Skip to content

ENH: expose the integer-grid refinement matrix - #899

Open
bingdongni wants to merge 1 commit into
PyWavelets:mainfrom
bingdongni:feature/integer-refinement-matrix
Open

bingdongni wants to merge 1 commit into
PyWavelets:mainfrom
bingdongni:feature/integer-refinement-matrix

Conversation

@bingdongni

Copy link
Copy Markdown

Related to #879. The low-pass filter already contains the integer-grid refinement operator, but callers currently have to assemble it and choose the orientation of the dual filter themselves. pywt.refinement_matrix(wavelet, dual=False) returns that operator for either scaling function. It keeps the filter grid, including leading or trailing zeros.

The tests compare matrix multiplication with subdivision by convolution and decimation. A cubic B-spline also checks the fixed point for its integer samples and the half-eigenvalue for its derivative. The function leaves eigenvector selection and normalization to the caller.

The Linux wheel test ran the installed suite (1,064 passed, two skipped), checked the docstring example, and compared both filters for all 106 built-in discrete wavelets.

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