io/conftest.py

@@ -16,10 +16,12 @@ from llama_stack.providers.inline.vector_io.chroma.config import ChromaVectorIOC
 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.milvus.config import MilvusVectorIOConfig, SqliteKVStoreConfig
+from llama_stack.providers.inline.vector_io.qdrant import QdrantVectorIOConfig
 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.chroma.chroma import ChromaIndex, ChromaVectorIOAdapter
 from llama_stack.providers.remote.vector_io.milvus.milvus import MilvusIndex, MilvusVectorIOAdapter
+from llama_stack.providers.remote.vector_io.qdrant.qdrant import QdrantVectorIOAdapter
 
 EMBEDDING_DIMENSION = 384
 COLLECTION_PREFIX = "test_collection"
@@ -94,7 +96,7 @@ def sample_embeddings_with_metadata(sample_chunks_with_metadata):
     return np.array([np.random.rand(EMBEDDING_DIMENSION).astype(np.float32) for _ in sample_chunks_with_metadata])
 
 
-@pytest.fixture(params=["milvus", "sqlite_vec", "faiss"])
+@pytest.fixture(params=["milvus", "sqlite_vec", "faiss", "chroma", "qdrant"])
 def vector_provider(request):
     return request.param
 
@@ -133,7 +135,7 @@ async def sqlite_vec_vec_index(embedding_dimension, tmp_path_factory):
     await index.initialize()
     index.db_path = db_path
     yield index
-    index.delete()
+    await index.delete()
 
 
 @pytest.fixture
@@ -276,14 +278,66 @@ async def chroma_vec_adapter(chroma_vec_db_path, mock_inference_api, embedding_d
     await adapter.shutdown()
 
 
+@pytest.fixture
+def qdrant_vec_db_path(tmp_path_factory):
+    import uuid
+
+    db_path = str(tmp_path_factory.getbasetemp() / f"test_qdrant_{uuid.uuid4()}.db")
+    return db_path
+
+
+@pytest.fixture
+async def qdrant_vec_adapter(qdrant_vec_db_path, mock_inference_api, embedding_dimension):
+    import uuid
+
+    config = QdrantVectorIOConfig(
+        path=qdrant_vec_db_path,
+        kvstore=SqliteKVStoreConfig(),
+    )
+    adapter = QdrantVectorIOAdapter(
+        config=config,
+        inference_api=mock_inference_api,
+        files_api=None,
+    )
+    collection_id = f"qdrant_test_collection_{uuid.uuid4()}"
+    await adapter.initialize()
+    await adapter.register_vector_db(
+        VectorDB(
+            identifier=collection_id,
+            provider_id="test_provider",
+            embedding_model="test_model",
+            embedding_dimension=embedding_dimension,
+        )
+    )
+    adapter.test_collection_id = collection_id
+    yield adapter
+    await adapter.shutdown()
+
+
+@pytest.fixture
+async def qdrant_vec_index(qdrant_vec_db_path, embedding_dimension):
+    import uuid
+
+    from qdrant_client import AsyncQdrantClient
+
+    from llama_stack.providers.remote.vector_io.qdrant.qdrant import QdrantIndex
+
+    client = AsyncQdrantClient(path=qdrant_vec_db_path)
+    collection_name = f"qdrant_test_collection_{uuid.uuid4()}"
+    index = QdrantIndex(client, collection_name)
+    yield index
+    await index.delete()
+
+
 @pytest.fixture
 def vector_io_adapter(vector_provider, request):
     """Returns the appropriate vector IO adapter based on the provider parameter."""
     vector_provider_dict = {
         "milvus": "milvus_vec_adapter",
         "faiss": "faiss_vec_adapter",
-        "sqlite_vec": "sqlite_vec_adapter",
+        "qdrant": "qdrant_vec_adapter",
         "chroma": "chroma_vec_adapter",
+        "sqlite_vec": "sqlite_vec_adapter",
     }
     return request.getfixturevalue(vector_provider_dict[vector_provider])
 

tests/unit/providers/vector_io/test_qdrant.py

@@ -23,6 +23,7 @@ from llama_stack.providers.inline.vector_io.qdrant.config import (
 from llama_stack.providers.remote.vector_io.qdrant.qdrant import (
     QdrantVectorIOAdapter,
 )
+from llama_stack.providers.utils.kvstore.config import SqliteKVStoreConfig
 
 # This test is a unit test for the QdrantVectorIOAdapter class. This should only contain
 # tests which are specific to this class. More general (API-level) tests should be placed in
@@ -36,7 +37,9 @@ from llama_stack.providers.remote.vector_io.qdrant.qdrant import (
 
 @pytest.fixture
 def qdrant_config(tmp_path) -> InlineQdrantVectorIOConfig:
-    return InlineQdrantVectorIOConfig(path=os.path.join(tmp_path, "qdrant.db"))
+    kvstore_config = SqliteKVStoreConfig(db_name=os.path.join(tmp_path, "test_kvstore.db"))
+
+    return InlineQdrantVectorIOConfig(path=os.path.join(tmp_path, "qdrant.db"), kvstore=kvstore_config)
 
 
 @pytest.fixture(scope="session")
@@ -50,6 +53,10 @@ def mock_vector_db(vector_db_id) -> MagicMock:
     mock_vector_db.embedding_model = "embedding_model"
     mock_vector_db.identifier = vector_db_id
     mock_vector_db.embedding_dimension = 384
+    # Mock model_dump_json to return a proper JSON string for kvstore persistence
+    mock_vector_db.model_dump_json.return_value = (
+        '{"identifier": "' + vector_db_id + '", "embedding_model": "embedding_model", "embedding_dimension": 384}'
+    )
     return mock_vector_db
 
 

tests/unit/providers/vector_io/test_vector_io_openai_vector_stores.py

@@ -30,12 +30,12 @@ async def test_initialize_index(vector_index):
 
 
 async def test_add_chunks_query_vector(vector_index, sample_chunks, sample_embeddings):
-    vector_index.delete()
-    vector_index.initialize()
+    await vector_index.delete()
+    await vector_index.initialize()
     await vector_index.add_chunks(sample_chunks, sample_embeddings)
     resp = await vector_index.query_vector(sample_embeddings[0], k=1, score_threshold=-1)
     assert resp.chunks[0].content == sample_chunks[0].content
-    vector_index.delete()
+    await vector_index.delete()
 
 
 async def test_chunk_id_conflict(vector_index, sample_chunks, embedding_dimension):