AGENT SKILL

wavelets-and-cwt

Routes PyWavelets users through wavelet-family inspection, custom wavelets, wavefun output, scale conversion, and continuous wavelet transforms.

License 许可
MIT
最近更新
2026-09-03
Stars
★ 351
Forks
27

01 前言 Front Matter

name
wavelets-and-cwt
license
MIT
metadata
{ "disco-role": "operating" }
description
Routes PyWavelets users through wavelet-family inspection, custom wavelets, wavefun output, scale conversion, and continuous wavelet transforms.
disable-model-invocation
true

02 资源文件 3

03 SKILL.md 正文

Wavelets and CWT

Use this sub-skill when the task is about the wavelet catalog, custom wavelet construction, wavelet-function inspection, or the continuous wavelet transform.

Route here when the task asks for

  • families() or wavelist()
  • Wavelet, ContinuousWavelet, or DiscreteContinuousWavelet
  • custom wavelets built from a filter_bank
  • wavefun, integrate_wavelet, central_frequency, scale2frequency, or frequency2scale
  • cwt and continuous-wavelet scaleograms
  • bundled demo signals such as pywt.data.demo_signal(...) or the sample pywt.data.nino() time series

Route elsewhere when the task is about

  • decimated transforms, SWT, MRA, coefficient packing, thresholding, padding, or fully separable transforms: go to ../discrete-transforms/SKILL.md
  • packet trees, packet path traversal, or packet reconstruction: go to ../wavelet-packets/SKILL.md

Start here

  • Read references/api-reference.md for the verified constructor signatures and the live cwt call shape.
  • Read references/workflows.md for recipes that choose a wavelet, inspect wavefun, and compute a CWT.
  • Read references/troubleshooting.md when wavelet names, continuous/dtype selection, or CWT scale/method choices fail.
  • Run ../../scripts/inspect_wavelet.py when you want a safe, no-plot inspection of a discrete or continuous wavelet.
  • Run ../../scripts/check_pywavelets_install.py when you want a broader smoke check that also exercises the core transform stack.

Common workflow anchors

  • Use pywt.families() to group the built-in discrete wavelet families.
  • Use pywt.wavelist(kind='continuous') to choose a continuous wavelet for cwt.
  • Use pywt.DiscreteContinuousWavelet(name) when you need a one-call resolver for discrete or continuous names.
  • Use pywt.ContinuousWavelet(name, dtype) when you want to force a continuous wavelet object directly.
  • Use pywt.scale2frequency(...) and pywt.frequency2scale(...) to convert between CWT scale and normalized frequency.
  • Use pywt.data.nino() or pywt.data.demo_signal('Doppler', n) for a small real-valued CWT input.

Important package facts

  • The live package accepts continuous wavelets such as morl, mexh, cmor, shan, fbsp, and cgau for cwt.
  • The live cwt implementation requires a continuous wavelet object and positive scales.
  • The code path currently accepts method='conv' and method='fft'; do not rely on auto as a runtime option.
  • wavefun returns 2, 3, or 5 outputs depending on whether the wavelet is continuous, orthogonal, or biorthogonal.

What to expect from the references

  • references/api-reference.md records the verified public constructors, helper functions, and cwt arguments.
  • references/workflows.md shows how to inspect wavelet objects and compute a scaleogram without guessing parameters.
  • references/troubleshooting.md covers the common continuous-wavelet, dtype, parameter-string, and scale errors.

04 同仓库更多 Skills

VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-triton-opt

Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-solve

END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-skill

PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-pattern-fusion

Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-orchestrate

Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-feedback

Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).

★ 351