Generalize GR tick label rotations

[BioTurboNick]

Resolves #3776

PR for reference images: JuliaPlots/PlotReferenceImages.jl#109

[BioTurboNick]

I think I have the offsets working well, let me know what you think.

[t-bltg]

I think I have the offsets working well, let me know what you think.

I have a few simplifications to propose, do you mind if I push the changes here after your commits ?
I don't want to mess things up 😉.

[BioTurboNick]

Sure!

[t-bltg]

Sure!

So I've simplified based on the assumption that the expected for y axis is what we do for x axis, rotated by 90°.

Also, I've removed the offsets (sorry about the time you spent on this), based on the look of ref29 which now looks like:

To me the 100 label is still to much pushed to the left:

, compared to the previous img http://docs.juliaplots.org/latest/generated/gr/#gr-ref29.

I you disagree on anything, I'd be happy to revert 😃.

[BioTurboNick]

Yeah, longer labels aren't adjusted the same, I'm seeing. Might be worth adding a graph with a long tick label to the test suite, like a DateTime, which would pick up on that.

Looking into the issue.

[t-bltg]

Yeah, longer labels aren't adjusted the same, I'm seeing. Might be worth adding a graph with a long tick label to the test suite, like a DateTime, which would pick up on that.

Looking into the issue.

To me, the rotated mwe and ref29 look OK without the offsets.

[BioTurboNick]

You convinced me it looks better with an offset haha. I think I have it working, and this generic positioning code allows the angle-based alignment code to be removed.

Only issue is the 3d plots' ticks aren't adjusted correctly anymore.

[isentropic]

Looks fine to me, but I'm not that proficient in GR

[t-bltg]

Pros:

• The new images at New images for Plots.jl #3782 PlotReferenceImages.jl#109 look good.

Cons:

• Pardon me for saying that, but the new code is a mess 😆 .

[BioTurboNick]

Oh it's definitely a mess.

I can try to simplify it, since the output is good.

[BioTurboNick]

I'm within inches of a much better positioning logic for especially 3d plots, but I keep tripping up on all the angles involved. I'll come back to this.

[BioTurboNick]

Well, I don't know if it is that much simpler to look at..., but the 3d plots positioning is now perfect IMO.

[BeastyBlacksmith]

@BioTurboNick can you make a PR with the most recent images?

[BioTurboNick]

It doesn't seem like there are any changes vs. the last one.

[BioTurboNick]

Oh I see, very tiny differences...

[BioTurboNick]

Okay I'm not sure what's going on. I keep seeing no difference on my end.

[BeastyBlacksmith]

@daschw, @t-bltg can you have a look if you can reproduce on your machine?

[codecov]

Codecov Report

Merging #3782 (ed9ed26) into master (a65cda8) will decrease coverage by 0.36%.
The diff coverage is 93.33%.

@@            Coverage Diff             @@
##           master    #3782      +/-   ##
==========================================
- Coverage   63.20%   62.84%   -0.37%     
==========================================
  Files          28       28              
  Lines        7491     7200     -291     
==========================================
- Hits         4735     4525     -210     
+ Misses       2756     2675      -81     

Impacted Files	Coverage Δ
src/recipes.jl	58.61% <60.00%> (-0.89%)	⬇️
src/colorbars.jl	90.19% <83.33%> (-0.55%)	⬇️
src/backends/gr.jl	88.85% <97.95%> (-0.41%)	⬇️
src/backends/unicodeplots.jl	74.13% <0.00%> (-1.26%)	⬇️
src/backends/pgfplotsx.jl	60.25% <0.00%> (-0.92%)	⬇️
src/args.jl	72.61% <0.00%> (-0.86%)	⬇️
src/output.jl	28.09% <0.00%> (-0.59%)	⬇️
src/pipeline.jl	90.09% <0.00%> (-0.55%)	⬇️
... and 15 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 9d56e72...ed9ed26. Read the comment docs.

[t-bltg]

1 passes, but I still don't know why ref40 does not pass on nightly. Oddly seems related to #3860.

[BeastyBlacksmith]

Thanks. Then we take it as is for now and need to look into 40 on nightly in another PR

[BioTurboNick]

Thanks all!