espdev

Results 58 comments of espdev

I have the same problem with aiofiles. My simple benchmark and results on linux/uvloop: ```python import time import asyncio import aiofiles import uvloop uvloop.install() async def _async_file_reader(file_name): async with aiofiles.open(file_name,...

Hi, The smoothing parameter is computed just like in MATLAB implementation. See: https://csaps.readthedocs.io/en/latest/formulation.html#definition > The calculation of the smoothing spline requires the solution of a linear system whose coefficient matrix...

Only the following description from Matlab documentation without any links and citations, unfortunately. ![2022-07-30_10-23-21](https://user-images.githubusercontent.com/1299189/181904257-3a845984-7768-4d0b-af36-2fda2b388f4e.png) Matlab implementation also has roughness weights through smoothing parameter vector. Currently, this is not implemented in...

Have you considered the possibility to use bivariate spline (`xs = u(t), ys = v(t)`) for your data with close X-values? For example this can be implemented something like this:...

@shusheer, Thank you for your report. You are right, periodic extrapolation from scipy PPoly seems to work incorrectly with smoothing spline. Unfortunately, I have never tested this feature. If we...

Thanks for the explanation and the code example! > There does appear to be some kind of discontinuity in the first derivative of the CubicSmoothingSpline with smoothing=1.0 case (orange line...

> My understanding of the csaps code is insufficiently deep to actually implement it, but I think in _sspumv.py at def _make_spline(x, y, w, smooth, shape) you would introduce this...

> There's no good reason for this to break the code I think explicit is better than implicit. Also if we try to set the same x values to CubicSpline...

Hello @Amit-Nayak-2000, I think in your case the factorization does not work for some reason. https://github.com/espdev/csaps-cpp/blob/41a67952e7ea71c05f76b6cb1f178825c576a981/src/csaps.cpp#L107-L121 See also notes from the [documentation](https://eigen.tuxfamily.org/dox/classEigen_1_1SparseLU.html): > Unlike the initial [SuperLU](https://eigen.tuxfamily.org/dox/classEigen_1_1SuperLU.html) implementation, there is...