Jupyter error "tried to read a stream that is not readable"

[quasars100]

Hi,
Sorry this is a basic question about installation - I followed all instructions (Julia was already installed, then I ran "pip install -U pysr" followed by " python -c 'import pysr; pysr.install()' "). There was no issues regarding importing Pysr in python; however, when I try running

model = PySRRegressor(
    niterations=40,  # < Increase me for better results
    binary_operators=["+", "*"],
    unary_operators=[
        "cos",
        "exp",
        "sin",
        "inv(x) = 1/x",
        # ^ Custom operator (julia syntax)
    ],
    extra_sympy_mappings={"inv": lambda x: 1 / x},
    # ^ Define operator for SymPy as well
    loss="loss(x, y) = (x - y)^2",
    # ^ Custom loss function (julia syntax)
)
model.fit(x,y)

the following error appeared:

RuntimeError: <PyCall.jlwrap (in a Julia function called from Python)
JULIA: tried to read a stream that is not readable
Stacktrace:
 [1] watch_stream(stream::Base.PipeEndpoint)
   @ SymbolicRegression.SearchUtilsModule ~/.julia/packages/SymbolicRegression/37l4B/src/SearchUtils.jl:79
 [2] _EquationSearch(::SymbolicRegression.CoreModule.ProgramConstantsModule.SRThreaded, datasets::Vector{SymbolicRegression.CoreModule.DatasetModule.Dataset{Float32}}; niterations::Int64, options::Options{typeof(loss), Int64, 0.86, 10}, numprocs::Nothing, procs::Nothing, addprocs_function::Nothing, runtests::Bool, saved_state::Nothing)
   @ SymbolicRegression ~/.julia/packages/SymbolicRegression/37l4B/src/SymbolicRegression.jl:418
 [3] EquationSearch(datasets::Vector{SymbolicRegression.CoreModule.DatasetModule.Dataset{Float32}}; niterations::Int64, options::Options{typeof(loss), Int64, 0.86, 10}, parallelism::String, numprocs::Nothing, procs::Nothing, addprocs_function::Nothing, runtests::Bool, saved_state::Nothing)
   @ SymbolicRegression ~/.julia/packages/SymbolicRegression/37l4B/src/SymbolicRegression.jl:383
 [4] EquationSearch(X::Matrix{Float32}, y::Matrix{Float32}; niterations::Int64, weights::Nothing, varMap::Vector{String}, options::Options{typeof(loss), Int64, 0.86, 10}, parallelism::String, numprocs::Nothing, procs::Nothing, addprocs_function::Nothing, runtests::Bool, saved_state::Nothing, multithreaded::Nothing)
   @ SymbolicRegression ~/.julia/packages/SymbolicRegression/37l4B/src/SymbolicRegression.jl:320
 [5] invokelatest(::Any, ::Any, ::Vararg{Any}; kwargs::Base.Pairs{Symbol, Any, NTuple{8, Symbol}, NamedTuple{(:weights, :niterations, :varMap, :options, :numprocs, :parallelism, :saved_state, :addprocs_function), Tuple{Nothing, Int64, Vector{String}, Options{typeof(loss), Int64, 0.86, 10}, Nothing, String, Nothing, Nothing}}})
   @ Base ./essentials.jl:731
 [6] _pyjlwrap_call(f::Function, args_::Ptr{PyCall.PyObject_struct}, kw_::Ptr{PyCall.PyObject_struct})
   @ PyCall ~/.julia/packages/PyCall/twYvK/src/callback.jl:32
 [7] pyjlwrap_call(self_::Ptr{PyCall.PyObject_struct}, args_::Ptr{PyCall.PyObject_struct}, kw_::Ptr{PyCall.PyObject_struct})
   @ PyCall ~/.julia/packages/PyCall/twYvK/src/callback.jl:44>

One reason might be that my Julia version is currently 1.8.5 instead of 0.6.0 (which isn't supported anymore, when I tried to downgrade the current version), or it could be another issue.
Is there any way to fix the issue?
When I tried running the code with an older Julia version (1.0), it said I had to update Julia first.
Thank you.

[MilesCranmer]

Any Julia version greater or equal to 1.6 should work fine (I am personally using 1.9.0-beta2). It could be that an older Julia version is being called instead? Are the older versions of Julia deleted from your machine?

Also, please provide system details (OS, Python version, whether you are using it in IPython/Jupyter/terminal).

[quasars100]

Of course - my environment is system python (version 3.9.0), regular Jupyter notebook (no virtual environment setup). The Julia version I am running is 1.8.5, installed using homebrew (brew install --cask julia); to make it compatible with python, I ran "import julia; julia.install()" in python3 (using terminal). Then, using

import pysr
pysr.install()

from julia import Julia; julia = Julia(compiled_modules=False)
from julia import Main
from julia.tools import redirect_output_streams

redirect_output_streams()

Pysr was able to run model.fit(x,y). Hope this helps!

[MilesCranmer]

Thanks! I think you shouldn’t need the from julia import Julia; julia = Julia(compiled_modules=False) - is that correct?

[quasars100]

That's correct (it works with and without that line),

[MilesCranmer]

Thanks. So PySR 0.11.12, once it's up, will fix this issue, so you won't need to manually tweak the startup. The fix was this: astroautomata/SymbolicRegression.jl#169

[quasars100]

Thank you!