SQL Server Storage Support for Indexing Pipeline

.semversioner/next-release/patch-20250417174516646520.json

@@ -0,0 +1,4 @@
+{
+  "type": "patch",
+  "description": "SQL server storage support for indexing"
+}

graphrag/config/defaults.py

@@ -324,6 +324,10 @@ class OutputDefaults:
     container_name: None = None
     storage_account_blob_url: None = None
     cosmosdb_account_url: None = None
+    database_name: None = None
+    database_server_name: None = None
+    autogenerate_tables: None = None
+    client_id: None = None
 
 
 @dataclass

graphrag/config/enums.py

@@ -66,6 +66,8 @@ class OutputType(str, Enum):
     """The blob output type."""
     cosmosdb = "cosmosdb"
     """The cosmosdb output type"""
+    sqlserver = "sqlserver"
+    """The sqlserver output type"""
 
     def __repr__(self):
         """Get a string representation."""

graphrag/config/models/output_config.py

@@ -36,3 +36,19 @@ class OutputConfig(BaseModel):
         description="The cosmosdb account url to use.",
         default=graphrag_config_defaults.output.cosmosdb_account_url,
     )
+    database_name: str | None = Field(
+        description="The database name to use for SQL Server.",
+        default=graphrag_config_defaults.output.database_name,
+    )
+    database_server_name: str | None = Field(
+        description="The database server name to use for SQL Server.",
+        default=graphrag_config_defaults.output.database_server_name,
+    )
+    autogenerate_tables: bool | None = Field(
+        description="Whether to automatically generate tables in SQL Server.",
+        default=graphrag_config_defaults.output.autogenerate_tables,
+    )
+    client_id: str | None = Field(
+        description="The client ID for a User-Assigned Managed Identity (SQL Server).",
+        default=graphrag_config_defaults.output.client_id,
+    )

graphrag/index/operations/summarize_communities/graph_context/context_builder.py

@@ -44,7 +44,6 @@ def build_local_context(
 ):
     """Prep communities for report generation."""
     levels = get_levels(nodes, schemas.COMMUNITY_LEVEL)
-
     dfs = []
 
     for level in progress_iterable(levels, callbacks.progress, len(levels)):

graphrag/storage/factory.py

@@ -12,6 +12,7 @@
 from graphrag.storage.cosmosdb_pipeline_storage import create_cosmosdb_storage
 from graphrag.storage.file_pipeline_storage import create_file_storage
 from graphrag.storage.memory_pipeline_storage import MemoryPipelineStorage
+from graphrag.storage.sql_server_pipeline_storage import create_sql_server_storage
 
 if TYPE_CHECKING:
     from graphrag.storage.pipeline_storage import PipelineStorage
@@ -45,6 +46,8 @@ def create_storage(
                 return create_cosmosdb_storage(**kwargs)
             case OutputType.file:
                 return create_file_storage(**kwargs)
+            case OutputType.sqlserver:
+                return create_sql_server_storage(**kwargs)
             case OutputType.memory:
                 return MemoryPipelineStorage()
             case _: