PORTABLE INDEPENDENT ITERATION06 REVIEW AND NUMERICAL FIGURES

Expected extracted bundle layout:

  efficiency/       frozen numerical sources and results
  reviews-r06/      these scripts, reviews, figure data and figures

Run from the extracted bundle root with Python and NumPy installed:

  python reviews-r06/independent_review.py
  python reviews-r06/review_matrix.py

The scripts detect the sibling efficiency directory. For a different layout,
pass its location explicitly, for example:

  python reviews-r06/review_matrix.py --efficiency-root path/to/efficiency

The independent engine review uses 60-digit standard-library Decimal
references, composite Simpson modal normalization, passive contact power and
energy checks, and a perturbed replay of the original three-mode engine.
The completed-matrix review verifies all 72 rows, their frozen inputs/source
hashes, budgets, commands, time resolution and reconstructed final energy.
It also records post-run conservative spatial displacement diagnostics.
Those diagnostics do not change the original study's acceptance rules.

Optional figure regeneration requires Matplotlib:

  python -m pip install matplotlib
  python reviews-r06/plot_iteration06.py

Matplotlib is needed ONLY to regenerate figures. The app and numerical model
do not depend on it. The static SVG and PNG figures open without Matplotlib.
The figures were rendered with Matplotlib 3.11.2. Regeneration in a different
rendering environment may produce different image hashes even when the exact
scientific data are unchanged; the source data and provenance are provided.

NUMERICAL_iteration06_ratios.svg/png shows all six selected conditions across
3, 5 and 7 retained modes. NUMERICAL_iteration06_peak_energies.svg/png uses a
common zero-based mJ axis in all six panels. Filled markers/lines are the
2.5-microsecond results; open circles are 5-microsecond results, often visually
coincident. They are deterministic outputs, not statistical error bars.

NUMERICAL_iteration06_figure_data.csv contains every plotted value at both
steps. FIGURE_PROVENANCE.json binds the study summary, 72 source rows, review,
plotting script, exact data and rendered figures by SHA-256. The scripts check
their source evidence before plotting. A failed hash or incomplete matrix is
not silently accepted. PORTABILITY-VERIFICATION.json records the isolated
bundle-layout reproduction check and PNG equality in the tested environment.

Re-running replaces generated local review/figure files and changes date and
runtime metadata. Preserve a snapshot first if the delivered evidence identity
must be retained. Frozen efficiency sources and results are read, never edited.

Interpretation: six selected numerical beam conditions, not six independent
physical experiments or a new performance probability. The controller commands
remain fixed. Adding beam modes does not validate fracture, material/contact
calibration, whole-device efficiency, throughput, or diamond/boulder capability.
SCIENTIFIC-ACCEPTANCE-ROADMAP.txt explains proposed next validation gates and
links to primary sources. These scripts perform no hardware control, network
request or publication. No additional dependency folder belongs in the bundle.
