Pygsti

Fixed

- Fixes pip installation issue caused by latest version of plotly (v3.8), which manifests itself as an inability to import _plot_html. This hotfix simply pins the required Plot.ly version to < 3.7.1; compatibility with Plot.ly version 3.8 will be included in the next non-hotfix release.

Fixed

- Minor bug fixes are also included.

Changed

- updates and extends pyGSTi's drift-detection protocols.

Fixed

- fixes an issue whereby copied Model objects aren't given new UUIDs. This is relevant when generating reports which view multiple models.

A significant update which shifts pyGSTi's focus to multiple qubits. Previous version, treated 1- and 2-qubit systems as the "normal" case, and n-qubit systems as the "special" case (evidenced by object and function naming and features, i.e. pyGSTi's API). This version takes a big step toward treating 1- and 2-qubit systems as special cases of a single, more general, n-qubit case.

This update breaks backward compatibility, which can mostly (often completely) be restored by running the included pyGSTi/scripts/update2v0.9.7.py script on old scripts and/or notebooks.

This update also revamps the tutorials, giving them into an intuitive network of hyper-linked notebooks. The updates will make both broadly learning what pyGSTi can do and located detailed guidance of specific topics much easier than it was in past versions. Begin by opening the pyGSTi/jupyter_notebooks/START_HERE.ipynb notebook.

Added

- Adds beta-level implementation of "Idle Tomography" (a new multi-qubit characterization protocol)
- Adds ability to compose gates by adding together their error generators.
- Adds support for exporting Circuit objects to IBM's OpenQASM specification and Rigetti Quantum Computing's Quil specification (more coming soon).
- Adds new "implicit" noise models designed for multi-qubit processors.

Fixed

- Bug fixes (particularly in the computation of gauge-invariant fidelities).

Changed

- the drift-detection sub-module is amidst an upgrade and is not completely functional in this release.

Added

- Updated and expanded data set comparison tools (e.g., more sophisticated hypothesis testing)
- adds Lindblad-parameterized SPAM operators.
- adds more gate & SPAM operator parameterizations, including affine and depolarization errors.

Added

- Direct Randomized Benchmarking (DRB) method. Includes sequence selection and data analysis routines (see new tutorials 17-18).
- Improved Clifford Randomized Benchmarking support (see tutorial 16).
- support for multi-qubit circuits (new Circuit object; see tutorial 15).
- ability to easily create LGST-only reports (see new example notebook).

Changed

- improvements to the DataSet object: you can now store "sparse" counts (where not every outcome is observed) and add meta-data to lines after a '' symbol.
- compatibility with plotly v3.0+ and cvxpy v1.0+.