CrumPy¶
Visualize quantum circuits with error propagation in a Jupyter widget.
This package builds off of the existing circuit visualizations of Crumble, a stabilizer circuit editor and sub-project of the open-source stabilizer circuit project Stim.
Installation¶
Requires: Python 3.11+
pip install crumpy
Usage¶
CrumPy provides a convenient Jupyter widget that makes use of Crumble’s ability to generate quantum circuit timeline visualizations with Pauli propagation from Stim circuit specifications.
Using CircuitWidget¶
CircuitWidget takes a
Stim circuit specification
in the form of a Python str. To convert from the official Stim package’s
stim.Circuit, use str(stim_circuit). If coming from another circuit type, it
is recommended to first convert to a stim.Circuit (e.g., with Stim’s
stimcirq package),
then to str. Note that not all circuits may be convertible to Stim.
from crumpy import CircuitWidget
your_circuit = """
H 0
CNOT 0 1
"""
circuit = CircuitWidget(stim=your_circuit)
circuit
Propagating Paulis¶
A useful feature of Crumble (and CrumPy) is the ability to propagate Paulis through a quantum circuit. Propagation is done automatically based on the specified circuit and Pauli markers. From the Crumble docs:
Propagating Paulis is done by placing markers to indicate where to add terms. Each marker has a type (X, Y, or Z) and an index (0-9) indicating which indexed Pauli product the marker is modifying.
Define a Pauli X, Y, or Z marker with the #!pragma MARK_ instruction. Note
that for compatibility with Stim, the ‘#!pragma ‘ is included as MARK_ is
not considered a standard Stim instruction:
z_error_on_qubit_2 = "#!pragma MARKZ(0) 2"
Legend¶
Example¶
circuit_with_error = """
#!pragma MARKX(0) 0
TICK
CNOT 0 1
TICK
H 0
"""
CircuitWidget(stim=circuit_with_error)
Notice how the single specified Pauli X marker propagates both through the control and across the connector of the CNOT, and gets transformed into a Pauli Z error when it encounters an H gate.
Local Development¶
See the contribution guidelines for a quick start guide and Python best practices.
Additional requirements¶
npm/Node.js (for bundling JavaScript); versions 11+/22+ recommended
Project Layout¶
CrumPy makes use of the widget creation tool AnyWidget. With AnyWidget, Python classes are able to take JavaScript and display custom widgets in Jupyter environments. CrumPy therefore includes both JavaScript and Python subparts:
crumpy/
├── src/
│ ├── crumpy/
│ │ ├── __init__.py # Python code for CircuitWidget
│ │ └── bundle.js # Bundled JavaScript will appear here
│ └── js/
│ ├── crumble/ # Modified Crumble code
│ │ ├── main.js # Main circuit visualization/setup
│ │ └── ...
│ └── package.json # Bundling setup and scripts
├── tests/ # Python tests
└── ...
src/crumpy contains the main Python package code.
src/crumpy/__init__.py contains the main class of the crumpy package,
CircuitWidget.
src/js/crumble/ contains the modified Crumble circuit visualization code that
will be rendered in the CircuitWidget widget.
src/js/crumble/main.js contains the main circuit visualization and setup logic
in the form of the render function
used by AnyWidget.
Bundling¶
To create a Jupyter widget, AnyWidget takes in JavaScript as an ECMAScript Module (ESM). Any changes made to the JavaScript code that backs the circuit visualization will require re-bundling into one optimized ESM file.
To bundle the JavaScript:
Navigate to
src/jsIf you haven’t yet, run
npm install(this will install the esbuild bundler)Run
npm run build(ornpm run watchfor watch mode)
A new bundle should appear at src/crumpy/bundle.js.
Note: If you are working in a Jupyter notebook and re-bundle the JavaScript, you may need to restart the notebook kernel and rerun any widget-displaying code for the changes to take effect.
Attribution¶
This package was created as part of Sam Zappa’s internship at Riverlane during the Summer of 2025. Thanks to Hyeok Kim, Leilani Battle, and Abe Asfaw for guidance and useful discussions.