UI はローカライズ済みです。スキルの紹介と説明文書は選択した言語へ AI が自動翻訳し、コード・コマンド・パス・専門用語は原文のままです。原作者の文書が優先されます。
スキルを探す
生産性・共同作業 · SKILL.md

wavelet-packets

wavelet-packets

Routes PyWavelets users through 1D, 2D, and ND wavelet packet tree construction, traversal, and reconstruction.

Independent GitHub snapshot · Automated classification · Original documentation governs

スキルの説明

GitHub

name: wavelet-packets description: "Routes PyWavelets users through 1D, 2D, and ND wavelet packet tree construction, traversal, and reconstruction." disable-model-invocation: true metadata: disco-role: operating license: MIT

Wavelet Packets

Use this sub-skill when the task is about packet trees rather than ordinary DWT/SWT coefficient lists.

Route here when the task asks for

  • WaveletPacket, WaveletPacket2D, or WaveletPacketND
  • BaseNode, Node, Node2D, NodeND
  • path-based lookup, tuple path access, or path_tuple
  • get_level, get_leaf_nodes, walk, walk_depth, decompose, reconstruct, __getitem__, __setitem__, or __delitem__
  • node naming conventions such as a/d, a/h/v/d, or ND packet path tuples

Route elsewhere when the task is about

  • ordinary decimated transforms, SWT, MRA, or coefficient packing: go to ../discrete-transforms/SKILL.md
  • wavelet catalogs, custom wavelets, or CWT: go to ../wavelets-and-cwt/SKILL.md

Start here

  • Read references/api-reference.md for the verified constructor signatures and packet-tree methods.
  • Read references/workflows.md for tree traversal, reconstruction, and path-selection examples.
  • Read references/troubleshooting.md when a path, axis, or reconstruction step fails.
  • Run ../../scripts/check_pywavelets_install.py when you want a quick no-network smoke check that includes packet reconstruction.

Common workflow anchors

  • WaveletPacket defaults to mode='symmetric' and one transformed axis.
  • WaveletPacket2D and WaveletPacketND default to mode='smooth'.
  • WaveletPacket2D uses a/h/v/d naming for the four children of each node.
  • WaveletPacketND uses path tuples or repeated child tokens for deeper trees.
  • get_level(..., order='freq') is available on the 1D and 2D packet classes for frequency-style ordering.
  • get_leaf_nodes(decompose=False) returns the currently existing leaves without forcing deeper decomposition.

Useful bundled data

  • pywt.data.aero() and pywt.data.camera() are good for 2D packet smoke checks.
  • pywt.data.ecg() is useful for 1D packet smoke checks.

What to expect from the references

  • references/api-reference.md lists the verified packet constructors and node methods.
  • references/workflows.md shows how to navigate, replace, delete, and reconstruct packet nodes.
  • references/troubleshooting.md covers invalid paths, axes, and shape-trimming surprises.
保存したスキル