JustInTimeDecoding

Data-generation used for simulating the JIT decoder in a cubic lattice, including twisted errors induced by the nonabelian stabilizers of the twisted quantum double model and the heralding heuristic proposed in arxiv/2604.02033.

Module layout

• geometry.py: lattice indexing, time-depth rule, last-step measured-edge mask, canonical output filename builder.
• lattice.py: incidence matrix construction, neighbor-edge lookup, vectorization masks, local edge-mask helper.
• decoder.py: logical-error checks and JIT decoding protocol.
• twisted.py: twisted-Z error generation and loop-closing matching builder.
• runner.py: high-level simulation entry points:
  • run_full_simulation
  • run_x_only_simulation
  • gather_effective_length_data
  • run_res_30d_grid
• cli.py: command-line wrapper preserving legacy argument order.

Quick usage

python -m shared_simulation.cli results/xz_errs 9 0.02 0.02 10000 0

Or from Python:

from shared_simulation.runner import run_full_simulation

counters = run_full_simulation(
    linear_size=9,
    px=0.02,
    pz=0.02,
    repetitions=10000,
    output_dir="results/xz_errs",
    boundary="OBC",
    run_id=0,
    use_jit=True,
)
print(counters)