juliacall.JuliaError: MethodError: Cannot `convert` an object of type PythonCall.Py to an object of type Tuple

[HerClau]

Example 2: Calling Julia from Python

Traceback (most recent call last):
  File "F:/Python/Python Scripts/Python module JuliaCall/Example_2_Calling_Julia_from_Python.py", line 22, in <module>
    jl.Vector[jl.Tuple]([batch(100) for _ in range(2000)]),
  File "C:\Users\user\.julia\packages\PythonCall\4eoCM\src\jlwrap\any.jl", line 201, in __call__
    return self._jl_callmethod($(pyjl_methodnum(pyjlany_call)), args, kwargs)
juliacall.JuliaError: MethodError: Cannot `convert` an object of type PythonCall.Py to an object of type Tuple
Closest candidates are:
  convert(::Type{T}, !Matched::T) where T<:Tuple at c:\users\user\appdata\local\programs\julia-1.7.2\share\julia\base\essentials.jl:336
  convert(::Type{T}, !Matched::Tuple{Vararg{Any, N}}) where {N, T<:Tuple} at c:\users\user\appdata\local\programs\julia-1.7.2\share\julia\base\essentials.jl:337
  convert(::Type{T}, !Matched::CartesianIndex) where T<:Tuple at c:\users\user\appdata\local\programs\julia-1.7.2\share\julia\base\multidimensional.jl:137

julia> versioninfo()
Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
OS: Windows (x86_64-w64-mingw32)
CPU: AMD Ryzen 7 5700X 8-Core Processor
WORD_SIZE: 64
LIBM: libopenlibm
LLVM: libLLVM-12.0.1 (ORCJIT, znver3)

(@v1.7) pkg> st Flux
Status C:\Users\user\.julia\environments\v1.7\Project.toml
[587475ba] Flux v0.13.3

(@v1.7) pkg> st PyCall
Status C:\Users\user\.julia\environments\v1.7\Project.toml
[438e738f] PyCall v1.93.1

Installed Pythons found by C:\Windows\py.exe Launcher for Windows
-3.8-64 C:\Users\user\AppData\Local\Programs\Python\Python38\python.exe *
pip list
juliacall 0.9.0
pip freeze
juliacall==0.9.0

[cjdoris]

Oh I need to update that example. The issue is this line:

jl.Vector[jl.Tuple]([batch(100) for _ in range(2000)])

you need to do

from juliacall import Main as jl, convert as jlconvert

then replace that line with

jlconvert(jl.Vector[jl.Tuple], [batch(100) for _ in range(2000)])

[HerClau]

I made the changes, but now another alert appears:

[juliapkg] Locating Julia ^1.6
[juliapkg] Using Julia 1.7.3 at C:\Users\user\AppData\Local\Programs\Julia-1.7.3\bin\julia.EXE
[juliapkg] Using Julia project at C:\Users\user\.julia\environments\pyjuliapkg
[juliapkg] Installing packages:
           julia> import Pkg
           julia> Pkg.add([Pkg.PackageSpec(name="PythonCall", uuid="6099a3de-0909-46bc-b1f4-468b9a2dfc0d")])
           julia> Pkg.resolve()

Traceback (most recent call last):
  File "F:\Python\Python Scripts\Python module JuliaCall\Example_2_Calling_Julia_from_Python.py", line 19, in <module>
    jl.Flux.train_b(
  File "C:\Users\user\.julia\packages\PythonCall\4eoCM\src\jlwrap\any.jl", line 201, in __call__
    return self._jl_callmethod($(pyjl_methodnum(pyjlany_call)), args, kwargs)
juliacall.JuliaError: UndefVarError: me not defined
Stacktrace:
  [1] adjoint
    @ C:\Users\user\.julia\packages\Zygote\ytjqm\src\lib\lib.jl:226 [inlined]
  [2] _pullback(__context__::Zygote.Context, 485::typeof(ZygoteRules.literal_getfield), x::Module, 486::Val{:me})
    @ Zygote C:\Users\user\.julia\packages\ZygoteRules\AIbCs\src\adjoint.jl:65
  [3] _pullback(cx::Zygote.Context, #unused#::typeof(getfield), x::Module, field_name::Symbol)
    @ Zygote C:\Users\user\.julia\packages\Zygote\ytjqm\src\lib\lib.jl:241
  [4] _pullback
    @ .\Base.jl:26 [inlined]
  [5] _pullback(::Zygote.Context, ::typeof(getproperty), ::Module, ::Symbol)
    @ Zygote C:\Users\user\.julia\packages\Zygote\ytjqm\src\compiler\interface2.jl:0
  [6] macro expansion
    @ C:\Users\user\.julia\packages\Zygote\ytjqm\src\lib\literal_getproperty.jl:83 [inlined]
  [7] _pullback
    @ C:\Users\user\.julia\packages\Zygote\ytjqm\src\lib\literal_getproperty.jl:44 [inlined]
  [8] _pullback
    @ .\none:1 [inlined]
  [9] _pullback(::Zygote.Context, ::var"#2#4"{Chain{Tuple{Dense{typeof(relu), Matrix{Float32}, Vector{Float32}}, Dense{typeof(relu), Matrix{Float32}, Vector{Float32}}, Dense{typeof(relu), Matrix{Float32}, Vector{Float32}}, Dense{typeof(identity), Matrix{Float32}, Vector{Float32}}}}}, ::PythonCall.PyArray{Float64, 2, true, false, Float64}, ::PythonCall.PyArray{Float64, 2, true, false, Float64})
    @ Zygote C:\Users\user\.julia\packages\Zygote\ytjqm\src\compiler\interface2.jl:0
 [10] _apply
    @ .\boot.jl:804 [inlined]
 [11] adjoint
    @ C:\Users\user\.julia\packages\Zygote\ytjqm\src\lib\lib.jl:200 [inlined]
 [12] _pullback
    @ C:\Users\user\.julia\packages\ZygoteRules\AIbCs\src\adjoint.jl:65 [inlined]
 [13] _pullback
    @ C:\Users\user\.julia\packages\Flux\js6mP\src\optimise\train.jl:120 [inlined]
 [14] _pullback(::Zygote.Context, ::Flux.Optimise.var"#37#40"{var"#2#4"{Chain{Tuple{Dense{typeof(relu), Matrix{Float32}, Vector{Float32}}, Dense{typeof(relu), Matrix{Float32}, Vector{Float32}}, Dense{typeof(relu), Matrix{Float32}, Vector{Float32}}, Dense{typeof(identity), Matrix{Float32}, Vector{Float32}}}}}, Tuple{PythonCall.PyArray{Float64, 2, true, false, Float64}, PythonCall.PyArray{Float64, 2, true, false, Float64}}})
    @ Zygote C:\Users\user\.julia\packages\Zygote\ytjqm\src\compiler\interface2.jl:0
 [15] pullback(f::Function, ps::Zygote.Params{Zygote.Buffer{Any, Vector{Any}}})
    @ Zygote C:\Users\user\.julia\packages\Zygote\ytjqm\src\compiler\interface.jl:352
 [16] gradient(f::Function, args::Zygote.Params{Zygote.Buffer{Any, Vector{Any}}})
    @ Zygote C:\Users\user\.julia\packages\Zygote\ytjqm\src\compiler\interface.jl:75
 [17] macro expansion
    @ C:\Users\user\.julia\packages\Flux\js6mP\src\optimise\train.jl:119 [inlined]
 [18] macro expansion
    @ C:\Users\user\.julia\packages\ProgressLogging\6KXlp\src\ProgressLogging.jl:328 [inlined]
 [19] train!(loss::Function, ps::Zygote.Params{Zygote.Buffer{Any, Vector{Any}}}, data::Vector{Tuple}, opt::ADAM; cb::Flux.Optimise.var"#38#41")
    @ Flux.Optimise C:\Users\user\.julia\packages\Flux\js6mP\src\optimise\train.jl:117
 [20] train!(loss::Function, ps::Zygote.Params{Zygote.Buffer{Any, Vector{Any}}}, data::Vector{Tuple}, opt::ADAM)
    @ Flux.Optimise C:\Users\user\.julia\packages\Flux\js6mP\src\optimise\train.jl:114
 [21] pyjlany_call(self::typeof(Flux.Optimise.train!), args_::PythonCall.Py, kwargs_::PythonCall.Py)
    @ PythonCall C:\Users\user\.julia\packages\PythonCall\4eoCM\src\jlwrap\any.jl:31
 [22] _pyjl_callmethod(f::Any, self_::Ptr{PythonCall.C.PyObject}, args_::Ptr{PythonCall.C.PyObject}, nargs::Int64)
    @ PythonCall C:\Users\user\.julia\packages\PythonCall\4eoCM\src\jlwrap\base.jl:69
 [23] _pyjl_callmethod(o::Ptr{PythonCall.C.PyObject}, args::Ptr{PythonCall.C.PyObject})
    @ PythonCall.C C:\Users\user\.julia\packages\PythonCall\4eoCM\src\cpython\jlwrap.jl:47

[cjdoris]

I don't know what that error is, it works fine for me - check you haven't made a typo. I have fixed the example.

[HerClau]

Reading the alert carefully, it seems that the error comes from calling the loss function:
"juliacall.JuliaError: UndefVarError: me not defined"
when changing the line:

loss = jl.seval("m -> (x, y) -> Flux.Losses.**ms**(m(x), y)")(model);

[my typo] by:

loss = jl.seval("m -> (x, y) -> Flux.Losses.mse(m(x), y)")(model)

The expected results are obtained:
loss = 0.015011342489806354 and the graph

[cjdoris]

Awesome 👍