📖 Debugger in Jupyterlab

JupyterLab ships with a Debugger front-end by default.

Usage

Here is a screencast showing how to enable the debugger and set up breakpoints. The steps are described in detail below.

Use a Kernel Supporting Debugger

First, check that the kernel supports debugging. If it does, the bug icon in the upper-right corner of the notebook will be enabled.

Debug Code in Notebook

To enable the debugger for a notebook, toggle the bug button in the upper-right corner of the notebook:

Once debugging is enabled, you can set breakpoints and step into the code.

Define a function that adds two elements:

def add(a, b):
   res = a + b
   return res

Call the function and print the result:

result = add(1, 2)
print(result)

3

To add a breakpoint, click on the gutter on line number 2 in the first code cell:

Execute the second code cell by clicking on the Run button:

Execution will stop where the breakpoint is set:

Explore the Code State

The debugger sidebar allows you to explore the code state. It includes a variable explorer, a list of breakpoints, a source preview, and the ability to navigate the call stack.

Variables

Variables can be explored using a tree view and a table view:

Call Stack

You can step into the code and continue execution using debug actions:

Breakpoints

New breakpoints can be added or removed while execution is stopped. They will appear in the list of breakpoints:

Source

The source panel shows the source of the current file being debugged:

Adapted from: jupyterlab documentation