You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I needed the first-kind Bessel function, so I tried using the scilib library. However, I found that the function scilib::math::bessel::j_nu(nu: f64, z: Complex64) -> Complex64 lacked sufficient precision, and the function scilib::math::bessel::j_n(n: i32, x: f64) -> f64 produced completely inaccurate, diverging results.
Switching to puruspe::bessel::Jn resolved the issue, and I’m satisfied with its higher precision compared to scilib.
That said, I’m still using scilib’s j_nu function for complex inputs, but I’d like to fully transition to puruspe if possible. Do you have any plans to support a complex input version of the first-kind Bessel function in puruspe?
Thank you for providing such an excellent library!
The text was updated successfully, but these errors were encountered:
I can't speak for @Axect but I was planning to submit PRs for a few complex functions over the winter break. I think we proposed adding num-complex as an optional dependency, behind a feature flag.
Thank you for your feedback! I'm really glad to hear that Puruspe's Bessel function is working well for your needs.
Regarding complex number support - yes, we've noticed increasing demand for this functionality. As @ethanbarry mentioned, we plan to implement this by introducing num-complex as an optional dependency behind a feature flag. While the exact timeline isn't set yet, it would be great to coordinate this with @ethanbarry's planned PRs during the winter break for a comprehensive update.
Looking forward to bringing these improvements to Puruspe!
Hello,
I needed the first-kind Bessel function, so I tried using the scilib library. However, I found that the function
scilib::math::bessel::j_nu(nu: f64, z: Complex64) -> Complex64
lacked sufficient precision, and the functionscilib::math::bessel::j_n(n: i32, x: f64) -> f64
produced completely inaccurate, diverging results.Switching to
puruspe::bessel::Jn
resolved the issue, and I’m satisfied with its higher precision compared toscilib
.That said, I’m still using
scilib
’sj_nu
function for complex inputs, but I’d like to fully transition topuruspe
if possible. Do you have any plans to support a complex input version of the first-kind Bessel function inpuruspe
?Thank you for providing such an excellent library!
The text was updated successfully, but these errors were encountered: