Revert "Remove auto_manage_storages_states option"

This reverts commit bfe6657b31.

README-zh.md

@@ -204,6 +204,7 @@ async def initialize_rag():
         embedding_func=openai_embed,
         llm_model_func=gpt_4o_mini_complete,
     )
+    await rag.initialize_storages()
     await initialize_pipeline_status()
     return rag
 
@@ -399,6 +400,7 @@ async def initialize_rag():
         )
     )
 
+    await rag.initialize_storages()
     await initialize_pipeline_status()
 
     return rag
@@ -545,6 +547,7 @@ async def initialize_rag():
         ),
     )
 
+    await rag.initialize_storages()
     await initialize_pipeline_status()
 
     return rag
@@ -762,6 +765,8 @@ async def initialize_rag():
         graph_storage="Neo4JStorage", #<-----------覆盖KG默认值
     )
 
+    # 初始化数据库连接
+    await rag.initialize_storages()
     # 初始化文档处理的管道状态
     await initialize_pipeline_status()
 
@@ -1188,6 +1193,9 @@ LightRAG 现已与 [RAG-Anything](https://github.com/HKUDS/RAG-Anything) 实现
                 )
             )
 
+            # 初始化存储（如果有现有数据，这将加载现有数据）
+            await lightrag_instance.initialize_storages()
+
             # 现在使用现有的 LightRAG 实例初始化 RAGAnything
             rag = RAGAnything(
                 lightrag=lightrag_instance,  # 传递现有的 LightRAG 实例

README.md

@@ -181,7 +181,8 @@ For a streaming response implementation example, please see `examples/lightrag_o
 
 ### ⚠️ Important: Initialization Requirements
 
-**LightRAG requires explicit initialization before use.** You must call `await initialize_pipeline_status()` after creating a LightRAG instance, otherwise you will encounter errors like:
+**LightRAG requires explicit initialization before use.** You must call both `await rag.initialize_storages()` and `await initialize_pipeline_status()` after creating a LightRAG instance, otherwise you will encounter errors like:
+- `AttributeError: __aenter__` - if storages are not initialized
 - `KeyError: 'history_messages'` - if pipeline status is not initialized
 
 ### A Simple Program
@@ -208,8 +209,9 @@ async def initialize_rag():
         embedding_func=openai_embed,
         llm_model_func=gpt_4o_mini_complete,
     )
-    # IMPORTANT: Initialize document processing pipeline status is required!
-    await initialize_pipeline_status()  #
+    # IMPORTANT: Both initialization calls are required!
+    await rag.initialize_storages()  # Initialize storage backends
+    await initialize_pipeline_status()  # Initialize processing pipeline
     return rag
 
 async def main():
@@ -399,6 +401,7 @@ async def initialize_rag():
         )
     )
 
+    await rag.initialize_storages()
     await initialize_pipeline_status()
 
     return rag
@@ -547,6 +550,7 @@ async def initialize_rag():
         ),
     )
 
+    await rag.initialize_storages()
     await initialize_pipeline_status()
 
     return rag
@@ -772,6 +776,8 @@ async def initialize_rag():
         graph_storage="Neo4JStorage", #<-----------override KG default
     )
 
+    # Initialize database connections
+    await rag.initialize_storages()
     # Initialize pipeline status for document processing
     await initialize_pipeline_status()
 
@@ -856,6 +862,8 @@ async def initialize_rag():
         graph_storage="MemgraphStorage", #<-----------override KG default
     )
 
+    # Initialize database connections
+    await rag.initialize_storages()
     # Initialize pipeline status for document processing
     await initialize_pipeline_status()
 
@@ -1231,6 +1239,8 @@ LightRAG now seamlessly integrates with [RAG-Anything](https://github.com/HKUDS/
                     ),
                 )
             )
+            # Initialize storage (this will load existing data if available)
+            await lightrag_instance.initialize_storages()
             # Now initialize RAGAnything with the existing LightRAG instance
             rag = RAGAnything(
                 lightrag=lightrag_instance,  # Pass the existing LightRAG instance
@@ -1423,16 +1433,24 @@ Valid modes are:
 
 ### Common Initialization Errors
 
-If you encounter the following error when using LightRAG:
+If you encounter these errors when using LightRAG:
 
-- **`KeyError: 'history_messages'`**
-- **Cause**: Pipeline status not initialized
-- **Solution**: Call `await initialize_pipeline_status()` after initializing storages
+1. **`AttributeError: __aenter__`**
+   - **Cause**: Storage backends not initialized
+   - **Solution**: Call `await rag.initialize_storages()` after creating the LightRAG instance
 
-```python
-rag = LightRAG(...)
-await initialize_pipeline_status()
-```
+2. **`KeyError: 'history_messages'`**
+   - **Cause**: Pipeline status not initialized
+   - **Solution**: Call `await initialize_pipeline_status()` after initializing storages
+
+3. **Both errors in sequence**
+   - **Cause**: Neither initialization method was called
+   - **Solution**: Always follow this pattern:
+   ```python
+   rag = LightRAG(...)
+   await rag.initialize_storages()
+   await initialize_pipeline_status()
+   ```
 
 ### Model Switching Issues
 

docs/rerank_integration.md

@@ -174,6 +174,7 @@ async def main():
         rerank_model_func=my_rerank_func,
     )
 
+    await rag.initialize_storages()
     await initialize_pipeline_status()
 
     # Insert documents

examples/lightrag_openai_compatible_demo.py

@@ -4,7 +4,7 @@ import inspect
 import logging
 import logging.config
 from lightrag import LightRAG, QueryParam
-from lightrag.llm.openai import openai_complete_if_cache, openai_embed
+from lightrag.llm.openai import openai_complete_if_cache
 from lightrag.llm.ollama import ollama_embed
 from lightrag.utils import EmbeddingFunc, logger, set_verbose_debug
 from lightrag.kg.shared_storage import initialize_pipeline_status
@@ -99,26 +99,6 @@ async def llm_model_func(
     )
 
 
-ollama_embedding_func = EmbeddingFunc(
-    embedding_dim=int(os.getenv("EMBEDDING_DIM", "1024")),
-    func=lambda texts: ollama_embed(
-        texts,
-        embed_model=os.getenv("EMBEDDING_MODEL", "bge-m3:latest"),
-        host=os.getenv("EMBEDDING_BINDING_HOST", "http://localhost:11434"),
-    ),
-)
-
-openai_embedding_func = EmbeddingFunc(
-    embedding_dim=int(os.getenv("EMBEDDING_DIM", "1024")),
-    func=lambda texts: openai_embed(
-        texts,
-        model=os.getenv("EMBEDDING_MODEL", "BAAI/bge-m3"),
-        base_url=os.getenv("EMBEDDING_BINDING_HOST", "https://api.deepseek.com"),
-        api_key=os.getenv("EMBEDDING_BINDING_API_KEY") or os.getenv("OPENAI_API_KEY"),
-    ),
-)
-
-
 async def print_stream(stream):
     async for chunk in stream:
         if chunk:
@@ -129,9 +109,18 @@ async def initialize_rag():
     rag = LightRAG(
         working_dir=WORKING_DIR,
         llm_model_func=llm_model_func,
-        embedding_func=openai_embedding_func,
+        embedding_func=EmbeddingFunc(
+            embedding_dim=int(os.getenv("EMBEDDING_DIM", "1024")),
+            max_token_size=int(os.getenv("MAX_EMBED_TOKENS", "8192")),
+            func=lambda texts: ollama_embed(
+                texts,
+                embed_model=os.getenv("EMBEDDING_MODEL", "bge-m3:latest"),
+                host=os.getenv("EMBEDDING_BINDING_HOST", "http://localhost:11434"),
+            ),
+        ),
     )
 
+    await rag.initialize_storages()
     await initialize_pipeline_status()
 
     return rag

lightrag/api/lightrag_server.py

@@ -149,6 +149,9 @@ def create_app(args):
         app.state.background_tasks = set()
 
         try:
+            # Initialize database connections
+            await rag.initialize_storages()
+
             await initialize_pipeline_status()
             pipeline_status = await get_namespace_data("pipeline_status")
 
@@ -398,6 +401,7 @@ def create_app(args):
             enable_llm_cache_for_entity_extract=args.enable_llm_cache_for_extract,
             enable_llm_cache=args.enable_llm_cache,
             rerank_model_func=rerank_model_func,
+            auto_manage_storages_states=False,
             max_parallel_insert=args.max_parallel_insert,
             max_graph_nodes=args.max_graph_nodes,
             addon_params={"language": args.summary_language},
@@ -427,6 +431,7 @@ def create_app(args):
             enable_llm_cache_for_entity_extract=args.enable_llm_cache_for_extract,
             enable_llm_cache=args.enable_llm_cache,
             rerank_model_func=rerank_model_func,
+            auto_manage_storages_states=False,
             max_parallel_insert=args.max_parallel_insert,
             max_graph_nodes=args.max_graph_nodes,
             addon_params={"language": args.summary_language},

lightrag/lightrag.py

@@ -334,7 +334,6 @@ class LightRAG:
     # Storages Management
     # ---
 
-    # TODO: Deprecated (LightRAG will always manage storages states)
     auto_manage_storages_states: bool = field(default=True)
     """If True, lightrag will automatically calls initialize_storages and finalize_storages at the appropriate times."""
 
@@ -532,14 +531,12 @@ class LightRAG:
 
         self._storages_status = StoragesStatus.CREATED
 
-        # Initialize storages
-        self._run_async_safely(self.initialize_storages, "Storage Initialization")
-
-        # Check and perform data migration if needed
-        self._run_async_safely(self._check_and_migrate_data, "Data Migration Check")
+        if self.auto_manage_storages_states:
+            self._run_async_safely(self.initialize_storages, "Storage Initialization")
 
     def __del__(self):
-        self._run_async_safely(self.finalize_storages, "Storage Finalization")
+        if self.auto_manage_storages_states:
+            self._run_async_safely(self.finalize_storages, "Storage Finalization")
 
     def _run_async_safely(self, async_func, action_name=""):
         """Safely execute an async function, avoiding event loop conflicts."""

reproduce/Step_1.py

@@ -35,6 +35,7 @@ if not os.path.exists(WORKING_DIR):
 async def initialize_rag():
     rag = LightRAG(working_dir=WORKING_DIR)
 
+    await rag.initialize_storages()
     await initialize_pipeline_status()
 
     return rag

reproduce/Step_1_openai_compatible.py

@@ -70,6 +70,7 @@ async def initialize_rag():
         embedding_func=EmbeddingFunc(embedding_dim=4096, func=embedding_func),
     )
 
+    await rag.initialize_storages()
     await initialize_pipeline_status()
 
     return rag