updated files to comply with graphrag-v2++

backend/graphrag_app/api/data.py

@@ -40,13 +40,15 @@ data_route = APIRouter(
     response_model=StorageNameList,
     responses={200: {"model": StorageNameList}},
 )
-async def get_all_data_containers():
+async def get_all_data_containers(
+    container_store_client=Depends(get_cosmos_container_store_client),
+):
     """
     Retrieve a list of all data containers.
     """
     items = []
     try:
-        container_store_client = get_cosmos_container_store_client()
+        # container_store_client = get_cosmos_container_store_client()
         for item in container_store_client.read_all_items():
             if item["type"] == "data":
                 items.append(item["human_readable_name"])
@@ -161,6 +163,8 @@ async def upload_files(
         })
         return BaseResponse(status="File upload successful.")
     except Exception as e:
+        # import traceback
+        # traceback.print_exc()
         logger = load_pipeline_logger()
         logger.error(
             message="Error uploading files.",

backend/graphrag_app/logger/__init__.py

@@ -9,16 +9,16 @@ from graphrag_app.logger.load_logger import load_pipeline_logger
 from graphrag_app.logger.pipeline_job_updater import PipelineJobUpdater
 from graphrag_app.logger.typing import (
     Logger,
-    PipelineAppInsightsReportingConfig,
-    PipelineReportingConfigTypes,
+    PipelineAppInsightsLogger,
+    # PipelineReportingConfigTypes,
 )
 
 __all__ = [
     "Logger",
     "ApplicationInsightsWorkflowCallbacks",
     "ConsoleWorkflowCallbacks",
-    "PipelineAppInsightsReportingConfig",
+    "PipelineAppInsightsLogger",
     "PipelineJobUpdater",
-    "PipelineReportingConfigTypes",
+    # "PipelineReportingConfigTypes",
     "load_pipeline_logger",
 ]

backend/graphrag_app/logger/load_logger.py

@@ -2,7 +2,7 @@
 # Licensed under the MIT License.
 
 import os
-from pathlib import Path
+from pathlib import PurePosixPath
 from typing import List
 
 from graphrag.callbacks.file_workflow_callbacks import FileWorkflowCallbacks
@@ -44,7 +44,9 @@ def load_pipeline_logger(
                     log_blob_name = os.path.join(logging_dir, log_blob_name)
                 # ensure the root directory exists; if not, create it
                 blob_service_client = azure_client_manager.get_blob_service_client()
-                container_root = Path(log_blob_name).parts[0]
+                container_root = PurePosixPath(log_blob_name).parts[0]
+                print(f"container_root={container_root}")
+                print(f"container_exist={blob_service_client.get_container_client(container_root).exists()}")
                 if not blob_service_client.get_container_client(
                     container_root
                 ).exists():

backend/graphrag_app/logger/typing.py

@@ -5,10 +5,7 @@ import logging
 from enum import Enum
 from typing import Literal
 
-from graphrag.index.config.reporting import (
-    PipelineReportingConfig,
-    PipelineReportingConfigTypes,
-)
+from graphrag.logger.base import StatusLogger
 from pydantic import Field as pydantic_Field
 
 
@@ -19,9 +16,7 @@ class Logger(Enum):
     APP_INSIGHTS = (4, "app_insights")
 
 
-class PipelineAppInsightsReportingConfig(
-    PipelineReportingConfig[Literal["app_insights"]]
-):
+class PipelineAppInsightsLogger(StatusLogger):
     """Represents the ApplicationInsights reporting configuration for the pipeline."""
 
     type: Literal["app_insights"] = Logger.APP_INSIGHTS.name.lower()
@@ -45,6 +40,6 @@ class PipelineAppInsightsReportingConfig(
 
 
 # add the new type to the existing PipelineReportingConfigTypes
-PipelineReportingConfigTypes = (
-    PipelineReportingConfigTypes | PipelineAppInsightsReportingConfig
-)
+# StatusLogger = (
+#     StatusLogger | PipelineAppInsightsReportingConfig
+# )

backend/graphrag_app/utils/azure_clients.py

@@ -11,6 +11,7 @@ from azure.cosmos import (
 from azure.identity import DefaultAzureCredential
 from azure.storage.blob import BlobServiceClient
 from azure.storage.blob.aio import BlobServiceClient as BlobServiceClientAsync
+from pathlib import PurePosixPath
 
 ENDPOINT_ERROR_MSG = "Could not find connection string in environment variables"
 
@@ -115,6 +116,9 @@ class AzureClientManager:
             _BlobServiceClientSingletonAsync.get_instance()
         )
 
+        # parse account hostname from the azure storage connection string or blob url
+        self.storage_account_hostname = PurePosixPath(self.storage_blob_url).parts[1]
+        
         # parse account name from the azure storage connection string or blob url
         if self.storage_connection_string:
             meta_info = {}
@@ -127,12 +131,8 @@ class AzureClientManager:
                 meta_info[m[0]] = m[1]
             self.storage_account_name = meta_info["AccountName"]
         else:
-            self.storage_account_name = self.storage_blob_url.split("//")[1].split(".")[
-                0
-            ]
+            self.storage_account_name = self.storage_account_hostname.split(".")[0]
 
-        # parse account hostname from the azure storage connection string or blob url
-        self.storage_account_hostname = self._blob_service_client.url.split("//")[1]
 
     def get_blob_service_client(self) -> BlobServiceClient:
         """

backend/pyproject.toml

@@ -51,7 +51,8 @@ fastapi = ">=0.110.0"
 fastapi-offline = ">=1.7.3"
 fastparquet = ">=2023.10.1"
 fsspec = ">=2024.2.0"
-graphrag = "==1.2.0"
+# graphrag = "==1.2.0"
+graphrag = { git = "https://github.com/microsoft/graphrag.git", rev = "3b1e70c06b5a22efe6ca38ac16e83d0428f660f0" }
 httpx = ">=0.25.2"
 kubernetes = ">=29.0.0"
 networkx = ">=3.2.1"