如何在 python Jupiter 笔记本中获取 Julia Makie 输出？

Question

运行以下笔记本单元

%load_ext julia.magic

%%julia
using GLMakie 

function plot_something()
    odeSol(x,y) = Point(-x, 2y)             # x'(t) = -x, y'(t) = 2y
    fig = Figure(resolution =(400,400))
    ax = Axis(fig)
    streamplot!(ax, odeSol, -2..2, -2..2, colormap = :plasma, 
        gridsize= (32,32), arrow_size = 0.07)
    fig[1,1] = ax
    fig
end

from julia import Main
Main.show_something()

产量`

目前可能的修复

function makie2image(fig)
    buf = IOBuffer()
    Makie.backend_show(GLMakie.GLBackend(), buf, MIME("image/png"), fig.scene)
    pyimport("IPython").display.Image(pybytes(buf.data))
end

用作：

Main.makie2image(Main.plot_something())

上述答案对我不起作用。相反，我必须使用更新的 JuliaCall 库

from juliacall import Main as jl

# Install the Makie package. This only needs to be run once then Main reloaded
import juliapkg
juliapkg.add("GLMakie", uuid="e9467ef8-e4e7-5192-8a1a-b1aee30e663a")

# Import GLMAkie into the Julia session
jl.seval("using GLMakie")

import numpy as np
x = np.linspace(0.1, 10, 30)
f = lambda x: x**2 * np.sin(x + 1/x)
y = f(x)

fig, ax, lns = jl.lines(x, y, label="Simple function")
jl.axislegend(position=jl.seval(':lt'))
fig

不幸的是，这失去了酷炫的

%%julia

细胞魔法

Answer 1

Answer 2

上述答案对我不起作用。相反，我必须使用更新的 JuliaCall 库

from juliacall import Main as jl

# Install the Makie package. This only needs to be run once then Main reloaded
import juliapkg
juliapkg.add("GLMakie", uuid="e9467ef8-e4e7-5192-8a1a-b1aee30e663a")

# Import GLMAkie into the Julia session
jl.seval("using GLMakie")

import numpy as np
x = np.linspace(0.1, 10, 30)
f = lambda x: x**2 * np.sin(x + 1/x)
y = f(x)

fig, ax, lns = jl.lines(x, y, label="Simple function")
jl.axislegend(position=jl.seval(':lt'))
fig

不幸的是，这失去了酷炫的

%%julia

细胞魔法