feat(agent): add agent module and update related components

- Add agent module to reme_ai package
- Update retrieve module to include agent-related functionality
- Modify summary module to use new agent features
- Update configuration to support agent operations

reme_ai/__init__.py

@@ -1,5 +1,6 @@
 from reme_ai import retrieve
 from reme_ai import summary
+from reme_ai import agent
 from reme_ai import vector_store
 
 __version__ = "0.1.0"

reme_ai/config/default.yaml

@@ -62,15 +62,15 @@ flow:
         description: "current query"
         required: true
 
-    vector_store:
-      flow_content: vector_store_action_op
-      description: "directly operate the vector store."
-      input_schema:
-        action:
-          type: "str"
-          description: "vector store operations"
-          required: true
-          enum: [ copy, delete, delete_ids, dump, load ]
+  vector_store:
+    flow_content: vector_store_action_op
+    description: "directly operate the vector store."
+    input_schema:
+      action:
+        type: "str"
+        description: "vector store operations"
+        required: true
+        enum: [ copy, delete, delete_ids, dump, load ]
 
 llm:
   default:

reme_ai/retrieve/task/build_query_prompt.yaml

@@ -3,4 +3,4 @@ query_build: |
     {execution_process}
     
     Read through the entire execution process to understand which part is currently being executed. 
-    Generate a `query` that reflects the current state, which will later be used to search for similar problems in the database and help resolve the issue at hand.
+    Generate a `query` that reflects the current state, which will later be used to search for similar problems in the database and help resolve the issue at hand.

reme_ai/summary/task/simple_summary_op.py

@@ -52,7 +52,7 @@ class SimpleSummaryOp(BaseLLMOp):
         return self.llm.chat(messages=[Message(content=summary_prompt)], callback_fn=parse_content)
 
     def execute(self):
-        trajectories: list = self.context.get("trajectories", [])
+        trajectories: list = self.context.trajectories
         trajectories: List[Trajectory] = [Trajectory(**x) if isinstance(x, dict) else x for x in trajectories]
 
         memory_list: List[BaseMemory] = []

reme_ai/summary/task/success_extraction_op.py

@@ -14,7 +14,7 @@ class SuccessExtractionOp(BaseLLMOp):
 
     def execute(self):
         """Extract task memories from successful trajectories"""
-        success_trajectories: List[Trajectory] = self.context.get("success_trajectories", [])
+        success_trajectories: List[Trajectory] = self.context.success_trajectories
         
         if not success_trajectories:
             logger.info("No success trajectories found for extraction")