search

HugoMVale / odrpack95

Software for Weighted Orthogonal Distance Regression with Bound Constraints
https://hugomvale.github.io/odrpack95/
BSD 3-Clause "New" or "Revised" License
6 stars 0 forks source link

Testing volunteer #1

Open markcampanelli opened 5 months ago

markcampanelli commented 5 months ago

@HugoMVale Thanks for tackling new Python bindings to ODRDPACK95. I make great use of ODR with the implicit-equation formulation. I also have a use case for bounded parameters. I would be happy to test out https://github.com/markcampanelli/pvfit using these Python bindings. Please ping me here when you think things are ready for such testing.

HugoMVale commented 5 months ago

@markcampanelli, thanks a lot for the offer. I hope a first draft in June. Keep you posted.

RobbieMackenzie commented 1 week ago

Great that you're doing this Hugo, I was thinking of rewriting ODRPACK95 in Julia a while back but I didn't have sufficient skill. Python bindings for ODRPACK95 is a great addition. Thanks.

HugoMVale commented 1 day ago

Thanks for the encouragement, @RobbieMackenzie.

My personal opinion is that mankind does not have the luxury of having enough people with enough time to afford a translation of a given library in every language. I just can't imagine having - and maintaining - a version of cvode (just to mention a well-known library) in C, Fortran, Julia, Rust, Go, Zig, Mojo, Carbon, ...

I think the odrpack source code is now in a presentable, modern state that can be understood, fixed, and improved by anyone that would be able to do do in any other language. Now, one should definitely work on bindings, as you say!

My hope/understanding is that the scipy team could take care of the python bindings, since they already have bindings for the "old" ODRPACK. Additionally, one could also do an interface for Julia. That would be really cool.!I just saw a nice example for PRIMA.