• Home
  • Features
  • Pricing
  • Docs
  • Announcements
  • Sign In

sandialabs / LoQS / 43545370531
90%

Build:
DEFAULT BRANCH: main
Ran 14 Nov 2026 02:29AM UTC
Jobs 1
Files 90
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

14 Nov 2026 02:16AM UTC coverage: 89.996% (+0.5%) from 89.494%
43545370531

push

github

web-flow
Add `StimEngine` for batched STIM runs of programs (#177)

* Add STIM migration helpers and fix STIM conversion gaps

STIM users had to write the model, record and decoder by hand, and some
STIM constructs were dropped or misrecorded; this adds helpers for the
three and fixes those gaps.
- Add `loqs.tools.stimtools.build_stim_model`: the ideal or
  `generated`-style Pauli-noisy model for the STIM or dense states.
- Add `stim_record_from_outcomes`, which rebuilds STIM's measurement
  record, detectors and observables from LoQS outcomes.
- Add `build_stim_decoder_instruction`, a PyMatching DEM decoder
  instruction that survives a program's save/load round trip.
- Honour `M(p)`-style measurement arguments on the STIM state and refuse
  them elsewhere, instead of warning and dropping them.
- Record `ZBasis*` measurements in `applied_stim_circuit_str`, stop
  pre/post ops wiping it, and read it from the new state.
- Let a physical-circuit instruction run with no `PatchLayout` unless it
  updates the Pauli frame.
- Minor: the noise-line warning shows once; tests for each fix and
  helper; flake8 per-file ignore for the new test file.

Co-authored-by: claude-sonnet-5-5 <noreply@anthropic.com>
Co-authored-by: claude-opus-5-5 <noreply@anthropic.com>

* Add StimEngine for batched STIM sampling of programs

Add an engine that traces one shot of a STIM-state program and samples
every shot with stim, returning per-shot arrays instead of histories.
- loqs.engines.StimEngine: compile() traces a shot on a private state
  that records internal-index text; run() samples and converts records.
- compile() refuses what the trace can't reproduce (other states,
  classical steps, per-shot draws, rebuilt states, truncated traces).
- loqs.core.CompiledProgram: plain, serializable steps keyed by the
  measurement record, each carrying its emitted text per target.
- loqs.core.OutputResults: backend-neutral per-shot outputs stored as
  values plus offsets, so ragged records sav... (continued)

4212 of 4952 branches covered (85.06%)

Branch coverage included in aggregate %.

13447 of 14670 relevant lines covered (91.66%)

0.92 hits per line

Coverage Regressions

Lines Coverage ∆ File
10
94.88
-0.22% backends/model/dictmodel.py
4
87.38
2.34% core/instructions/builders.py
2
66.67
-8.33% tools/__init__.py
1
93.84
0.11% backends/state/stimstate.py
1
92.05
0.16% core/history.py
Jobs
ID Job ID Ran Files Coverage
1 43545370531.1 14 Nov 2026 02:29AM UTC 90
90.0
GitHub Action Run
Source Files on build 43545370531
  • Tree
  • List 90
  • Changed 14
  • Source Changed 0
  • Coverage Changed 14
Coverage ∆ File Lines Relevant Covered Missed Hits/Line Branch Hits Branch Misses
  • Back to Repo
  • Github Actions Build #43545370531
  • 084ce57e on github
  • Prev Build on main (#43287455094)
STATUS · Troubleshooting · Open an Issue · Sales · Support · CAREERS · ENTERPRISE · START FREE TRIAL · SCHEDULE DEMO
ANNOUNCEMENTS · TWITTER · TOS & SLA · Supported CI Services · What's a CI service? · Automated Testing

© 2026 Coveralls, Inc