spectral-derivatives - Support for more bases

Added Legendre polynomials and Bernstein polynomials. Both are fit with least squares and therefore take significantly longer to run than the FFT-based methods for Fourier and Chebyshev, but more diversity is good. Bernstein polynomials in particular are shaped quite differently and add something unique.

- Jupyter Notebook
Published by pavelkomarov about 1 year ago

spectral-derivatives - Using Series-to-Series Rule Directly

This release is a product of #16 and subsequent work, largely codified in this comparison of methods. Due to realizations from my noise study, DCT-II support is revoked. But the good news is the series-based method is more flexible, so users can now supply any domain points they like, although they will be warned about increased computational cost.

- Jupyter Notebook
Published by pavelkomarov about 1 year ago

spectral-derivatives - DCT-II support

• added support for taking derivatives with the DCT-II on its half-index grid
• improved error messages to support this extension
• improved endpoint-finding code and added a parameter to bypass solving for the endpoints, in case they are not needed

- Jupyter Notebook
Published by pavelkomarov over 1 year ago

spectral-derivatives - increased allclose atol when checking for cosine-spaced points

- Jupyter Notebook
Published by pavelkomarov over 1 year ago

spectral-derivatives - nu -> order

warning for >4th order derivative now links to the notebook that solves #1

- Jupyter Notebook
Published by pavelkomarov over 1 year ago

spectral-derivatives - antiderivative support for Fourier

- Jupyter Notebook
Published by pavelkomarov over 1 year ago

spectral-derivatives - Filtering support

Bug fix so filtering works in higher dimension

- Jupyter Notebook
Published by pavelkomarov over 1 year ago

spectral-derivatives - filtering support

- Jupyter Notebook
Published by pavelkomarov over 1 year ago

spectral-derivatives - arbitrary domains support

Technically you don't need to take t_n as input to the functions, but giving it and treating a shift/scale of domain helps make the code a lot more user-friendly. Also added numerous error messages to steer users in the right direction vis-a-vis sampling, since it's easy to goof up.

- Jupyter Notebook
Published by pavelkomarov over 1 year ago

spectral-derivatives - minor improvements

- Jupyter Notebook
Published by pavelkomarov over 1 year ago

spectral-derivatives - There are now tests

- Jupyter Notebook
Published by pavelkomarov over 1 year ago

spectral-derivatives - Initial release

still no pytests or coverage, but I want to get it out

- Jupyter Notebook
Published by pavelkomarov over 1 year ago