TopolearnJupyterLab is a full workspace; using it as a single-notebook viewer leaves most of its productivity on the table.
JupyterLab arranges documents and tools in tabs you can drag to split the main area, for example a notebook beside the CSV it reads, or two views of the same notebook.
Features worth knowing:
- File browser (left sidebar): open notebooks, CSV files (shown as a grid), images, Markdown with live preview, and text editors with syntax highlighting.
- Consoles: right-click a notebook and choose "New Console for Notebook" to get a scratch area attached to the same kernel, ideal for quick checks without cluttering the notebook.
- Terminals: a full shell in the browser, useful on remote servers.
- Running panel: lists open kernels and terminals so you can shut down forgotten ones holding memory.
- Table of contents: built from Markdown headings, for navigating long notebooks.
- Debugger: enable the bug icon on a notebook to set breakpoints by clicking line numbers, then inspect variables and the call stack (requires an ipykernel with debugger support, which current versions have).
- Contextual help (Ctrl+I) shows documentation for the object under the cursor.
- Collapsible cells and outputs, and "Render side-by-side" for long outputs.
- Real-time collaboration via the
jupyter-collaborationextension.
Extensions add features such as Git integration (jupyterlab-git), language servers for completion and linting (jupyterlab-lsp) and formatters. In JupyterLab 4 most are installed with pip, with no build step; jupyter labextension list shows what is installed.
Example
jupyter lab --version # 4.6.x
jupyter labextension list # installed front-end extensions
# Common extensions are plain pip packages in JupyterLab 4
pip install jupyterlab-git # Git panel
pip install jupyterlab-lsp python-lsp-server # completion, hover, linting
pip install jupyter-collaboration # real-time collaboration
# Settings live in JSON and can be versioned per user
ls ~/.jupyter/lab/user-settings/@jupyterlab/Remember
- Split the main area to work on several documents side by side
- Attach a console to a notebook's kernel for scratch work
- Use the Running panel, debugger and table of contents; add extensions with pip
Watch out for
- Leaving many kernels running in the background, each holding data in memory.