§1 — Problem (fully defined)

The thermal-fusion neutron-rate StandardName family is internally inconsistent. All members carry unit Hz (a count rate, s⁻¹), but they disagree on physical_base and the structural tree is broken. Current live/superseded members:

• total_neutron_source_rate_due_to_thermal_fusion — base=source_rate, dd-sourced, accepted. This is the canonical name (the correct base for a Hz neutron rate).
• deuterium_deuterium_neutron_flux_due_to_thermal_fusion — base=flux, derived, accepted; structural parent = neutron_flux_due_to_thermal_fusion.
• tritium_tritium_neutron_flux_due_to_thermal_fusion — base=flux, derived, accepted; structural parent = neutron_flux_due_to_thermal_fusion.
• neutron_flux_due_to_thermal_fusion — base=flux, superseded. This is the structural PARENT of the two derived children above — but it is dead.
• flux_due_to_thermal_fusion — base=flux, manual, superseded onto the canonical during sn-pedestal-source-recovery (a redundant duplicate; already retired).

Two defects: (a) the two per-reaction children use base=flux, but "flux" denotes an areal density (m⁻²·s⁻¹); a total count rate in Hz is a source_rate, as the canonical name correctly uses. (b) Their structural parent neutron_flux_due_to_thermal_fusion is superseded, so two live derived names hang off a dead parent — an orphaned derivation the structural fixpoints cannot re-anchor.

§2 — Why it matters (motivation)

Mixed physical_base for one physical concept (a neutron count rate) means the same quantity is named two different ways depending on whether it is the total or a per-reaction channel — the catalog reads as if "flux" and "source_rate" are different physics when they are the same dimension. The dead structural parent is a latent ledger/graph inconsistency: a live derived name whose HAS_PARENT target is superseded cannot be regenerated or validated cleanly, and breaks the "derived child ⇒ live parent" expectation. It is small (2 live names) but a clear correctness/consistency fix. Surfaced by sn-pedestal-source-recovery (f-psr-006) while retiring the redundant manual flux_due_to_thermal_fusion.

§3 — The decision + approach (reference)

The concept and correct base already exist (source_rate, on the canonical total). The fix aligns the per-reaction children to it and restores a live parent:

• Rename flux→source_rate for the two children via the rename cascade: deuterium_deuterium_neutron_flux_due_to_thermal_fusion → deuterium_deuterium_neutron_source_rate_due_to_thermal_fusion (and the T-T analogue), validating that each target is ISN-grammar-valid and round-trips.
• Restore a live parent — the renamed children strip to a neutron_source_rate_due_to_thermal_fusion parent; ensure that parent is live (re-derived by the structural fixpoint, or the canonical total_neutron_source_rate_due_to_thermal_fusion serves as the group parent). Decide the exact parent below.
• Leave the already-superseded flux_due_to_thermal_fusion and neutron_flux_due_to_thermal_fusion as-is (retired).

Mechanism: the deterministic rename cascade (sn run --rename OLD:NEW --include-accepted, --dry-run first), never a --focus recompose. Verify 0 orphans / 0 desyncs and that the live children point at a live parent.

Project: imas-standard-names (grammar) + imas-codex (standard_names) + imas graph
Plan: sn-fusion-neutron-rate-naming (http://localhost:8765/imas-standard-names/sn-fusion-neutron-rate-naming.html)
Section: rename flux children to source_rate + restore live parent
Tier: sonnet

Context
  total_neutron_source_rate_due_to_thermal_fusion (base=source_rate, dd, accepted) is canonical. Two
  accepted DERIVED children — deuterium_deuterium_neutron_flux_due_to_thermal_fusion and
  tritium_tritium_neutron_flux_due_to_thermal_fusion — still use base=flux [Hz] and hang off the
  SUPERSEDED parent neutron_flux_due_to_thermal_fusion. Honour the fusion-child-parent decision shown
  live above before acting.

State to read (CODE / FILES / DATA)
  imas-codex graph_ops.py: _execute_rename_cascade, HAS_PARENT/HAS_SUCCESSOR, structural rederive
  (seed_parent_sources / rederive_structural_edges). CLI: sn run --rename OLD:NEW --include-accepted
  (--dry-run first). ISN grammar: parse_standard_name / render.compose must accept + round-trip the
  targets (verified: deuterium_deuterium_neutron_rate_due_to_thermal_fusion parses, base=rate). Graph:
    MATCH (n:StandardName) WHERE n.id CONTAINS 'due_to_thermal_fusion'
    OPTIONAL MATCH (n)-[:HAS_PARENT]->(p) RETURN n.id,n.physical_base,n.unit,n.name_stage,p.id

Scope locks / constraints (non-decision)
  - Rename cascade only, NEVER --focus recompose. Every target ISN-grammar-valid + round-tripping.
  - Preserve provenance (derived sources) and content; keep 0 orphans / 0 desyncs.
  - Do not resurrect the superseded flux_due_to_thermal_fusion / neutron_flux_due_to_thermal_fusion;
    the children must end up under a LIVE source_rate parent.
  - Note: 'source_rate' is the valid base; 'rate' also parses — pick the base consistent with the
    canonical total_neutron_source_rate_due_to_thermal_fusion.

Done-when
  1. The two per-reaction children carry base=source_rate (name aligned to the canonical), grammar-valid.
  2. They hang off a LIVE structural parent (per the decision).
  3. 0 orphans / 0 desyncs; provenance preserved.
  4. This followup resolved.