Continuing the discussion from Wrap Julia package in Python: which is the best option?:
When you wrap a Julia package, as in 2. or 3., yes it’s a wrapper, but also called a binding, when used from another language like Python or R. I’ve used the former word more, but trying to use the technically correct latter word myself (the former needs not be across languages).
Those are the two main binding packages I recall (and JuliaCall used for both, pyjulia before, and I believe JuliaCall best to use, and the other JuliaCall for R package best to to implement R bindings), 2. diffeqpy, is also available as an R binding. @MilesCranmer, I believe PySR doesn’t have an equivalent R binding? I feel like I recall some binding for R, but I might be confusing with a binding in other direction:
Can you also name some more for Python, or R or other language?
A lot of people embed Julia in C, C++, Rust; and even C# or Java (or there are packages similar to PythonCall.jl/JuliaCall for C#/.NET and Java to call from at least, then I assume some call Julia from those too, maybe even F#).
There must be some C bindings, but since there’s no standard C (or C++) package manager, I just don’t know or where to look.
https://crates.io/crates/jlrs/reverse_dependencies
jlrs provides bindings to the Julia C API that enable Julia code to be called from Rust and more.
jlrs-ndarray is listed as only dependant, and it’s really just a related package in it’s docs “jlrs currently supports Linux and Windows.” which is outdated, since it supports macOS too. SO I see no good reason jlrs doesn’t enable a lot of bindings, they might just all be private, just not (yet) in a repo/registered crate.
I’m just curious, mainly how many Julia packages have been wrapper for bindings, e.g. for Python. There’s also some interop with Ruby and Lua, but I’m just not explored those communities. And for JavaScript, mainly for Julia’s benefit? There’s node-julia to call Juia, but I think not yet common or easy to call from a web browser, JavaScript that way.
[A lot of Julia packages reuse other languages, most often C, or C++ e.g. GitHub - xtensor-stack/xtensor-julia: Julia bindings for xtensor but I’m not asking for that, rather which other languages use Julia.]
To partially answer my question I now found another Python binding (for CellListMap.jl):
though I still want more examples, the more the better… and even better, like @biona001 trying to make binding for more than one language (was it successful?), and then which:
I think you need to make bindings separate, i.e. the wrapped code is just Julia, and the most code, and all bindings some overhead, but is there some way to make bindings for 2 or more language easier (in Julia or even other languages), something for automating making them?