import numpy as np
# Load and clean the dataset
df = pd.read_csv('data/train.csv')
df = df.dropna()
Open-source Cursor for Notebooks — agent inside every cell. Chat with your notebook, auto-fix errors, refactor across cells. Notebook界的Cursor — 每个单元格都有AI代理。与你的Notebook对话,自动修复错误,跨单元格重构代码。
A real agent that reads your cells, runs them, sees the output, and edits them back. 真正的AI代理:读取单元格、运行代码、查看输出、编辑修改
Select code, describe the change in natural language. Accept the diff with Enter, reject with Esc. Works inside any cell.选中代码,自然语言描述需求。Enter接受改动,Esc拒绝。适用于任意单元格。
Core核心Multi-step plan → execute → verify loop with cell-level tools: read_cell, edit_cell, insert_cell, run_cell.多步规划 → 执行 → 验证闭环。配备单元格级工具:读取、编辑、插入、运行。
AgentAgent@cell, @file references, slash commands, full notebook context, streaming responses.支持@cell、@file引用、斜杠命令、完整Notebook上下文、流式响应。
Chat对话Copilot-style inline completion native to JupyterLab. Start typing, see suggestions in gray.JupyterLab原生支持的Copilot风格行内补全。开始输入,灰色提示实时呈现。
Completion补全One click on the 🐛 button after an error. The agent diagnoses and patches the cell automatically.出错后点击🐛按钮。AI自动诊断并修复单元格。
Auto-Fix自动修复Anthropic, OpenAI, Google, Azure, Ollama, vLLM, any OpenAI-compatible endpoint. Local-first & privacy-first.支持Anthropic、OpenAI、Google、Azure、Ollama、vLLM及任意OpenAI兼容端点。本地优先,隐私优先。
Flexible灵活See why Jupyter Studio is the best choice for AI-powered notebook development. 了解Jupyter Studio为何是AI驱动Notebook开发的最佳选择
Watch the agent think, search, read, and edit — all inside your notebook. 观察Agent思考、搜索、读取、编辑 — 全程在你的Notebook中完成
Agent analyzes your request and plans the approachAgent分析你的需求并制定计划
Searches across cells for relevant context在单元格中搜索相关上下文
Reads cell contents and understands dependencies读取单元格内容,理解代码依赖
Makes precise edits and runs to verify精确编辑并运行验证
Connect to any LLM provider — cloud or local. Your code never leaves your machine unless you want it to. 连接任意LLM提供商 — 云端或本地。除非你想,否则你的代码不会离开你的设备。
Join thousands of data scientists and ML engineers using Jupyter Studio. 加入成千上万使用Jupyter Studio的数据科学家和机器学习工程师