Hi folks, including Andreas and the other language founders (I hope you read this post)!
Since Nim can talk to C libraries, is there some easy way to have a Nim program use SciKit-learn?
Let me see if I can describe how much growth in popularity this might create for Nim...facts:
I'd say that being able to use compiled Nim code with SciKit-learn could be the thing that pushes Nim over the top in terms of recognition and popularity, just like data science is one of the main reasons Python is so popular.
I'm a casual programmer myself, so I'm not skilled enough at programming to know how to construct this...does it require a Nim library or package to use seamlessly? Is there documentation on how to create the necessary package?
Thanks and regards! Dave Sisk, MS in Data Science & Data Architect
nimpy, which lets you easily use Python libraries in your nim programs as well as create Python extensions written in nim (instead of in C or in Cython)
Nimpy won't make Scikit-learn faster. Even if you call Python from Nim it's still Python code being executed.
I'm a casual programmer myself, so I'm not skilled enough at programming to know how to construct this...does it require a Nim library or package to use seamlessly? Is there documentation on how to create the necessary package?
There are 2 way : you can either re-implement Scikit-learn in Nim native code. This is probably a long & complex task. Another possiblity is doing something similar to Flambeau binding the C++ shared library used by Scikit-learn and then make a Nim API resembling the Python API on top of it.