Changelog¶
[Unreleased]¶
A documentation site, at qoroquantum.github.io/deltakit-maestro, built with Sphinx: the README’s overview, a guide, the example notebook with its plots, each example script with the output of a run made by the build, an API reference from the docstrings, and this changelog. The guide holds what the README used to cover in depth: the noise modes, the decoder’s detector error model, coherent noise, the bond dimension, performance and gotchas. The Docs workflow builds the site on every pull request, failing on any Sphinx warning or failing example, and publishes it from
main.pip install -e ".[docs,examples]"installs what the build needs, and the README and the PyPI Documentation link point at the site.
[0.1.0]¶
First release.
MaestroBackendadapter for Riverlane Deltakit QEC circuits and Stim circuits overmaestro.sinter.Every run is simulated by Maestro (
qoro-maestro>=0.3.14), with the simulation type picked from the noise:Pauli-only and noiseless runs on
Stabilizer, which samples them on Pauli frames.T1 decay without coherent noise on
Stabilizer, shot by shot (T1 is not a Pauli channel).Coherent over-rotations on
MatrixProductState(MPS).An explicit
simulation_typeoverrides the choice. The chosen type is pinned (fixed_backend=True), so Maestro does not swap it for Pauli frames; passfixed_backend=Falseto allow that.
Shots come back in random order, so any slice of them is a fair sample.
Noise model translation:
Extracts gate noise, idle noise, reset noise, and readout errors from Deltakit circuits.
Supports Pauli twirling of coherent channels and hardware noise layers for detector error model (DEM) generation.
Sinter integration:
as_sinter_taskproduces Sinter-compatible tasks with matched DEMs and JSON metadata.sinter_decodersprovides configured decoders (including PyMatching).compile_samplerfor direct sampling of detection events and observable flips.
Source distribution packaging includes the full test suite and examples, verified with
scripts/verify_dist.py.Supports Python 3.10 to 3.13 on macOS and Linux.