Problem with contourpy package

[Marie59]

Hi @Alexander-Barth and @ctroupin,
It's me again, I got the tool running on a local galaxy and now I'm trying to run the entire notebook "90-full-analysis.ipynb" to be sure that everything is in order and to finalize the docker image (still some small modifications to be made after some testing).
At the step "2.1 Choice of bathymetry" when I try to run the figure just below I get the following error :

`PyError ($(Expr(:escape, :(ccall(#= /home/jovyan/.julia/packages/PyCall/twYvK/src/pyfncall.jl:43 =# @pysym(:PyObject_Call), PyPtr, (PyPtr, PyPtr, PyPtr), o, pyargsptr, kw))))) <class 'ImportError'>
ImportError("/opt/julia-1.8.3/bin/../lib/julia/libstdc++.so.6: version `GLIBCXX_3.4.30' not found (required by /opt/conda/lib/python3.10/site-packages/contourpy/_contourpy.cpython-310-x86_64-linux-gnu.so)")
 File "/opt/conda/lib/python3.10/site-packages/matplotlib/pyplot.py", line 2527, in contour
   __ret = gca().contour(
 File "/opt/conda/lib/python3.10/site-packages/matplotlib/__init__.py", line 1442, in inner
   return func(ax, *map(sanitize_sequence, args), **kwargs)
 File "/opt/conda/lib/python3.10/site-packages/matplotlib/axes/_axes.py", line 6451, in contour
   contours = mcontour.QuadContourSet(self, *args, **kwargs)
 File "/opt/conda/lib/python3.10/site-packages/matplotlib/contour.py", line 769, in __init__
   kwargs = self._process_args(*args, **kwargs)
 File "/opt/conda/lib/python3.10/site-packages/matplotlib/contour.py", line 1395, in _process_args
   import contourpy
 File "/opt/conda/lib/python3.10/site-packages/contourpy/__init__.py", line 7, in <module>
   from contourpy._contourpy import (


Stacktrace:
 [1] pyerr_check
   @ ~/.julia/packages/PyCall/twYvK/src/exception.jl:75 [inlined]
 [2] pyerr_check
   @ ~/.julia/packages/PyCall/twYvK/src/exception.jl:79 [inlined]
 [3] _handle_error(msg::String)
   @ PyCall ~/.julia/packages/PyCall/twYvK/src/exception.jl:96
 [4] macro expansion
   @ ~/.julia/packages/PyCall/twYvK/src/exception.jl:110 [inlined]
 [5] #107
   @ ~/.julia/packages/PyCall/twYvK/src/pyfncall.jl:43 [inlined]
 [6] disable_sigint
   @ ./c.jl:473 [inlined]
 [7] __pycall!
   @ ~/.julia/packages/PyCall/twYvK/src/pyfncall.jl:42 [inlined]
 [8] _pycall!(ret::PyCall.PyObject, o::PyCall.PyObject, args::Tuple{StepRangeLen{Float64, Base.TwicePrecision{Float64}, Base.TwicePrecision{Float64}, Int64}, StepRangeLen{Float64, Base.TwicePrecision{Float64}, Base.TwicePrecision{Float64}, Int64}, Matrix{Float64}, Vector{Float64}}, nargs::Int64, kw::PyCall.PyObject)
   @ PyCall ~/.julia/packages/PyCall/twYvK/src/pyfncall.jl:29
 [9] _pycall!(ret::PyCall.PyObject, o::PyCall.PyObject, args::Tuple{StepRangeLen{Float64, Base.TwicePrecision{Float64}, Base.TwicePrecision{Float64}, Int64}, StepRangeLen{Float64, Base.TwicePrecision{Float64}, Base.TwicePrecision{Float64}, Int64}, Matrix{Float64}, Vector{Float64}}, kwargs::Base.Pairs{Symbol, Any, Tuple{Symbol, Symbol}, NamedTuple{(:colors, :linewidths), Tuple{String, Float64}}})
   @ PyCall ~/.julia/packages/PyCall/twYvK/src/pyfncall.jl:11
[10] pycall(::PyCall.PyObject, ::Type{PyCall.PyAny}, ::StepRangeLen{Float64, Base.TwicePrecision{Float64}, Base.TwicePrecision{Float64}, Int64}, ::Vararg{Any}; kwargs::Base.Pairs{Symbol, Any, Tuple{Symbol, Symbol}, NamedTuple{(:colors, :linewidths), Tuple{String, Float64}}})
   @ PyCall ~/.julia/packages/PyCall/twYvK/src/pyfncall.jl:83
[11] contour(::StepRangeLen{Float64, Base.TwicePrecision{Float64}, Base.TwicePrecision{Float64}, Int64}, ::Vararg{Any}; kws::Base.Pairs{Symbol, Any, Tuple{Symbol, Symbol}, NamedTuple{(:colors, :linewidths), Tuple{String, Float64}}})
   @ PyPlot ~/.julia/packages/PyPlot/H01LC/src/PyPlot.jl:194
[12] top-level scope
   @ In[17]:5`

Did you already have that ? Any idea of where it could be coming from ?

[Alexander-Barth]

I confirm this issue. I think this is this issue:
JuliaLang/julia#34276

Simply removing the file /opt/julia-1.8.3/bin/../lib/julia/libstdc++.so.6 worked for me to force loading the system libstdc++.so (used by python and which is newer that the julia libstdc++.so). This will be the default for the upcoming version 1.9 of julia:

https://github.com/JuliaLang/julia/blob/master/HISTORY.md#external-dependencies

I will make a new docker image with the /opt/julia-1.8.3/bin/../lib/julia/libstdc++.so.6 removed.

To reproduce the issue quickly:

using PyPlot; contourf(randn(10,10))

[Alexander-Barth]

Does it work for you now with abarth/divand-jupyterhub:2023-04-20T0902 ?

[Marie59]

I added a line removing /opt/julia-1.8.3/bin/../lib/julia/libstdc++.so.6 as you said in my local dockerfile of divand. It's working !! Thank you !
The only issue I had was in "3.1 World Ocean Database" with this linewod_all_data_filename = download("https://dox.ulg.ac.be/index.php/s/Px6r7MPlpXAePB2/download")
But I think the issue is just coming from the fact that download needs a 2nd argument with a name for the file (so I tried with a random name .nc) and it worked. The entire notebook seems to be running correctly !

[ctroupin]

Hello Marie,
in fact the function download can work with only 1 argument, in that case the file that is downloaded is located in a tmp directory; and the output of download (here: "wod_all_data_filename") is the path of the file.

julia> filename = download("https://dox.ulg.ac.be/index.php/s/Px6r7MPlpXAePB2/download"); @show(filename);
filename = "/tmp/jl_iTkDKbOvWV"

but maybe on Galaxy, it is not possible to write files in /tmp/.

[Alexander-Barth]

To debug this, one can use the julia function tempdir() (which is indeed /tmp per default). tempdir() should return be a writable directory. It could also be just a temporary network issue.