feat(grammar): W17B vocabulary extension from PWI+waves absorbed-power evidence

[Simon-McIntosh]

Summary

Add 3 physical_base tokens from imas-codex W17B mining of low-scoring standard names
(reviewer_score_name < 0.55, N=86 total analyzed, 0 VocabGap graph nodes — tokens
identified by manual parsing of compound physical_base residues).

Source: imas-codex W17B wave-heating + PWI campaign synthesis.
Evidence gate: N ≥ 3 distinct names (policy: commit 1982016).
Branch: w17b-vocab-extension from main (does not touch PR #11's vocab-rotation-wave7-wave9 branch).

---

Evidence for new vocabulary tokens

Tokens proposed

Token	Segment	Vocab file	N (distinct low-score names)	Sample names
absorbed_wave_power	physical_base	physical_bases.yml	4	fast_electron_absorbed_wave_power_per_toroidal_mode, thermal_ion_absorbed_wave_power_per_toroidal_mode, fast_ion_absorbed_wave_power_per_toroidal_mode, thermal_electron_absorbed_wave_power_per_toroidal_mode
wave_absorbed_power_density	physical_base	physical_bases.yml	5	wave_absorbed_power_density_per_toroidal_mode, fast_electron_wave_absorbed_power_density_per_toroidal_mode, thermal_electron_wave_absorbed_power_density_per_toroidal_mode, fast_ion_wave_absorbed_power_density_per_toroidal_mode, thermal_ion_wave_absorbed_power_density_per_toroidal_mode
wave_absorbed_power_inside_flux_surface	physical_base	physical_bases.yml	3	fast_ion_wave_absorbed_power_inside_flux_surface_per_toroidal_mode, thermal_electron_wave_absorbed_power_inside_flux_surface_per_toroidal_mode, thermal_ion_wave_absorbed_power_inside_flux_surface_per_toroidal_mode

Evidence gathered from imas-codex VocabGap analysis: parsing 86 low-scoring SNs
(reviewer_score_name < 0.55) and extracting unrecognised physical_base residue
strings. N counts number of distinct low-scoring names that produce the given
unrecognised residue when the known subject prefix is stripped.

Why existing tokens do not suffice

• absorbed_wave_power: wave_absorbed_power exists in ISN (canonical word order), but the imas-codex pipeline generates names with absorbed_wave_power word order (adjective-noun vs noun-adjective). This token is the evidence-backed variant form needed to satisfy the open-fallback validator.
• wave_absorbed_power_density: wave_absorbed_power covers volumetric power integrals; wave_absorbed_power_density (W·m⁻³) is the distinct volumetric density form absent from the vocabulary. Five species variants (fast_electron, thermal_electron, fast_ion, thermal_ion, plus bare) all require it.
• wave_absorbed_power_inside_flux_surface: Parallel to the already-present wave_driven_absorbed_power_inside_flux_surface. Three species-qualified names need the base concept of absorbed wave power cumulated inside a flux surface (W), distinguished from the wave_driven_ (current-drive context) and the _density (W·m⁻³) variants.

---

Tokens explicitly rejected (preposition-form ban)

Token	Reason
per_toroidal_mode (postfix)	Grammar policy bans changing existing operator kind; per_toroidal_mode is defined as unary_prefix with test coverage (test_d5_vocabulary.py::test_per_toroidal_mode_power_round_trip). Postfix usage in corpus names is a name-generation bug in imas-codex, not a vocabulary gap.
inside_flux_surface as preposition	inside is a locative preposition; ISN only allows of/at/over locus relations.
per_ style postfixes	Explicitly banned by ISN grammar policy (_per_ is a preposition).

Tokens deferred (N < 3)

Token	Segment	N	Notes
wave_power	physical_base	1	Only fast_electron_wave_power_inside_flux_surface_per_toroidal_mode — just one name.
peak_power_flux	physical_base	2	peak_power_flux_of_inner_divertor_target + outer — just misses gate.
count_rate	physical_base	2	minimum_count_rate_of_neutron_detector_mode + maximum — just misses gate.
bidirectional_reflectance_distribution_function	physical_base	1	Single BRDF name; highly domain-specific. Re-evaluate after PWI coverage expansion.
incident_angle	physical_base	1	Only sputtering_physical_coefficient_incident_angle. sputtering_angle_of_incidence already in vocab; rename preferred.
vessel_annular_outer_outline_point	position	1	Single name; vessel_outline_point already in locus_registry (PR #11).
logarithmic_density_gradient	physical_base	1	Only normalized_logarithmic_density_gradient.

Mining summary

• 86 low-score names analysed (score < 0.55)
• 0 VocabGap nodes in graph (VocabGap node type not populated yet in this codex version)
• 273 compound physical_base residues extracted (after subject/locus stripping)
• 3 tokens meet N ≥ 3 gate and are additive (no grammar-rule changes)
• 1 grammar fix identified but deferred (per_toroidal_mode prefix→postfix would break test_d5_vocabulary.py)

---

Changes

• imas_standard_names/grammar/vocabularies/physical_bases.yml — add absorbed_wave_power, wave_absorbed_power_density, wave_absorbed_power_inside_flux_surface
• imas_standard_names/schemas/entry_schema.json — schema-version string bump (auto-generated)

Testing

• All existing tests pass (1108 passed, 68 xfailed, 0 failed)
• Grammar codegen drift check passes
• New tests not required (vocab-only additions, no grammar-rule changes)

Checklist

• I have read the CONTRIBUTING.md guidelines
• Conventional commit message format used
• N ≥ 3 evidence provided above for all included tokens
• Branch is from main, does not conflict with PR feat(vocab): add evidence-gated tokens from imas-codex rotation waves 7 and 9 #11 (vocab-rotation-wave7-wave9)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ All tests successful. No failed tests found.

📢 Thoughts on this report? Let us know!