plot error?

[Wangwei-Lan]

Hello,

I plotted a circuit in jupyter notebook in one cell, it worked fine. However, I continue to plot another circuit in the next cell, it always give me error. See below
circuit1 = chain(1, put(1=>H)); plot(circuit1)
circuit2 = chain(1, put(1=>X)); plot(circuit2)
The first one gives me the correct H gate, but I run the second code, circuit2 always gives me the same plot as circuit1, which is incorrect since it is an X gate. I am not quite sure whether it is a bug or it is because the setup on my computer is wrong. Could you help me verify that?

[GiggleLiu]

Maybe I know which editor are you using? I tried VSCode, it plots correctly.

[Wangwei-Lan]

I am not using an editor, I used Jupiter notebook. For Vscode, it works fine.

[Egg12138]

I met a problem alike yesterday.

(Both in Jupyter notebook & VSCode builtin
jupyter notebook)
platform: WSL Ubuntu20.04LTS.
using Yao, YaoPlots

step1

In one block, I wrote:
chainYHZ = chain(3, put(1=>Y), put(2=>Z), put(3=>Z), repeat(Z, 1:3))
plot(chainYHZ)

output is as expected:
Y-Z
Z-Z
Z-Z

step2

then I ran another block of codes:
chainYHH = chain(3, put(1=>Y), put(2=>H), put(3=>H))
plot(chainYHH)

output is wrong:
Y
Z
Z

step3

I clean the output of the step1 block
the output of step2 immediately changed into the correct chain:
Y
H
H

step4

I re-executed the step1 block,
step2 block changed again:
Y
Z
Z

step5

I cleared output of step1 block again, (yeah, the step2 output remain correct: Y H H )
then
I added a new block after step2 block, denoted by step5 block.
chain5 = chain(3, put(1=>X), put(2=>H), put(3=>X))
plot(chain5)

the output is:
Y
H
Y

step6

repeat: I clear the step2 output,
the step5 block plot correctly:
X
H
X

step7

I insert a new block after step1 block, wrong plots happened again.

step8

I use println(chainXXX) , the output is always correct but julia plots incorrectly when plot outputs of other code blocks exist.
I don't know why.
(I'm new for Yao.jl so I'm not sure whether it's a feature...)

step9

even if I using Plots, same things happened.

[GiggleLiu]

The plotting backend we are using is Luxor. IIUC, it is a wield interaction between Luxor and Jupyter.
@cormullion do you have any idea?
Maybe this issue is related: JuliaGraphics/Luxor.jl#226

[GiggleLiu]

FYI: the reltated code block is https://github.com/QuantumBFS/YaoPlots.jl/blob/0157a3e8d57113d6219a8dba9d2f320f99fc1395/src/vizcircuit.jl#L463

[Egg12138]

The plotting backend we are using is Luxor. IIUC, it is a wield interaction between Luxor and Jupyter.
@cormullion do you have any idea?
Maybe this issue is related: JuliaGraphics/Luxor.jl#226

thanks. I'll try the solution in the quoted issues.

[cormullion]

I think this is the infamous "SVG/Jupyter cell interaction" problem, summarised by knuesel here: MakieOrg/Makie.jl#952 (comment)

ie

SVGs in different cells in a Jupyter notebook share definitions, which is usually not what you want.

A related issue was found in Pluto (JuliaGraphics/Luxor.jl#228) - but Pluto handles SVGs more carefully than Jupyter and doesn't have this problem.

[cormullion]

Since this is a common problem people encounter when using Jupyter, I'll add a workaround to the next Luxor release. (Very large SVGs might take a bit longer to display, perhaps...)

[GiggleLiu]

Thanks, that would be great!

[Egg12138]

Since this is a common problem people encounter when using Jupyter, I'll add a workaround to the next Luxor release. (Very large SVGs might take a bit longer to display, perhaps...)

That's awesome.Thanks!

[cormullion]

hopefully fixed in Luxor 3.6.0.