qupulse: A Quantum compUting PULse parametrization and SEquencing framework

The qupulse project aims to produce a software toolkit facilitating experiments involving pulse driven state manipulation of physical qubits.

It provides a high-level hardware-independent representation of pulses as well as means to translate this representation to hardware-specific device instructions and waveforms, execute these instructions and perform corresponding measurements.

Pulses can be assembled from previously defined subpulses, allowing easy construction of high-level from low-level pulses and re-use of previous work. Additionally, all pulses are parameterizable allowing users to fine-tune and adapt pulse templates to specific hardware or functionality without redefining an entire pulse sequence. To ensure meaningful parameter values, constraints can be put on parameters on a per-pulse basis.

Status and stability

The qupulse library is used productively by the Quantum Technology Group at the 2nd Institute of Physics at the RWTH Aachen University. As such, some features - such as pulse definition - are mostly complete and tested and interfaces are expected to remain largely stable (or changes to be backward compatible). A key goal is that experiments should be repeatable with new versions of qupulse. However, it is still possible for existing portions of the code base to be redesigned if this will increase the usability long-term.

The current feature list is as follows:

• Definition of complex (arbitrarily deep nested and looped pulses) parameterized pulses in Python (including measurement windows)
• Mathematical expression evaluation (based on sympy) for parameter values and parameter constraints
• Serialization of pulses (to allow storing into permanent storage)
• Hardware model representation
• High-level pulse to hardware configuration and waveform translation routines
• Hardware drivers for Tabor Electronics, Tektronix and Zurich Instruments AWGs and AlazarTech Digitizers
• MATLAB interface to access qupulse functionality

Pending changes are tracked in the changes.d subdirectory and published in RELEASE_NOTES.rst on release using the tool towncrier.

Installation

qupulse is available on PyPi and the latest release can be installed by executing:

python -m pip install qupulse[default]

which will install all required and optional dependencies except for hardware support. qupulse version numbers follow the Semantic Versioning conventions.

Alternatively, the current development version of qupulse can be installed by executing

python -m pip install -e git+https://github.com/qutech/qupulse.git#egg=qupulse[default]

which will clone the github repository to ./src/qupulse and do an editable/development install.

Requirements and dependencies

qupulse requires at least Python 3.10 and is tested on 3.10, 3.11 and 3.12. It relies on some external Python packages as dependencies. We intentionally did not restrict versions of dependencies in the install scripts to not unnecessarily prevent usage of newer releases of dependencies that might be compatible. However, if qupulse does encounter problems with a particular dependency version please file an issue.

The backend for TaborAWGs requires packages that can be found here.

The data acquisition backend for AlazarTech cards needs a package that unfortunately is not open source (yet). If you need it or have questions contact simon.humpohl@rwth-aachen.de.

Documentation

You can find documentation on how to use this library on readthedocs and IPython notebooks with examples in this repo. You can build it locally with hatch run docs:html.

Folder Structure

The repository primarily consists of the folders qupulse (toolkit core code) and tests (toolkit core tests). Additional parts of the project reside in MATLAB (MATLAB interface) and doc (configuration and source files to build documentation)

qupulse contains the entire Python source code of the project and is further partitioned the following packages of related modules

• pulses which contains all modules related to pulse representation.
• hardware containing classes for hardware representation as well as hardware drivers
• utils containing miscellaneous utility modules or wrapping code for external libraries
• _program contains general and hardware specific representations of instantiated (parameter free) pulses. It is private because there is no stability guarantee.

Contents of tests mirror the structure of qupulse. For every <module> somewhere in qupulse there should exist a <module>Tests.py in the corresponding subdirectory of tests.

Development

qupulse uses hatch as development tool which provides a convenient interface for most development tasks. The following should work.

• hatch build: Build wheel and source tarball
• hatch version X.X.X: Set version
• hatch run docs:html: Build documentation (requires pandoc)
• hatch run docs:clean-notebooks to execute all example notebooks that start with 00-03 and clean all metadata.
• hatch run changelog:draft and hatch run changelog:release to preview or update the changelog.

License

The current version of qupulse is available under the GPL-3.0-or-later license. Versions up to and including 0.10 were licensed under the MIT license. If you require different licensing terms, please contact us to discuss your needs.