Pin weaviate-client version

Author: ChristianZaccaria

llama_stack/providers/registry/vector_io.py

@@ -503,7 +503,7 @@ def available_providers() -> list[ProviderSpec]:
             Api.vector_io,
             AdapterSpec(
                 adapter_type="weaviate",
-                pip_packages=["weaviate-client"],
+                pip_packages=["weaviate-client>=4.16.5"],
                 module="llama_stack.providers.remote.vector_io.weaviate",
                 config_class="llama_stack.providers.remote.vector_io.weaviate.WeaviateVectorIOConfig",
                 provider_data_validator="llama_stack.providers.remote.vector_io.weaviate.WeaviateRequestProviderData",

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

@@ -98,7 +98,7 @@ async def query_vector(self, embedding: NDArray, k: int, score_threshold: float)
         Returns:
             QueryChunksResponse with chunks and scores
         """
-        log.info(
+        log.debug(
             f"WEAVIATE VECTOR SEARCH CALLED: embedding_shape={embedding.shape}, k={k}, threshold={score_threshold}"
         )
         sanitized_collection_name = sanitize_collection_name(self.collection_name, weaviate_format=True)
@@ -135,7 +135,7 @@ async def query_vector(self, embedding: NDArray, k: int, score_threshold: float)
             chunks.append(chunk)
             scores.append(score)
 
-        log.info(f"WEAVIATE VECTOR SEARCH RESULTS: Found {len(chunks)} chunks with scores {scores}")
+        log.debug(f"WEAVIATE VECTOR SEARCH RESULTS: Found {len(chunks)} chunks with scores {scores}")
         return QueryChunksResponse(chunks=chunks, scores=scores)
 
     async def delete(self, chunk_ids: list[str] | None = None) -> None:
@@ -166,7 +166,7 @@ async def query_keyword(
         Returns:
             QueryChunksResponse with chunks and scores
         """
-        log.info(f"WEAVIATE KEYWORD SEARCH CALLED: query='{query_string}', k={k}, threshold={score_threshold}")
+        log.debug(f"WEAVIATE KEYWORD SEARCH CALLED: query='{query_string}', k={k}, threshold={score_threshold}")
         sanitized_collection_name = sanitize_collection_name(self.collection_name, weaviate_format=True)
         collection = self.client.collections.get(sanitized_collection_name)
 
@@ -199,7 +199,7 @@ async def query_keyword(
             chunks.append(chunk)
             scores.append(score)
 
-        log.info(f"WEAVIATE KEYWORD SEARCH RESULTS: Found {len(chunks)} chunks with scores {scores}.")
+        log.debug(f"WEAVIATE KEYWORD SEARCH RESULTS: Found {len(chunks)} chunks with scores {scores}.")
         return QueryChunksResponse(chunks=chunks, scores=scores)
 
     async def query_hybrid(
@@ -223,7 +223,7 @@ async def query_hybrid(
         Returns:
             QueryChunksResponse with combined results
         """
-        log.info(
+        log.debug(
             f"WEAVIATE HYBRID SEARCH CALLED: query='{query_string}', embedding_shape={embedding.shape}, k={k}, threshold={score_threshold}, reranker={reranker_type}"
         )
         sanitized_collection_name = sanitize_collection_name(self.collection_name, weaviate_format=True)
@@ -265,11 +265,10 @@ async def query_hybrid(
             if score < score_threshold:
                 continue
 
-            log.info(f"Document {chunk.metadata.get('document_id')} has score {score}")
             chunks.append(chunk)
             scores.append(score)
 
-        log.info(f"WEAVIATE HYBRID SEARCH RESULTS: Found {len(chunks)} chunks with scores {scores}")
+        log.debug(f"WEAVIATE HYBRID SEARCH RESULTS: Found {len(chunks)} chunks with scores {scores}")
         return QueryChunksResponse(chunks=chunks, scores=scores)
 
 
@@ -297,7 +296,7 @@ def __init__(
 
     def _get_client(self) -> weaviate.WeaviateClient:
         if "localhost" in self.config.weaviate_cluster_url:
-            log.info("using Weaviate locally in container")
+            log.info("Using Weaviate locally in container")
             host, port = self.config.weaviate_cluster_url.split(":")
             key = "local_test"
             client = weaviate.connect_to_local(

pyproject.toml

@@ -25,8 +25,8 @@ classifiers = [
 ]
 dependencies = [
     "aiohttp",
-    "fastapi>=0.115.0,<1.0", # server
-    "fire", # for MCP in LLS client
+    "fastapi>=0.115.0,<1.0",                  # server
+    "fire",                                   # for MCP in LLS client
     "httpx",
     "huggingface-hub>=0.34.0,<1.0",
     "jinja2>=3.1.6",
@@ -43,13 +43,12 @@ dependencies = [
     "tiktoken",
     "pillow",
     "h11>=0.16.0",
-    "python-multipart>=0.0.20", # For fastapi Form
-    "uvicorn>=0.34.0", # server
-    "opentelemetry-sdk>=1.30.0", # server
+    "python-multipart>=0.0.20",               # For fastapi Form
+    "uvicorn>=0.34.0",                        # server
+    "opentelemetry-sdk>=1.30.0",              # server
     "opentelemetry-exporter-otlp-proto-http>=1.30.0", # server
-    "aiosqlite>=0.21.0", # server - for metadata store
-    "asyncpg", # for metadata store
-    "weaviate-client>=4.16.5",
+    "aiosqlite>=0.21.0",                      # server - for metadata store
+    "asyncpg",                                # for metadata store
 ]
 
 [project.optional-dependencies]

tests/unit/providers/vector_io/remote/test_milvus.py

@@ -324,6 +324,3 @@ async def test_query_hybrid_search_default_rrf(
     call_args = mock_milvus_client.hybrid_search.call_args
     ranker = call_args[1]["ranker"]
     assert ranker is not None
-
-
-# TODO: Write tests for the MilvusVectorIOAdapter class.