Updating to 2.0 with MLJ?

[snowysnowcones]

Hi!
Congrats on the 2.0 release last month. I've been using SymbolicRegression.jl with the MLJ interface. It's been a long time since i've managed my packages and it doesn't seem like I have SymbolicRegression installed ("pkg status" does not list it) Instead I have "import SymbolicRegression: SRRegressor". How do I update SR to 2.0 with this interface?
Should I abandon and use the low level interface since it doesn't seem like MLJ is needed anymore?
Thanks for your advice

[MilesCranmer]

What is the error you are seeing?

I am not sure I 100% follow. But you can use the normal Julia package install command to get SymbolicRegression, i.e., Pkg.add("SymbolicRegression").

And yes in SymbolicRegression you can import the necessary components directly from SR, no need for a separate MLJ install

import SymbolicRegression: SRRegressor, machine, fit!, predict, report

# Dataset with two named features:
X = (a = rand(500), b = rand(500))

# and one target:
y = @. 2 * cos(X.a * 23.5) - X.b ^ 2

# with some noise:
y = y .+ randn(500) .* 1e-3

model = SRRegressor(
    niterations=50,
    binary_operators=[+, -, *],
    unary_operators=[cos],
)
mach = machine(model, X, y)
fit!(mach)
report(mach)

[snowysnowcones]

I don't see an error. So what I think was happening is I had SymbolicRegression loaded in my default environment that I no longer use. I guess "import SymbolicRegression" took it from the default environment.

Here is my environment I actively develop in. As you can see, no SR:

But here is my Default env:

I will just remove the package from my default environment and re-install on my actively developed environment. That should clear up the issue.

[MilesCranmer]

Ah I see. Yeah the current SR version is 2.4.1