Dev 2.1.2

[xumi1993]

No description provided.

[Copilot]

🟡 Changes recommended

Critical derivative-chain and optimization-coordinate inconsistencies remain unresolved.

Get a fresh assessment by requesting another Copilot review.

Pull request overview

Updates SurfATT to version 2.1.2 with Rayleigh kernel changes, radial-anisotropy handling, and kernel-density output.

Changes:

• Adjusts Rayleigh derivative scaling.
• Adds radial kernel handling.
• Stores and exposes per-iteration kernel density.
• Bumps the project version.

File summaries

File	Summary
src/xdmf.cpp	Adds kernel-density XDMF attributes; moderate, 1 vote: density visibility is incorrectly coupled to gradient output.
src/surfker/rlker.cpp	Updates Rayleigh derivative scaling; critical, 1 vote: derivative conventions remain inconsistent with optimization updates.
src/preproc.cpp	Adds radial-anisotropy kernel handling; critical, 2 votes: the radial chain-rule contributions are incomplete.
src/optimize.cpp	Revises optimizer derivative documentation; nit, 2 votes: the explanation contradicts the implementation.
src/inversion.cpp	Stores weighted kernel density per iteration.
include/inversion.h	Declares kernel-density storage support.
CMakeLists.txt	Bumps the version to 2.1.2.

Review details

Suppressed comments (1)

src/xdmf.cpp:98

• store_kernel_density() writes kernel_density_N independently of gradient snapshots, but this attribute is emitted only when it <= last_grad_iter. With the default steepest-descent configuration (output_in_process_model false), the HDF5 dataset is created but XDMF never references it; even when model output is enabled, density visibility is incorrectly coupled to gradient output. Emit the attribute based on the density dataset/iteration and update the XDMF after writing it.

            if (IP.postproc().is_kden) {
                out << attr("kernel_density", "kernel_density" + sfx);
            }

• Files reviewed: 7/7 changed files
• Comments generated: 3
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.