Harden qet-bench publication release candidate

.gitignore

@@ -23,8 +23,10 @@ dist/
 # Generated benchmark artifacts
 qet-bench/results/
 
+# Private planning notes outside the publishable package tree
+/docs/
+
 # Editor and OS noise
 .DS_Store
 .idea/
 .vscode/
-

README.md

@@ -0,0 +1,44 @@
+# qet-bench
+
+This repository hosts `qet-bench`, a reproducible benchmark suite for the
+minimal two-qubit Quantum Energy Teleportation (QET) protocol.
+
+The package lives in [`qet-bench/`](qet-bench/). It implements the exact
+Hamiltonian model, Alice measurement, classical feedforward, Bob rotation,
+energy ledger, null tests, count estimators, deterministic artifact scripts,
+an independent pure-Python reference branch, and an optional Qiskit/Aer
+interoperability bridge.
+
+The scientific scope is intentionally narrow: `qet-bench` demonstrates
+Bob-side local energy extraction from ground-state correlations while keeping
+the global ledger explicit. It does not claim autonomous vacuum power or a
+hardware energy source.
+
+## Quickstart
+
+```bash
+cd qet-bench
+python -m pip install -e ".[dev]" -c requirements-lock.txt
+python -m pytest
+python scripts/reproduce_benchmarks.py
+python scripts/make_all_figures.py
+```
+
+Optional Qiskit/Aer smoke check:
+
+```bash
+cd qet-bench
+python -m pip install -e ".[qiskit]" -c requirements-lock.txt
+python scripts/run_qiskit_bridge.py
+```
+
+## Review Materials
+
+- [Package README](qet-bench/README.md)
+- [Artifact manifest](qet-bench/docs/artifact_manifest.md)
+- [Release validation log](qet-bench/docs/release_validation.md)
+- [Research roadmap register](qet-bench/docs/research_roadmap.md)
+- [Blocker ledger](qet-bench/docs/blocker_ledger.md)
+- [Manuscript draft](qet-bench/paper/paper.md)
+- [ReScience-oriented manuscript](qet-bench/paper/rescience.md)
+- [Manuscript build notes](qet-bench/paper/README.md)

qet-bench/CITATION.cff

@@ -1,20 +1,20 @@
 cff-version: 1.2.0
 message: "If you use qet-bench, please cite this software."
-title: "qet-bench: Exact two-qubit Quantum Energy Teleportation benchmarks"
-version: "0.2.0-noise-atlas"
+title: "qet-bench: Reproducible exact two-qubit Quantum Energy Teleportation benchmarks"
+version: "0.2.0"
 date-released: 2026-04-16
 authors:
   - family-names: "Gueguen"
     given-names: "Francois"
     affiliation: "Lucent Lab"
 license: MIT
-abstract: "A reproducible benchmark suite for minimal two-qubit Quantum Energy Teleportation with explicit energy accounting."
+abstract: "A reproducible benchmark suite for minimal two-qubit Quantum Energy Teleportation with explicit ground-state energy accounting."
 repository-code: "https://github.com/lucent-lab/qet-bench"
 url: "https://github.com/lucent-lab/qet-bench"
 preferred-citation:
   type: software
-  title: "qet-bench: Exact two-qubit Quantum Energy Teleportation benchmarks"
-  version: "0.2.0-noise-atlas"
+  title: "qet-bench: Reproducible exact two-qubit Quantum Energy Teleportation benchmarks"
+  version: "0.2.0"
   date-released: 2026-04-16
   authors:
     - family-names: "Gueguen"

qet-bench/README.md

@@ -45,8 +45,13 @@ See [docs/artifact_manifest.md](docs/artifact_manifest.md) for the full
 CSV/figure provenance table. See
 [docs/research_roadmap.md](docs/research_roadmap.md) for the maintained
 roadmap table that separates reproduction work from new research territory and
-records milestone tags. See [docs/blocker_ledger.md](docs/blocker_ledger.md)
-for the maintained review/submission blocker ledger.
+records milestone tags. See
+[docs/release_validation.md](docs/release_validation.md) for the current local
+validation log while hosted GitHub Actions runners are unavailable. See
+[docs/vacuum_investigation_roadmap.md](docs/vacuum_investigation_roadmap.md)
+for the step-by-step path from QET benchmarks toward deeper vacuum-energy
+accounting audits. See [docs/blocker_ledger.md](docs/blocker_ledger.md) for
+the maintained review/submission blocker ledger.
 
 ## Optional Qiskit/Aer Bridge
 

qet-bench/docs/artifact_manifest.md

@@ -10,8 +10,8 @@ except `benchmarks.csv`, which is generated by
 ## Generated-Artifact Policy
 
 The committed `results/` files are release artifacts for reviewer convenience:
-they are generated outputs, but they are intentionally tracked in the `v0.1.0`
-snapshot so the benchmark values and figures can be inspected without rerunning
+they are generated outputs, but they are intentionally tracked in release
+snapshots so the benchmark values and figures can be inspected without rerunning
 the package. The source of truth remains the Python package, scripts, tests, and
 pinned install path. Regenerating the artifacts with the commands below should
 produce no git diff:

qet-bench/docs/blocker_ledger.md

@@ -20,12 +20,18 @@ Reference policy:
 - Prefer DOI, Software Heritage ID, hosted CI run URL, or paper DOI when
   available.
 
+Current release identity:
+
+- `v0.2.0` is the active publication-hardening review target.
+- `0.2.0` is the package and citation metadata version.
+- Earlier `v0.2.0-*` references are historical milestone tags.
+
 | Blocker | Why it matters | Status | Scope | Resolution / Next Action | Reference |
 |---|---|---|---|---|---|
 | Missing exact benchmark implementation and regression fixtures | Without exact values, the package cannot serve as a reproducible QET benchmark. | Closed | Core software | Implemented exact Hamiltonian, ground state, QET ledger, benchmark CSV, figures, and regression tests. | `v0.1.0` |
 | No explicit no-free-energy accounting guardrail | QET is easy to overstate if Alice injection and global accounting are omitted. | Closed | Scientific framing | README, docs, tests, and paper state that Bob extraction is paid for by Alice measurement/control resources and check `E_A >= E_B` for the exact benchmark. | `v0.1.0` |
 | Missing artifact provenance | Reviewers need to know which script generated each CSV/figure and what each artifact validates. | Closed | Reproducibility | Added artifact manifest mapping outputs to generators, validation coverage, and supported claims. | `v0.1.0` |
-| Hosted CI unavailable because of quota | Hosted CI gives reviewers an independent, visible validation path; local-only validation is weaker. | Open | Review infrastructure | Re-run GitHub Actions when quota permits and record the run URL or badge. Keep local Python 3.11/3.12 CI as the interim validation record. | pending |
+| Hosted CI unavailable because of quota | Hosted CI gives reviewers an independent, visible validation path; local-only validation is weaker. | Open | Review infrastructure | Hosted jobs currently fail before execution because account billing/spending limit blocks runners. Re-run GitHub Actions when quota permits and record the successful run URL or badge. Keep local Python 3.11/3.12 CI as the interim validation record. | blocked run <https://github.com/lucent-lab/qet-bench/actions/runs/24507161016> |
 | Release archive DOI or Software Heritage ID missing | A GitHub tag is useful, but archival review generally needs a persistent citation target. | Open | Publication metadata | Archive the selected review snapshot, then update `CITATION.cff`, `paper.md`, `paper/README.md`, `research_roadmap.md`, and this ledger. | pending |
 | Repository access for external review | Reviewers cannot inspect a private repository unless explicit access is granted. | Open | Review logistics | Make the repository public or grant reviewer access before submission. | pending |
 | Placeholder software citation metadata | Citation metadata must identify a real citable artifact and author record. | Mitigated | Publication metadata | Replaced placeholder contributor author with provisional author metadata and added preferred software citation fields. Add DOI once archived. | `v0.2.0-paper-review` |
@@ -38,9 +44,9 @@ Reference policy:
 | Noise-survival wording could imply universal thresholds | Fixed-protocol simulated diagnostics are not theorem-level thresholds or hardware phase boundaries. | Mitigated | Noise diagnostics | Added anchor-aware atlas wording, interval-average definitions, and repeated non-universal/fixed-protocol language. | `v0.2.0-noise-atlas` |
 | Roadmap status was not maintained in a single table | Without a maintained status table, reproduction work and new research territory blur over time. | Closed | Project management | Added `research_roadmap.md` with stage status, classification, deliverable, and release reference. | `v0.2.0-roadmap-register` |
 | Blockers were only reported in chat, not tracked in-repo | Chat-only blocker lists are easy to lose and cannot be audited per release. | Processing | Project management | Maintain this blocker ledger and update it when a blocker changes status or is resolved by a tag, commit, DOI, run URL, or paper reference. | pending |
-| Optional Aer smoke artifact with Qiskit installed missing | The optional bridge has no positive-path artifact unless Qiskit/Aer is installed and run. | Processing | Interoperability | Run the optional Qiskit extra in a disposable environment and add or document the smoke artifact if stable. | pending |
+| Optional Aer smoke artifact with Qiskit installed missing | The optional bridge has no positive-path artifact unless Qiskit/Aer is installed and run. | Mitigated | Interoperability | Local Python 3.12.8 smoke passed with `qiskit==2.3.1` and `qiskit-aer==0.17.2`; `qiskit_interop.md` records the generated CSV path and observed absolute `E_B` errors. Decide during archival whether to force-add optional generated artifacts. | pending merge/tag |
 | JOSS public-history / external-impact evidence is thin | Some venues screen for sustained public development and evidence of research use. | Open | Venue fit | Prefer ReScience-style reproducibility review first, or wait for public history/adoption before JOSS submission. | pending |
-| ReScience-specific reproduced-result framing missing | ReScience-style review needs a crisp statement of the original result being reproduced. | Processing | Manuscript | Adapt the JOSS-style source into a ReScience-oriented submission package naming the reproduced QET benchmark and artifacts. | pending |
+| ReScience-specific reproduced-result framing missing | ReScience-style review needs a crisp statement of the original result being reproduced. | Mitigated | Manuscript | Added `paper/rescience.md`, which names the reproduced two-qubit QET result, artifact bundle, reproduction commands, and interpretation limits. Final author metadata and public repository access remain separate blockers. | pending merge/tag |
 | Full microscopic controller/environment ledger missing | Conservative bookkeeping is not a microscopic hardware or thermodynamic environment model. | Open | New methods research | Design and implement an explicit controller, memory, pulse, reset, and environment ledger before making stronger thermodynamic claims. | none |
 
 ## Maintenance Checklist

qet-bench/docs/qiskit_interop.md

@@ -47,3 +47,18 @@ The optional Aer script is therefore a simulator-interoperability diagnostic,
 not a replacement for the exact ledger and not a hardware execution protocol.
 As in the rest of the package, no result is interpreted as free energy: Bob's
 local extraction is reported against the Alice-side measurement/control ledger.
+
+## Local Smoke Record
+
+For the `v0.2.0` review target, the optional bridge was smoke-tested locally on
+Python 3.12.8 with `qiskit==2.3.1` and `qiskit-aer==0.17.2`:
+
+```bash
+python -m pip install -e ".[qiskit]" -c requirements-lock.txt
+python scripts/run_qiskit_bridge.py
+```
+
+The script wrote `results/data/qiskit_bridge_benchmarks.csv` with absolute
+`E_B` errors of approximately `0.01553` for `(h=1, k=0.5)` and `0.00279` for
+`(h=1, k=1)`. This CSV is generated under the ignored `results/` tree unless a
+future release deliberately archives optional interoperability artifacts.

qet-bench/docs/release_validation.md

@@ -0,0 +1,37 @@
+# Release Validation Log
+
+This log records local validation for review snapshots while hosted GitHub
+Actions jobs are unavailable. Hosted CI remains the preferred independent
+validation path once account billing/quota allows runners to start.
+
+## v0.2.0 Review Target
+
+Date: 2026-04-16
+
+Hosted CI status:
+
+- Workflow: `.github/workflows/ci.yml`
+- Representative hosted run checked:
+  <https://github.com/lucent-lab/qet-bench/actions/runs/24507161016>
+- Result: jobs did not start because the account billing/spending limit blocked
+  hosted runners.
+
+Local CI mirror:
+
+| Python | Environment | Commands | Result |
+|---|---|---|---|
+| 3.11.6 | `/tmp/qet-bench-py311` | `pip install -e ".[dev]" -c requirements-lock.txt`; `pytest`; `scripts/reproduce_benchmarks.py`; `scripts/make_all_figures.py` | `40 passed, 1 skipped`; benchmark and figure scripts completed with no tracked artifact diff. |
+| 3.12.8 | `.venv` | `pip install -e ".[dev,qiskit]" -c requirements-lock.txt`; `pytest`; `scripts/reproduce_benchmarks.py`; `scripts/make_all_figures.py`; `scripts/run_qiskit_bridge.py`; Pandoc citation smoke build | `39 passed, 2 skipped`; benchmark and figure scripts completed with no tracked artifact diff; optional Qiskit/Aer bridge completed; Pandoc wrote `/tmp/qet-bench-manuscript.html`. |
+
+Optional Qiskit/Aer bridge:
+
+- `qiskit==2.3.1`
+- `qiskit-aer==0.17.2`
+- Generated file: `results/data/qiskit_bridge_benchmarks.csv`
+- Absolute `E_B` errors:
+  - `(h=1, k=0.5)`: approximately `0.01553`
+  - `(h=1, k=1)`: approximately `0.00279`
+
+The optional bridge artifact lives under the ignored `results/` tree unless a
+future release deliberately force-adds optional interoperability outputs.
+

qet-bench/docs/research_roadmap.md

@@ -4,7 +4,10 @@ This register tracks the project roadmap at publication granularity. It should
 be updated whenever a stage changes status, a release tag is cut, or an archive
 DOI becomes available.
 
-For review risks and submission blockers, see `blocker_ledger.md`.
+For the step-by-step path from the QET benchmark into broader void/vacuum
+energy-accounting work, see `vacuum_investigation_roadmap.md`. For review risks
+and submission blockers, see `blocker_ledger.md`. For the current local
+validation record while hosted CI is unavailable, see `release_validation.md`.
 
 Status values:
 
@@ -18,23 +21,30 @@ Reference policy:
 - Add an archive DOI, Software Heritage ID, or paper DOI when available.
 - For active work without a tag, use "pending" and replace it when merged.
 
+Current release identity:
+
+- `v0.2.0` is the active publication-hardening review target.
+- `0.2.0` is the package and citation metadata version.
+- Earlier `v0.2.0-*` entries in the table are historical milestone tags.
+
 | Stage | Status | Reproduction Or New Territory | Deliverable | Reference |
 |---|---|---|---|---|
 | Exact two-qubit QET benchmark package | Done | Reproduction | Minimal exact simulator, Hamiltonians, ledgers, null tests, count estimators, deterministic benchmark scripts, generated artifacts, artifact manifest. | `v0.1.0` |
-| Repository hygiene and local CI path | Done | Reproduction hardening | `.gitignore`, GitHub Actions workflow, pinned requirements lock, local CI recipe for Python 3.11 and 3.12. Hosted CI remains quota-blocked until re-run. | `v0.1.0` |
+| Repository hygiene and local CI path | Done | Reproduction hardening | `.gitignore`, GitHub Actions workflow, pinned requirements lock, local CI recipe for Python 3.11 and 3.12. Hosted CI remains billing/quota-blocked until re-run. | `v0.1.0` |
 | Controller accounting extension | Done | Accounting extension; not new microscopic physics | Conservative nonnegative controller-cost ledger and documentation. Reports diagnostic net battery gain without claiming a hardware battery model. | `v0.1.0` |
 | Manuscript release provenance | Done | Publication hardening | Paper build notes, release-provenance wording, local verification record, and submission-risk notes. | `v0.2.0-manuscript` |
 | Independent pure-Python reference branch | Done | Reproduction cross-check | Branch-amplitude implementation independent of the NumPy matrix propagation path, with benchmark and branch-audit artifacts. | `v0.2.0-manuscript` |
 | Optional Qiskit/Aer interoperability bridge | Done | Reproduction/interoperability | Lazy optional Qiskit/Aer adapter, count-key canonicalizer, branch-resolved ideal-circuit script, docs, and tests that pass without Qiskit installed. | `v0.2.0-qiskit-bridge` |
 | Fixed-protocol noise-survival atlas | Done | New simulation diagnostic; not a universal threshold result | Anchor-aware survival curves and summaries across a predeclared `(h, k)` set, interval-averaged survival ratios, and small-multiples figures. | `v0.2.0-noise-atlas` |
 | JOSS-style manuscript hardening | Done | Publication hardening | JOSS-shaped `paper.md`, stricter venue notes, claim-to-artifact map, quality-control table, expanded citation metadata, and contribution guidelines. | `v0.2.0-paper-review` |
 | Public archive and DOI | Processing | Publication hardening | Archive the chosen review snapshot with Zenodo, Software Heritage, or equivalent; update `CITATION.cff`, `paper.md`, and this register with DOI/SWHID. | pending |
-| Hosted CI revalidation | Processing | Reproduction hardening | Re-run GitHub Actions once quota permits; record the run URL or badge target in README/paper notes. | pending |
-| Optional Aer smoke artifact with Qiskit installed | Processing | Reproduction/interoperability | Run `pip install -e ".[qiskit]"` in a disposable environment and save or document the Aer bridge comparison artifact if the dependency stack is stable. | pending |
-| ReScience-style reproducibility package | Processing | Reproduction publication | Adapt the manuscript to a ReScience/comparable reproducibility venue, explicitly naming the reproduced prior result and artifact bundle. | pending |
+| Hosted CI revalidation | Processing | Reproduction hardening | Re-run GitHub Actions once billing/quota permits; recent hosted jobs were blocked before execution, so `release_validation.md` records the local Python 3.11/3.12 validation substitute. | pending; blocked run <https://github.com/lucent-lab/qet-bench/actions/runs/24507161016> |
+| Optional Aer smoke artifact with Qiskit installed | Processing | Reproduction/interoperability | Local Python 3.12.8 smoke passed with `qiskit==2.3.1` and `qiskit-aer==0.17.2`; document the generated comparison artifact and close after merge/tag. | pending |
+| ReScience-style reproducibility package | Processing | Reproduction publication | `paper/rescience.md` frames the work as an exact reproduction of the two-qubit QET benchmark with validation artifacts and bounded novelty claims; close after merge/tag and final author metadata confirmation. | pending |
 | Full microscopic controller/environment ledger | Not done | New methods territory | Replace conservative bookkeeping diagnostics with an explicit modeled controller, memory reset, pulse, and environment exchange ledger. | none |
 | Many-qubit QET extension | Not done | New physics/simulation territory | Extend beyond two qubits to chains or network geometries, with explicit local-energy partitions and finite-size/scaling diagnostics. | none |
 | Finite-temperature or mixed-state QET | Not done | New physics/simulation territory | Study thermal or mixed resource states and compare extraction/accounting behavior against the pure ground-state benchmark. | none |
+| DCE/Casimir/ultrastrong-coupling audit module | Not done | New vacuum-mechanism audit territory | Add a separate module that tracks pump, switching, reset, and environment work for one non-QET vacuum-adjacent mechanism. | none |
 | Hardware or hardware-faithful execution | Not done | New experimental/hardware-simulation territory | Define a circuit/backend protocol, calibration assumptions, count-processing path, and bounded interpretation for real or hardware-faithful execution. | none |
 
 ## Maintenance Checklist