reconfigure vector store size (microsoft#2281)

Author: dayesouza

.semversioner/next-release/patch-20260315024056229023.json

@@ -0,0 +1,4 @@
+{
+  "type": "patch",
+  "description": "reconfigure vector store size by embedding model"
+}

packages/graphrag/graphrag/index/validate_config.py

@@ -6,12 +6,16 @@
 import asyncio
 import logging
 import sys
+from typing import TYPE_CHECKING
 
 from graphrag_llm.completion import create_completion
 from graphrag_llm.embedding import create_embedding
 
 from graphrag.config.models.graph_rag_config import GraphRagConfig
 
+if TYPE_CHECKING:
+    from graphrag_llm.types import LLMEmbeddingResponse
+
 logger = logging.getLogger(__name__)
 
 
@@ -29,13 +33,41 @@ def validate_config_names(parameters: GraphRagConfig) -> None:
     for id, config in parameters.embedding_models.items():
         embed_llm = create_embedding(config)
         try:
-            asyncio.run(
+            response = asyncio.run(
                 embed_llm.embedding_async(
                     input=["This is an LLM Embedding Test String"]
                 )
             )
             logger.info("Embedding LLM Config Params Validated")
+
+            if id == parameters.embed_text.embedding_model_id:
+                _sync_vector_store_dimensions(parameters, response)
+
         except Exception as e:  # noqa: BLE001
             logger.error(f"Embedding configuration error detected.\n{e}")  # noqa
             print(f"Failed to validate embedding model ({id}) params", e)  # noqa: T201
             sys.exit(1)
+
+
+def _sync_vector_store_dimensions(
+    parameters: GraphRagConfig,
+    response: "LLMEmbeddingResponse",
+) -> None:
+    """Sync vector store dimensions to match the actual embedding model output."""
+    detected = len(response.first_embedding)
+    if detected == 0:
+        return
+
+    configured = parameters.vector_store.vector_size
+    if detected == configured:
+        return
+
+    logger.warning(
+        "Embedding model produces %d-dimensional vectors but vector_size is "
+        "configured as %d. Overriding vector_size to match the model.",
+        detected,
+        configured,
+    )
+    parameters.vector_store.vector_size = detected
+    for schema in parameters.vector_store.index_schema.values():
+        schema.vector_size = detected