[pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Author: pre-commit-ci[bot]

EdgeCraftRAG/docs/Agentic_RAG_Guide.md

@@ -64,4 +64,4 @@ While DeepSearch agent is working, it will make plans on its own and execute the
 ![alt text](../assets/img/agent_deepsearch_plan.png)
 
 When all planned steps completed, DeepSearch agent will generate the final answer.
-![alt text](../assets/img/agent_deepsearch_answer.png)
+![alt text](../assets/img/agent_deepsearch_answer.png)

EdgeCraftRAG/docs/Experience_Guide.md

@@ -3,14 +3,17 @@
 This guide explains how to enable, configure, and use Experience capabilities within EdgeCraftRAG (EC-RAG). It covers the creation and management of experience to enhance the answer quality
 
 ## Overview
+
 Experience refers to the knowledge and skills acquired through practical involvement, trial, and reflection, serving as a key foundation for solving real-world problems. In EC-RAG we now provide abilities to add such experience for the reference of your Chat.
 
-Befor use Experience, please make sure Milvus VectorDB is enabled and below env is set:
+Before use Experience, please make sure Milvus VectorDB is enabled and below env is set:
+
 ```bash
 export METADATA_DATABASE_URL="http://${HOST_IP}:19530"
 ```
 
 ## Creating An Experience
+
 Upon successful deployment of EC-RAG, we may navigate to `Knowledge Base` tab from the left panel:
 ![alt text](../assets/img/experience_access.png)
 
@@ -35,10 +38,11 @@ In the `Create Experience` window, we could add experience and its corresponding
 
 ### View and Edit Experiences
 
-After experience added, they will be shown in list under `Experience` tab. We could toggle the `+` sign to show experience content details. Use the three dots next to `Experience` to activate or deactive a Experience Base:
+After experience added, they will be shown in list under `Experience` tab. We could toggle the `+` sign to show experience content details. Use the three dots next to `Experience` to activate or deactivate a Experience Base:
 ![alt text](../assets/img/experience_management_4.png)
 
 ## Examples
+
 Here's a comparison with experience activated and not activated:
 
 Without experience:

EdgeCraftRAG/docs/Experience_Guide_Zh.md

@@ -6,6 +6,7 @@
 
 Experience（经验）是指个人或团队在实践过程中积累的知识和技能，通常通过实际操作、试错和反思获得，是解决实际问题的重要依据。EC-RAG 现已支持添加此类Experience，为您的对话提供更多参考。
 使用Experience之前，请确保Mivlus服务已经启动，并且请指定如下的环境变量:
+
 ```bash
 export METADATA_DATABASE_URL="http://${HOST_IP}:19530"
 ```
@@ -27,13 +28,15 @@ export METADATA_DATABASE_URL="http://${HOST_IP}:19530"
 ![alt text](../assets/img/experience_management_1_zh.png)
 
 ### 添加 Experience
+
 当您需要添加新 Experience 时，请从右侧面板选择`新建经验`：
 ![alt text](../assets/img/experience_management_2_zh.png)
 
 在`新建经验`窗口中，您可以根据需要添加经验及其对应的内容，EC-RAG 支持创建多个经验条目：
 ![alt text](../assets/img/experience_management_3_zh.png)
 
 ### 查看与修改
+
 添加经验后，它们将以列表形式显示在`经验`选项卡下。您可以点击`+`号展开查看 Experience 内容的详细信息。使用左侧面板里`Experience`旁边的三个点可以激活或停用某个 Experience 库：
 ![alt text](../assets/img/experience_management_4_zh.png)
 

EdgeCraftRAG/edgecraftrag/api/v1/agent.py

@@ -5,8 +5,8 @@
 import os
 import time
 
-from edgecraftrag.api_schema import AgentCreateIn
 from edgecraftrag.api.v1.knowledge_base import Synchronizing_vector_data
+from edgecraftrag.api_schema import AgentCreateIn
 from edgecraftrag.base import AgentType
 from edgecraftrag.config_repository import MilvusConfigRepository, save_agent_configurations
 from edgecraftrag.context import ctx
@@ -188,4 +188,4 @@ async def manage_agent_bound_pipeline(bound_pl_idx, request):
 
     cur_active_pipeline = pl_manager.get_active_pipeline()
     if prev_active_pipeline and cur_active_pipeline and prev_active_pipeline.idx != cur_active_pipeline.idx:
-        await Synchronizing_vector_data(prev_active_pipeline, cur_active_pipeline)
+        await Synchronizing_vector_data(prev_active_pipeline, cur_active_pipeline)

EdgeCraftRAG/edgecraftrag/api/v1/chatqna.py

@@ -31,19 +31,13 @@ async def retrieval(request: ChatCompletionRequest):
                 status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
                 detail="Retrieval needs to have an active knowledgebase",
             )
-        contexts = await ctx.get_pipeline_mgr().run_retrieve_postprocess(
-            chat_request=request
-        )
+        contexts = await ctx.get_pipeline_mgr().run_retrieve_postprocess(chat_request=request)
         serialized_contexts = serialize_contexts(contexts)
 
-        ragout = RagOut(
-            query=request.messages, contexts=serialized_contexts, response=""
-        )
+        ragout = RagOut(query=request.messages, contexts=serialized_contexts, response="")
         return ragout
     except Exception as e:
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR, detail=str(e)
-        )
+        raise HTTPException(status_code=status.HTTP_500_INTERNAL_SERVER_ERROR, detail=str(e))
 
 
 # ChatQnA
@@ -62,9 +56,7 @@ async def chatqna(request: ChatCompletionRequest):
         generator = active_pl.get_generator(GeneratorType.CHATQNA)
         inference_type = generator.inference_type if generator else "local"
 
-        request.input = ctx.get_session_mgr().concat_history(
-            sessionid, inference_type, request.messages
-        )
+        request.input = ctx.get_session_mgr().concat_history(sessionid, inference_type, request.messages)
 
         # Run agent if activated, otherwise, run pipeline
         if ctx.get_agent_mgr().get_active_agent():
@@ -74,9 +66,7 @@ async def chatqna(request: ChatCompletionRequest):
         else:
             generator = active_pl.get_generator(GeneratorType.CHATQNA)
             if not generator:
-                raise Exception(
-                    "code:0000Please make sure chatqna generator is available in pipeline."
-                )
+                raise Exception("code:0000Please make sure chatqna generator is available in pipeline.")
             request.model = generator.model_id
 
         if request.stream:
@@ -110,9 +100,7 @@ async def ragqna(request: ChatCompletionRequest):
         if generator:
             request.model = generator.model_id
         if request.stream:
-            res_gen, contexts = await ctx.get_pipeline_mgr().run_pipeline(
-                chat_request=request
-            )
+            res_gen, contexts = await ctx.get_pipeline_mgr().run_pipeline(chat_request=request)
 
             # Escape newlines for json format as value
             async def res_gen_json():
@@ -125,26 +113,18 @@ async def res_gen_json():
             s_contexts = json.dumps(serialize_contexts(contexts))
             context_gen = stream_generator('"contexts":' + s_contexts + ',"response":"')
             final_gen = stream_generator('"}')
-            output_gen = chain_async_generators(
-                [query_gen, context_gen, res_gen_json(), final_gen]
-            )
+            output_gen = chain_async_generators([query_gen, context_gen, res_gen_json(), final_gen])
 
             return StreamingResponse(output_gen, media_type="text/plain")
         else:
-            ret, contexts = await ctx.get_pipeline_mgr().run_pipeline(
-                chat_request=request
-            )
+            ret, contexts = await ctx.get_pipeline_mgr().run_pipeline(chat_request=request)
             serialized_contexts = serialize_contexts(contexts)
 
-            ragout = RagOut(
-                query=request.messages, contexts=serialized_contexts, response=str(ret)
-            )
+            ragout = RagOut(query=request.messages, contexts=serialized_contexts, response=str(ret))
             return ragout
 
     except Exception as e:
-        raise HTTPException(
-            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR, detail=str(e)
-        )
+        raise HTTPException(status_code=status.HTTP_500_INTERNAL_SERVER_ERROR, detail=str(e))
 
 
 # Detecting if vllm is connected

EdgeCraftRAG/edgecraftrag/api/v1/pipeline.py

@@ -81,7 +81,7 @@ async def get_pipeline_benchmark():
 async def get_pipeline_benchmarks(name):
     pl = ctx.get_pipeline_mgr().get_pipeline_by_name_or_id(name)
     if pl and pl.benchmark:
-        bench_res = {'pipeline_bench': pl.benchmark.benchmark_data_list, 'llm_bench': pl.benchmark.llm_data_list}
+        bench_res = {"pipeline_bench": pl.benchmark.benchmark_data_list, "llm_bench": pl.benchmark.llm_data_list}
         return bench_res
 
 
@@ -314,10 +314,14 @@ async def update_pipeline_handler(pl, req):
                 # Use weakref to achieve model deletion and memory release
                 model_ref = weakref.ref(model)
                 if gen.generator_type == GeneratorType.CHATQNA:
-                    pl.generator.append(QnAGenerator(model_ref, gen.prompt_path, gen.inference_type, gen.vllm_endpoint, gen.prompt_content))
+                    pl.generator.append(
+                        QnAGenerator(
+                            model_ref, gen.prompt_path, gen.inference_type, gen.vllm_endpoint, gen.prompt_content
+                        )
+                    )
                 elif gen.generator_type == GeneratorType.FREECHAT:
                     pl.generator.append(FreeChatGenerator(model_ref, gen.inference_type, gen.vllm_endpoint))
-                
+
                 if pl.enable_benchmark:
                     if "tokenizer" not in locals() or tokenizer is None:
                         _, tokenizer, bench_hook = ctx.get_model_mgr().load_model_ben(gen.model)

EdgeCraftRAG/edgecraftrag/components/agent.py

@@ -72,7 +72,9 @@ async def run_pipeline_chatqna(self, request):
     async def _run_pipeline_generate(self, request):
         pl = self.get_bound_pipeline()
         if pl is not None:
-            return await pl.run(cbtype=CallbackType.GENERATE, chat_request=request, generator_type=GeneratorType.FREECHAT)
+            return await pl.run(
+                cbtype=CallbackType.GENERATE, chat_request=request, generator_type=GeneratorType.FREECHAT
+            )
 
     async def run_pipeline_retrieve_and_rerank(self, request):
         pl = self.get_bound_pipeline()

EdgeCraftRAG/edgecraftrag/components/benchmark.py

@@ -6,7 +6,7 @@
 from typing import Any, List, Optional
 
 import requests
-from edgecraftrag.base import BaseComponent, CompType, InferenceType, BenchType
+from edgecraftrag.base import BaseComponent, BenchType, CompType, InferenceType
 from prometheus_client.parser import text_string_to_metric_families
 from pydantic import BaseModel, Field, model_serializer
 

EdgeCraftRAG/edgecraftrag/components/generator.py

@@ -3,8 +3,8 @@
 
 import asyncio
 import json
-import time
 import os
+import time
 import urllib.request
 from concurrent.futures import ThreadPoolExecutor
 from urllib.parse import urlparse
@@ -123,6 +123,7 @@ async def stream_generator(llm, prompt_str, unstructured_str, benchmark=None, be
         result_error = str(e)[start_idx:]
         yield f"code:0000{result_error}"
 
+
 def clone_generator(src_generator: BaseComponent, dst_generator_cfg: dict = None):
     if not dst_generator_cfg:
         # If no config is provided, do a pure clone.
@@ -148,16 +149,19 @@ def clone_generator(src_generator: BaseComponent, dst_generator_cfg: dict = None
             return None
         # For QnAGenerator, we also need prompt-related info
         qna_args = shared_args.copy()
-        qna_args.update({
-            "prompt_template_file": src_generator.prompt_template_file,
-            "prompt_content": src_generator.prompt_content,
-        })
+        qna_args.update(
+            {
+                "prompt_template_file": src_generator.prompt_template_file,
+                "prompt_content": src_generator.prompt_content,
+            }
+        )
         new_generator = QnAGenerator(**qna_args)
     elif generator_type == GeneratorType.FREECHAT:
         new_generator = FreeChatGenerator(**shared_args)
 
     return new_generator
 
+
 class QnAGenerator(BaseComponent):
 
     def __init__(self, llm_model, prompt_template_file, inference_type, vllm_endpoint, prompt_content, **kwargs):

EdgeCraftRAG/edgecraftrag/components/model.py

@@ -48,13 +48,15 @@ def ser_model(self):
 
 class OpenAIEmbeddingModel(BaseModelComponent, OpenAIEmbedding):
     def __init__(self, model_id, api_base, **kwargs):
-        api_base = api_base+"/v1" if api_base and not api_base.endswith("/v1") else api_base
+        api_base = api_base + "/v1" if api_base and not api_base.endswith("/v1") else api_base
         super().__init__(
             model_id=model_id,
             api_base=api_base,
             **kwargs,
         )
-        OpenAIEmbedding.__init__(self, model_id_or_path=model_id, model_name=model_id, api_base=api_base, api_key="unused")
+        OpenAIEmbedding.__init__(
+            self, model_id_or_path=model_id, model_name=model_id, api_base=api_base, api_key="unused"
+        )
         self.comp_type = CompType.MODEL
         self.comp_subtype = ModelType.VLLM_EMBEDDING
         self.model_id = model_id