Skip to content

Propagate Raman invariant fix to atom-Fourier interpolation - #34

Merged
mesonepigreco merged 1 commit into
atom-fourier-interpolationfrom
fix/raman-invariants-atom-fourier
Jul 30, 2026
Merged

Propagate Raman invariant fix to atom-Fourier interpolation#34
mesonepigreco merged 1 commit into
atom-fourier-interpolationfrom
fix/raman-invariants-atom-fourier

Conversation

@mesonepigreco

Copy link
Copy Markdown
Collaborator

Summary

Propagate the Raman invariant hotfix from #33 to atom-fourier-interpolation.

This branch contains the exact hotfix commit with recorded provenance:

620c527c3cf9c71df9c964014c74d03656885ef7

QSpaceAtomFourierLanczos inherits the corrected Q-space Raman preparation and accumulated-modulus behavior from QSpaceLanczos; no interpolation-specific implementation is added.

Root cause and impact

The prior real-space Raman implementation overwrote accumulated diagonal combinations for unpolarized channels 0–3, and both backend add=True paths measured only the latest projected component. The shared builder now constructs each invariant once and preserves consistent xy, xz, yz ordering.

Previously saved channels 0–3 must be recomputed. Channels 4–6 remain reusable. IR formulas and scaling are unchanged.

Validation

  • focused Raman/IR and inheritance tests: 12 passed
  • atom-Fourier and interpolation tests: 36 passed
  • the source commit also passed the full main-line real-space/Julia and Q-space suites reported in Fix Raman invariant preparation #33

(cherry picked from commit 620c527)
@mesonepigreco
mesonepigreco marked this pull request as ready for review July 30, 2026 11:38
@mesonepigreco
mesonepigreco merged commit cc76e26 into atom-fourier-interpolation Jul 30, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant