You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Julia Version 1.12.1
Commit ba1e628ee4 (2025-10-17 13:02 UTC)
Build Info:
Official https://julialang.org release
Platform Info:
OS: Windows (x86_64-w64-mingw32)
CPU: 14 × Intel(R) Core(TM) Ultra 7 255U
WORD_SIZE: 64
LLVM: libLLVM-18.1.7 (ORCJIT, goldmont)
GC: Built with stock GC
Threads: 1 default, 1 interactive, 1 GC (on 14 virtual cores)
Output running the following in Julia 0.7 or greater: using Pkg; pkg"status" (if you are on Julia 0.6 or earlier run Pkg.status())
Status `C:\***\environments\v1.12\Project.toml`
[76e01b6b] FlexUnits v0.2.8
⌅ [682c06a0] JSON v0.21.4
[0f8b85d8] JSON3 v1.14.3
[f0f68f2c] PlotlyJS v0.18.17
[91a5bcdd] Plots v1.41.1
[295af30f] Revise v3.11.0
[b543fe20] TimeRecords v2.0.2
Info Packages marked with ⌅ have new versions available but compatibility constraints restrict them from upgrading. To see why use `status --outdated`
Reacted by t-bltg, Etienne Ott, Alexey Stukalov and Roman Lebedev
I gave up on PlotlyJS extra features and switched to PlotlyKaleido for just off-line rendering. WebIO seems the stumbling block - I assume they’ll fix it eventually.
I spent some more time on this, but this is progressing slowly ...
Using PlotlyBase + PlotlyKaleido is the better choice anyhow, because it has far less dependencies.
Is there any way I could use this combo a backend through the Plots interface? I only really use PlotlJS for the zooming capability. I eventually and up saving files anyway, but it's faster to just zoom in on the plot than repeatedly changing axes programmatically to inspect long timeseries for disturbance patterns. Most of my code-base already uses the Plots.jl interface.
FWIW - Use PlotlyKaleido to generate an HTML and that HTML will be interactive (w/ zoom and pan and so on). Not as powerful as PlotlyJS and not as nicely integrated w/ IDEs and REPLs, but...
Today I could get PlotlyJS running!
It is still brittle, as the Blink version relies on JSExpr 0.5. But I've now moved to JSExpr 1, which is breaking.
The good thing is, I started from scratch from the master branches and made only minor changes. The strange thing is that Blink's @js macro stopped working. Currently I import JSExpr and rely on JSExpr's @js macro, which works differently.
But there's light at the end of the tunnel.
Reacted by Michele Zaffalon, Simon Christ, MChristianBF, Gabgobie and Math, Python & Julia
Activity
BeastyBlacksmith commentedon Oct 23, 2025
xref: #512 (comment)
fhagemann commentedon Dec 9, 2025
+1
PhyX-Meow commentedon Jan 25, 2026
+1 Please, I have to give up using PlotlyJS because this. It conflicts with a lot of packages now.
jd-foster commentedon Feb 3, 2026
An example to guide this: jump-dev/MathOptInterface.jl#2918
Also, this advice here: sglyon/PlotlyBase.jl#121 (comment)
hhaensel commentedon Mar 21, 2026
See also JuliaGizmos/WebIO.jl#518
orenbenkiki commentedon Mar 25, 2026
Also running into this. JSON is at 1.4 now, PlotlyJS pins it to 0.2[01]. JSON is a pretty popular package so this causes a lot of conflicts.
hhaensel commentedon Mar 25, 2026
Most of is is ready to be rolled out, but we need 4 packages to bump their compats. I submitted PRs in all of them.
Deduction42 commentedon Jun 12, 2026
Any luck with getting the dependencies to bump their compats?
orenbenkiki commentedon Jun 13, 2026
I gave up on PlotlyJS extra features and switched to PlotlyKaleido for just off-line rendering. WebIO seems the stumbling block - I assume they’ll fix it eventually.
hhaensel commentedon Jun 13, 2026
I spent some more time on this, but this is progressing slowly ...
Using PlotlyBase + PlotlyKaleido is the better choice anyhow, because it has far less dependencies.
Deduction42 commentedon Jun 13, 2026
Is there any way I could use this combo a backend through the Plots interface? I only really use PlotlJS for the zooming capability. I eventually and up saving files anyway, but it's faster to just zoom in on the plot than repeatedly changing axes programmatically to inspect long timeseries for disturbance patterns. Most of my code-base already uses the Plots.jl interface.
hhaensel commentedon Jun 13, 2026
You could use either VSCode's ability to display plots. Just execute
at VSCode's REPL.
Or you could use a Genie App with StipplePlotly ;-)
orenbenkiki commentedon Jun 13, 2026
FWIW - Use PlotlyKaleido to generate an HTML and that HTML will be interactive (w/ zoom and pan and so on). Not as powerful as PlotlyJS and not as nicely integrated w/ IDEs and REPLs, but...
hhaensel commentedon Jun 13, 2026
Today I could get PlotlyJS running!
It is still brittle, as the Blink version relies on JSExpr 0.5. But I've now moved to JSExpr 1, which is breaking.
The good thing is, I started from scratch from the master branches and made only minor changes. The strange thing is that Blink's
@jsmacro stopped working. Currently I import JSExpr and rely on JSExpr's@jsmacro, which works differently.But there's light at the end of the tunnel.