Extend plot support for `entanglementplot` and `transferplot` to `Makie.jl` by borisdevos · Pull Request #428 · QuantumKitHub/MPSKit.jl

borisdevos · 2026-05-20T07:22:57Z

Deals with #249. This works locally, below I attach the results for both Plots.jl and CairoMakie.jl*. I didn't really test any other Makie backend, but I think CairoMakie is the one you want anyways in this context.

As you can see in the attachment, I didn't really bother to make the Plots.jl plots any prettier than they were, and they remain equally blurry. Anyone willing to prettify these, feel free to do so. I put slightly more effort in the Makie plots, but even there plenty of improvements can be added.

I added todo's here and there, mostly stylistic choices.

*There are two plots per function, just to show that keyword arguments are being taken correctly:

# gs is the ground-state of the Ising Hamiltonian at g = 0.1
entanglementplot(gs)
entanglementplot(gs, expand_symmetry = true, sortby = maximum, sector_formatter = c -> "Sector $c")

transferplot(gs)
transferplot(gs; sectors = [Z2Irrep(0), Z2Irrep(1)], transferkwargs = (; num_vals = 1))

someplots.pdf

…o bd/plotting

lkdvos

This definitely looks great!
I have a couple questions, mostly since it has been a while since I looked at Makie:

Do you know if the other attributes can still be set/changed? For example, what happens if I want to override the default title/ticks/...?

I seem to recall also that there was some system with observables etc, where various parts where dynamically recalculated. Do you know if this is still relevant?

lkdvos · 2026-05-20T17:24:13Z

 Adapt = "79e6a3ab-5dfb-504d-930d-738a2a938a0e"
 Aqua = "4c88cf16-eb10-579e-8560-4a9242c79595"
 BlockTensorKit = "5f87ffc2-9cf1-4a46-8172-465d160bd8cd"
+CairoMakie = "13f3f980-e62b-5c42-98c6-ff1f3baf88f0"


This should probably be a weakdep?

lkdvos · 2026-05-20T17:26:06Z

 DocStringExtensions = "ffbed154-4ef7-542d-bbb7-c09d3a79fcae"
 HalfIntegers = "f0d1745a-41c9-11e9-1dd9-e5d34d218721"
 KrylovKit = "0b1a1467-8014-51b9-945f-bf0ae24f4b77"
+LaTeXStrings = "b964fa9f-0449-5b57-a5c2-d3ea65f4040f"


Does this make sense to have as a default dependency? We could in principle also have a default xlabel that uses Unicode, and require people that wish Latex strings to specify the xlabel themselves (which I would assume people typically want anyways)

At least for Makie I could keep the default latexstring expressions and put it as a weakdep, since it's explicitly in the deps of Makie. I put it here in deps to get it working for Plots, but we can indeed default to whatever and let the user overwrite themselves (if possible for Plots, I think I've got it working for Makie right now)

borisdevos · 2026-05-21T15:01:08Z

I just tried some random keyword arguments for the Makie plots, and they seem to work. I'll test tomorrow if I can do the same for Plots.
(Also I realised I was running Z2-symmetric MPS in the Z2 SSB phase 🫠)

…o bd/plotting

borisdevos · 2026-05-26T09:24:20Z

I don't understand the current aqua test failing. It seems to be complaining about missing compat bounds for every package under extras of the main project.toml (but are specified in the test project.toml), but this previously wasn't failing.

codecov · 2026-05-27T10:02:39Z

Codecov Report

❌ Patch coverage is 80.86957% with 44 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
ext/MPSKitMakieExt.jl	82.05%	28 Missing ⚠️
ext/MPSKitPlotsExt.jl	78.37%	16 Missing ⚠️

Files with missing lines	Coverage Δ
src/MPSKit.jl	`100.00% <ø> (ø)`
ext/MPSKitPlotsExt.jl	`78.37% <78.37%> (ø)`
ext/MPSKitMakieExt.jl	`82.05% <82.05%> (ø)`

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

…ng for in-place methods

borisdevos · 2026-09-08T19:18:15Z

Okay, I finally managed to look into this again. I'm happy with how things are in terms of being able to use these backends with some customisation. What remains is some stylistic decisions. In particular, I don't know how useful it is to plot transfer matrix spectra the way we are now. These phases don't matter in many cases, so if we allow some toggle to not plot them, there's then a way to plot the different sectors next to each other, similar to the current entanglement plots.

Some other style choices are in todos, though I'm sure there are many more.

lkdvos · 2026-09-15T17:58:57Z

+#TODO: remove once the recipes publish their axis attributes instead of setting them
+function with_current_axis(f, target)
+    target isa Makie.AbstractAxis || return f()
+    previous = Makie.current_axis()
+    Makie.current_axis!(target)
+    try
+        return f()
+    finally
+        isnothing(previous) || Makie.current_axis!(previous)
+    end
+end
+
+# overwrite user-provided axis attributes
+function apply_plotkwargs!(ax, plotkwargs)
+    ax isa Makie.AbstractAxis || return ax
+    for (k, v) in pairs(plotkwargs)
+        setproperty!(ax, k, v)
+    end
+    return ax
+end


This looks a little backwards, I feel like it should be possible to make the attributes respect what was provided, and somehow it feels like we are doing too much work with providing ticks etc, for which there should just be a default that is sensible and otherwise user-overridable?

So this part of the code exists to deal with the in-place versions of the plot functions, but you want these things to be attributes of the default?

The workaround here was because some of the default attributes I set are of the axis, and not of the scatter plot itself. I looked at the whole ComputeGraph machinery you once mentioned to me, and I think that's exactly you want. I'll try this out

lkdvos · 2026-09-15T18:07:47Z

+
+    ax.ylabel = L"\log(\lambda)"
+    ax.ylabelsize = 24
+    smallest = minimum(Iterators.filter(>(0), Iterators.flatten(spectrum)); init = 1.0) # safety net


Does this mean that we are effectively hiding pathological cases? Do you think there is a way of still showing them so we can diagnose the issue?

I can filter on ==0, give a warning but still plot the way it is currently?

…o bd/plotting

borisdevos added 11 commits February 16, 2026 12:54

first attempt at extension packages

cf5effa

plots weakdep

0f295ba

only recipesbase weakdep

15a9022

tests

49bd34d

Merge branch 'main' of https://github.com/QuantumKitHub/MPSKit.jl int…

b0395fc

…o bd/plotting

get makie working

f583b6f

get plots working

babffbf

minor

d66b9ce

update docstrings

97729ac

Merge branch 'main' of https://github.com/QuantumKitHub/MPSKit.jl int…

28b8b17

…o bd/plotting

remove todo

59779ca

lkdvos reviewed May 20, 2026

View reviewed changes

borisdevos added 2 commits May 21, 2026 16:56

get makie plot kwargs working

ed6054a

potential project.toml test fix

a6bf25f

borisdevos added 5 commits May 22, 2026 17:52

remove latexstrings dep in plots.jl, but keep in makie for now

b0ff466

docstrings

5be4fc4

fix makie tests

7d90389

i have no clue how to fix aqua tests

1595a47

Merge branch 'main' of https://github.com/QuantumKitHub/MPSKit.jl int…

a74a333

…o bd/plotting

lkdvos reviewed May 26, 2026

View reviewed changes

Comment thread Project.toml Outdated

remove extras and targets

14c37e2

lkdvos mentioned this pull request Jul 6, 2026

refactor: transfer_spectrum and correlation_length default to sector-centered implementation #452

Merged

borisdevos added 5 commits September 3, 2026 15:03

merge, dealing with previous refactors

5978d87

restore 2-arg form of transferplot

c2d9cbd

in-place methods easily defined for plots.jl

8422b3b

annoying workaround for makie in-place methods

5343d2e

safety net on 0 schmidt value, remove ambiguity, get plotkwargs worki…

794e004

…ng for in-place methods

borisdevos added 4 commits September 6, 2026 22:16

add in-place and symmetric mps tests

7b9501c

have the plots backend also work with plotkwargs

cc74519

allow moving legend in makie transferplots

6f88ab4

add to changelog

259e9ee

custom workers for plot backends

4f62d65

lkdvos reviewed Sep 15, 2026

View reviewed changes

Comment thread test/runtests.jl

lkdvos reviewed Sep 15, 2026

View reviewed changes

Comment thread Project.toml Outdated

lkdvos reviewed Sep 15, 2026

View reviewed changes

borisdevos added 4 commits September 18, 2026 14:21

fix bad merge

676e778

warnings on 0 schmidt value, same for plots

0418d74

work with p.attributes to style axis with :axis_info node

6c3b445

Merge branch 'main' of https://github.com/QuantumKitHub/MPSKit.jl int…

5d34bff

…o bd/plotting

Conversation

borisdevos commented May 20, 2026

Uh oh!

lkdvos left a comment

Choose a reason for hiding this comment

Uh oh!

lkdvos May 20, 2026

Choose a reason for hiding this comment

Uh oh!

lkdvos May 20, 2026

Choose a reason for hiding this comment

Uh oh!

borisdevos May 21, 2026

Choose a reason for hiding this comment

Uh oh!

borisdevos commented May 21, 2026

Uh oh!

borisdevos commented May 26, 2026

Uh oh!

Uh oh!

codecov Bot commented May 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

borisdevos commented Sep 8, 2026

Uh oh!

Uh oh!

Uh oh!

lkdvos Sep 15, 2026

Choose a reason for hiding this comment

Uh oh!

borisdevos Sep 18, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

lkdvos Sep 15, 2026

Choose a reason for hiding this comment

Uh oh!

borisdevos Sep 18, 2026

Choose a reason for hiding this comment

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

codecov Bot commented May 27, 2026 •

edited

Loading