菜单
Use when working in SIMPLE's src/main/pca subsystem, including PCA, PPCA, complex PPCA, kernel PCA, diffusion maps, and SVD-backed dimensionality-reduction code used in analysis and denoising workflows.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use when working in SIMPLE's src/main/nano subsystem, including nanoparticle and molecule data structures, graphene and time-series helpers, and the SINGLE-specific domain logic that supports liquid-cell and nanoparticle workflows.
Use when working on SIMPLE's abinitio2D workflow and its cluster2D stages, including 2D search strategies, probabilistic 2D sampling, class-average restoration, distributed partial class-average assembly, online matcher particle I/O, project updates, and streaming/rejection helpers.
Use when working on SIMPLE's model_cavgs_rejection application, simple_cavg_quality feature-space selector, class-average quality selection, quality_mode=analyze outputs, cavgs_quality feature/reference tables, chunk vs pool rejection tuning, or microchunk quality-vector integration.
Use when studying, explaining, or modifying SIMPLE's fractional-update and trailing-reconstruction workflow, including update_frac sampling, sampled/updatecnt bookkeeping, matcher reconstruction I/O, partial even/odd reconstruction handoff, rec3D/volassemble ownership, downsampling compatibility, obsfield mirrors, and stage-to-stage previous-reference contracts.
Use when working in SIMPLE's src/main/stream subsystem, including online/mini-stream orchestration, chunk and microchunk processing, staged pipeline steps, watchers, queue/message definitions, and streaming variants of preprocessing and cluster2D workflows.
Use when working on SIMPLE's streaming microchunk class-average rejection, simple_microchunked2D lifecycle sentinels, pass-1/pass-2/refchunk/match generation, skip_pass_2 behavior, simple_cluster2D_rejector criteria, or boundaries with model_cavgs_rejection.
| name | simple-main-pca |
| description | Use when working in SIMPLE's src/main/pca subsystem, including PCA, PPCA, complex PPCA, kernel PCA, diffusion maps, and SVD-backed dimensionality-reduction code used in analysis and denoising workflows. |
src/main/pcaThis folder owns PCA-family methods.
simple_pca.f90simple_pca_svd.f90simple_ppca.f90simple_complex_ppca.f90simple_kpca_svd.f90simple_diffusion_maps.f90simple_diff_map_denoise.f90image/, project/, and numeric utilitiesDifferent PCA variants may share concepts but not assumptions; avoid forcing a single abstraction unless the code already wants it.