Restore 0.1.5 version from stash

LightRAG/lightrag/lightrag.py

@@ -7,7 +7,7 @@ import inspect
 import os
 import time
 import warnings
-from dataclasses import asdict, dataclass, field
+from dataclasses import dataclass, field
 from datetime import datetime, timezone
 from functools import partial
 from typing import (
@@ -516,7 +516,8 @@ class LightRAG:
             )
 
         # Fix global_config now
-        global_config = asdict(self)
+        # 避免 deepcopy 带来的 asyncpg __reduce__ 异常，统一使用浅复制配置
+        global_config = self._global_config()
 
         _print_config = ",\n  ".join([f"{k} = {v}" for k, v in global_config.items()])
         logger.debug(f"LightRAG init with param:\n  {_print_config}\n")
@@ -655,6 +656,10 @@ class LightRAG:
 
         self._storages_status = StoragesStatus.CREATED
 
+    def _global_config(self) -> dict[str, Any]:
+        """浅拷贝全局配置，避免 dataclasses.asdict 深拷贝底层连接对象。"""
+        return dict(self.__dict__)
+
     async def initialize_storages(self):
         """Storage initialization must be called one by one to prevent deadlock"""
         if self._storages_status == StoragesStatus.CREATED:
@@ -1863,12 +1868,16 @@ class LightRAG:
                             await asyncio.gather(*first_stage_tasks)
 
                             # Stage 2: Process entity relation graph (after text_chunks are saved)
-                            entity_relation_task = asyncio.create_task(
-                                self._process_extract_entities(
-                                    chunks, pipeline_status, pipeline_status_lock
+                            # 可通过环境变量关闭实体/关系抽取，避免非必要的 KG 处理带来的失败
+                            if not os.getenv("LIGHTRAG_DISABLE_ENTITY_RELATION", "false").lower() == "true":
+                                entity_relation_task = asyncio.create_task(
+                                    self._process_extract_entities(
+                                        chunks, pipeline_status, pipeline_status_lock
+                                    )
                                 )
-                            )
-                            chunk_results = await entity_relation_task
+                                chunk_results = await entity_relation_task
+                            else:
+                                chunk_results = []
                             file_extraction_stage_ok = True
 
                         except Exception as e:
@@ -1951,24 +1960,25 @@ class LightRAG:
                                         )
 
                                 # Use chunk_results from entity_relation_task
-                                await merge_nodes_and_edges(
-                                    chunk_results=chunk_results,  # result collected from entity_relation_task
-                                    knowledge_graph_inst=self.chunk_entity_relation_graph,
-                                    entity_vdb=self.entities_vdb,
-                                    relationships_vdb=self.relationships_vdb,
-                                    global_config=asdict(self),
-                                    full_entities_storage=self.full_entities,
-                                    full_relations_storage=self.full_relations,
-                                    doc_id=doc_id,
-                                    pipeline_status=pipeline_status,
-                                    pipeline_status_lock=pipeline_status_lock,
-                                    llm_response_cache=self.llm_response_cache,
-                                    entity_chunks_storage=self.entity_chunks,
-                                    relation_chunks_storage=self.relation_chunks,
-                                    current_file_number=current_file_number,
-                                    total_files=total_files,
-                                    file_path=file_path,
-                                )
+                                if not os.getenv("LIGHTRAG_DISABLE_ENTITY_RELATION", "false").lower() == "true":
+                                    await merge_nodes_and_edges(
+                                        chunk_results=chunk_results,  # result collected from entity_relation_task
+                                        knowledge_graph_inst=self.chunk_entity_relation_graph,
+                                        entity_vdb=self.entities_vdb,
+                                        relationships_vdb=self.relationships_vdb,
+                                        global_config=self._global_config(),
+                                        full_entities_storage=self.full_entities,
+                                        full_relations_storage=self.full_relations,
+                                        doc_id=doc_id,
+                                        pipeline_status=pipeline_status,
+                                        pipeline_status_lock=pipeline_status_lock,
+                                        llm_response_cache=self.llm_response_cache,
+                                        entity_chunks_storage=self.entity_chunks,
+                                        relation_chunks_storage=self.relation_chunks,
+                                        current_file_number=current_file_number,
+                                        total_files=total_files,
+                                        file_path=file_path,
+                                    )
 
                                 # Record processing end time
                                 processing_end_time = int(time.time())
@@ -2139,7 +2149,7 @@ class LightRAG:
         try:
             chunk_results = await extract_entities(
                 chunk,
-                global_config=asdict(self),
+                global_config=self._global_config(),
                 pipeline_status=pipeline_status,
                 pipeline_status_lock=pipeline_status_lock,
                 llm_response_cache=self.llm_response_cache,
@@ -2553,7 +2563,7 @@ class LightRAG:
             actual data is nested under the 'data' field, with 'status' and 'message'
             fields at the top level.
         """
-        global_config = asdict(self)
+        global_config = self._global_config()
 
         # Create a copy of param to avoid modifying the original
         data_param = QueryParam(
@@ -2671,7 +2681,7 @@ class LightRAG:
         """
         logger.debug(f"[aquery_llm] Query param: {param}")
 
-        global_config = asdict(self)
+        global_config = self._global_config()
 
         try:
             query_result = None
@@ -3462,7 +3472,7 @@ class LightRAG:
                         relationships_vdb=self.relationships_vdb,
                         text_chunks_storage=self.text_chunks,
                         llm_response_cache=self.llm_response_cache,
-                        global_config=asdict(self),
+                        global_config=self._global_config(),
                         pipeline_status=pipeline_status,
                         pipeline_status_lock=pipeline_status_lock,
                         entity_chunks_storage=self.entity_chunks,