I met a strange issue when run a Spark job in DataProc Image 2.0 which Spark Version is 3.1.3. Let’s introduce this problem and how I can debug and resolve it.
简单介绍一个结合 Milvus(向量数据库)、Elasticsearch(全文检索/BM25),并通过 LangChain 构建的 Retrieval 方案。
我创建了一个名为CAT(代码助手工具)的简单项目,用来学习Continue.dev的工作原理。
注意: 该项目目前仅实现了基础的消息传递架构和基础设施。尚未实现实际的代码助手功能。重点是建立组件之间的通信框架。
I created a simple project called CAT (Code Assistant Tool) to learn how Continue.dev works.
Note: This project currently only implements the basic messaging architecture and infrastructure. The actual code assistant functionality has not yet been implemented. The focus is on establishing the communication framework between components.
MCP 最近很火,自己也为公司内部的Code Assistant Tool 实现了一个MCP Client。因为网上几乎全是教大家怎么写MCP Server的,很少有关于MCP Client,kotlin 实现的就更少了。所以我觉得有必要记录一下过程,以及遇到的问题,希望可以帮助大家。