RFC for ergonomically passing Python variables to Julia

[MilesCranmer]

I propose one of the following two APIs as a way to more ergonomically pass Python variables to Julia code:

jl.teval("const MY_CONST = $(x)", x=my_python_const)

a "template" eval, or

with jl.let(x=my_python_const):
    jl.seval("const MY_CONST = x")

which is a juliacall "let."

Would be interested in hearing other ideas too.

This is designed to address the following issue. Currently, passing Python objects to Julia via seval requires creating closure functions:

jl.seval("x -> @eval const MY_CONST = $x")(my_python_const)

While Python's PEP 750 proposes Template Strings which could offer one solution:

jl.seval(t"const MY_CONST = {my_python_const}")

We could implement something similar with a new teval method:

jl.teval("const MY_CONST = {my_python_const}")

However, this has two key issues:

1. The {} syntax conflicts with Julia type signatures (e.g., Vector{Float64}), which could cause subtle bugs. Even if t-strings are eventually merged to Python, we would still face this issue!
2. teval wouldn't have access to local variables to reference my_python_const

Let's look at the two proposed approaches in more detail:

The first approach makes variable passing explicit but ergonomic:

jl.teval("const MY_CONST = $(x)", x=my_python_const)

The second approach uses a context manager for temporary variable binding:

with jl.let(x=my_python_const):
    jl.seval("const MY_CONST = x")  # Would access the bound value

Basically the seval would have access to this stack of Python objects passed via let, and put them into the Julia context via a Julia-evaluated let statement.

Thoughts? @cjdoris @mkitti

Going to post this on the Julia discourse too. Edit: https://discourse.julialang.org/t/rfc-for-ergonomically-passing-python-variables-to-julia/124266