Group singleshot emulator result extraction

[DennisWayo]

Summary

This is a follow-up to qiboteam/qibolab #1505 and addresses the singleshot result-handling cleanup described in qiboteam/qibolab #1223.

The change extracts measured subsystem states by grouping acquisitions that share the same sampled full-system state. For each unique measurement sample, the full-system state is unraveled once, then each acquisition reads only the subsystem index it needs.

This replaces the previous diagonal-style extraction through a larger temporary array:

np.stack(np.unravel_index(sampled[inverse_map], dims))[qubit_indices]

with a more direct helper:

_sampled_measurements(sampled, dims, inverse_map, qubit_indices)

The public result shape and clipping behavior are unchanged.

Context

#1505 simplified the cyclic/non-shot probability path. This PR keeps the next result-handling step narrow by touching only the singleshot extraction path.

I kept this independent from #1413, because #1413 still depends on the confusion-matrix path from draft qiboteam/qibolab #1401.

Follow-Up

If this PR lands, my next planned step is to use the merged Qibocal emulator platforms from qiboteam/qibocal #1551 to run Qibocal workflows with the Dynamiqs emulator engine and report concrete validation findings. That follow-up does not depend on #1401 or #1413 (unless @alecandido wants us to take a second look); those remain separate result/confusion-matrix work.

Validation

• pytest tests/instruments/emulator/test_results.py -q
• pytest tests/instruments/emulator/test_emulator.py tests/instruments/emulator/test_results.py -q
• pytest tests/instruments/emulator -q
• ruff check src/qibolab/_core/instruments/emulator/results.py tests/instruments/emulator/test_results.py
• ruff format --check src/qibolab/_core/instruments/emulator/results.py tests/instruments/emulator/test_results.py
• Qibocal Rabi smoke on merged Qibocal platform: qubit acquisition + fit passed
• Qibocal Rabi smoke on merged Qibocal platform: qutrit acquisition + fit passed

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 39.47%. Comparing base (326d9a1) to head (960c348).
⚠️ Report is 19 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1509      +/-   ##
==========================================
+ Coverage   39.41%   39.47%   +0.05%     
==========================================
  Files         115      115              
  Lines        6152     6158       +6     
==========================================
+ Hits         2425     2431       +6     
  Misses       3727     3727              

Flag	Coverage Δ
unittests	39.47% <100.00%> (+0.05%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull request overview

This PR refactors the emulator single-shot result extraction path to group acquisitions that share the same sampled full-system state, reducing repeated unraveling work while preserving the public result shape and clipping behavior.

Changes:

• Introduces _sampled_measurements to extract measured subsystem states by grouping acquisitions by shared sampled state.
• Updates _singleshot_results to use _sampled_measurements instead of the previous diagonal-style extraction through a larger temporary array.
• Adds a focused unit test validating the acquisition grouping/extraction behavior.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
src/qibolab/_core/instruments/emulator/results.py	Adds _sampled_measurements and switches single-shot extraction to use it.
tests/instruments/emulator/test_results.py	Adds a unit test for grouped single-shot measurement extraction.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[DennisWayo]

Thanks, this is a fair point. I applied the preallocation change manually rather than accepting the suggestion directly, so I could keep the patch aligned with the existing helper and rerun the local checks before pushing.

_sampled_measurements() now preallocates the output array and fills it per sample, avoiding the intermediate dictionary/stack pass. I also added a small guard for mismatched inverse_map/indices lengths and a focused test for that case.

cc @alecandido

[lballerio]

good for me

[DennisWayo]

Agreed @lballerio, thanks for catching this. I removed the guard and its dedicated test since inverse_map and indices are both built from the acquisition order in _singleshot_results(). The helper still keeps the preallocated result array and fills it per sampled state.

[lballerio]

the comment before was not critical, now the uninfluent check has been removed, for me is OK