stabstream 0.1.0

Python bindings for stabstream — high-performance QEC syndrome stream deserializer

stabstream

High-performance QEC syndrome stream deserializer and real-time Union-Find decoder for Python — built in Rust via PyO3.

Installation

pip install stabstream

Quick start

Zero-config (Stim users)

import stim
import stabstream

circuit = stim.Circuit.from_file("surface_d5.stim")

with stabstream.from_stim_circuit("recording.qssf", circuit) as stream:
    for frame in stream:
        apply_correction(frame.observable_flips)

Explicit DEM

with stabstream.open("recording.qssf", decoder="surface_d5.dem") as stream:
    for frame in stream:
        print(frame.observable_flips)

Async (live hardware, multiple sources)

import asyncio, stabstream

async def main():
    circuit = stim.Circuit.from_file("surface_d5.stim")
    async with stabstream.from_stim_circuit("tcp://fpga:9000", circuit) as stream:
        async for frame in stream:
            await apply_correction(frame.observable_flips)

asyncio.run(main())

Raw NumPy arrays

with stabstream.open("recording.qssf") as stream:
    for frame in stream:
        events = frame.to_numpy_detector_events()  # shape (ancilla_count,), dtype bool
        meas   = frame.to_numpy_meas_results()     # shape (ancilla_count,), dtype int8

Key classes

Class	Description
SyndromeStream	Dual sync+async stream with integrated UF decoder
SyndromeFrame	Single parsed frame — NumPy arrays, observable_flips
SyndromeWindow	Sliding multi-round detector matrix
DetectorErrorModel	Stim DEM parser + PyMatching bridge
LogicalErrorAccumulator	Lock-free p_L accumulator
StabstreamError	Exception raised on decode/stream errors

Source

github.com/reggswilgoxson/stabstream