This repository contains the statistical source data and plotting scripts needed to reproduce the main-text figures of the paper. For every figure panel there is (1) a plain-text/CSV dataset and (2) a self-contained Python script that reads the dataset and regenerates the panel.
The figshare deposit (figshare_upload/) contains only the main-text figure
source data (Figs 1–4: panels 1c, 2a, 2b, 2c, 2e, 3b, 3c, 4a, 4c, 4d, 4e, 4f),
as open-format CSV/TXT. It deliberately excludes analysis code, supplementary-figure
data, and published figure images.
The analysis code, notebooks, and supplementary material live in this GitHub
repository and are referenced from the paper's Code Availability statement. Data
underlying several supplementary figures depends on raw .h5 measurement files that
are not redistributed here.
Suggested statements:
- Data availability → figshare deposit (main-text source data).
- Code availability → this GitHub repository.
fig1/ Fig 1(c) mode frequency spectrum & coherences
fig2/ Fig 2(a) f0g1 (Q<->B) Rabi trace [pending raw trace]
Fig 2(b) B<->S2 beam-splitter swap trace [pending raw trace]
Fig 2(c) beam-splitter RB, all modes M1-S1..7
Fig 2(e) per-mode SWAP infidelity bar chart
fig3/ Fig 3(b) random read fidelity vs # accesses (7 modes + ref) [done]
Fig 3(c) random read infidelity vs RAQM size [done]
fig4/ Fig 4(a) cross-Kerr matrix [done]
Fig 4(c) state-dependent access error matrix [done]
Fig 4(d) spectator-access dephasing matrix [done]
Fig 4(e) many-body dephasing rate per mode [done]
Fig 4(f) random-read error budget [done]
All data panels are reproduced and verified against the published values. Pending only: Fig 2(a) and Fig 2(b) raw time traces (awaiting the source files).
(Schematic / circuit / device panels — 1a,b, 2d, 3a, 4b — contain no statistical data and are not included.)
raqm_style.py defines a shared muted/earthy color theme (cream background;
terracotta, teal, slate-blue, gold) used by all plotting scripts — an
alternative palette to the one in the published figures. Edit it to restyle
every panel at once.
SOURCES.md— per-panel map of repo dataset → slab-drive source file → analysis notebook.analysis/— the actual fitting/extraction functions (RB fits, gate-fidelity conversion, dual-rail parity populations) used to produce the figure data.notebooks/— reference copies of the analysis notebooks (as run).
python >= 3.9, numpy, pandas, matplotlib, scipy
cd fig2 && python plot_fig2c.py # writes fig2c_M1-S{1..7}.pdf/.pngEach script writes a .pdf and .png next to itself. Datasets are CSV with a
header row; column meanings are documented at the top of each plotting script
and in the per-figure README.md.