New recipe: lens #2372

BeastyBlacksmith · 2020-01-28T22:28:55Z

The lens recipe should magnify a certain region and show it in an inset plot.
Current API is

pl = plot((1:5).^2)
plot!(pl, [1.2, 2.8], [3.4, 7.8], seriestype=:lens, inset=(1,bbox(0.4,0.,0.5,0.5)))

Where you pass the magnifying region as x and y values and the location of the inset as usually.
Currently only relative bounding boxes are supported. Location of the linking line has to be improved.
Also drawing of the linking line should actually be outside of any plot, so it doesn't infer with axis limits of the parent plot, but I don't know if this is possible.
fixes #541

mkborregaard · 2020-01-29T08:08:04Z

src/recipes.jl

@@ -801,6 +801,50 @@ end
 # note: don't add dependencies because this really isn't a drop-in replacement


+# ---------------------------------------------------------------------------
+# lens! - magnify a region of a plot
+@recipe function f(::Type{Val{:lens}}, plt::AbstractPlot)


I'm surprised this works - it's the type recipe signature, which is usually just used for type conversions to vector. Should this not be a "user recipe" with a @userplot shorthand defined?

I thought, this is the signature of a plot recipe ( and a valid usecase )

also I would like to only define the lens! shorthand, since lens by itself doesn't make too much sense.

Ah you're right. I don't think they are used anywhere in the ecosystem, but I may be wrong (it lists MarginalHist as an example, but looking down on that page clearly shows that marginalhist is a "user recipe"). Defining a lens! shorthand makes sense - but I feel that calling it with seriestype as done here feels unideomatic.

What would be a more ideomatic alternative? I don't bother too much to change this.

Just the lens! shorthand? Sorry I have zero experience using plot recipes

Okay, now I got it. You meant the way to call the recipe is unideomatic, I thought the implementation with using the seriestype. I will add the lens! shorthand.

BeastyBlacksmith · 2020-01-30T22:08:19Z

I thought, adding a ghost subplot for the connecting line would help, but, since the bounding box of the parent box seems to get calculated afterwards it still affects his axis limits. Also some backends have problems if the inset plot is outside of the parent plot.

BeastyBlacksmith · 2020-01-30T22:18:15Z

Easiest course of action would be to get rid of the connecting line altogether and just have the box + inset.

BeastyBlacksmith · 2020-01-31T22:27:32Z

This is the best I can do without having access to the bounding box of the parent plot.
Please give it a try.

mkborregaard · 2020-02-29T11:35:59Z

@BeastyBlacksmith are you waiting for me here?
BTW would this close #541 ?

BeastyBlacksmith · 2020-03-01T16:46:14Z

@mkborregaard Would be nice if you can confirm that I can't get the bounding box of the current plot at the recipe stage or if I am looking for the wrong thing and layout information is stored elsewhere.
Plus I am looking for a way to have the keywords passed be applied for only for the inset. I can make only parent or both but failed to do only inset.
Other than that #541 would be closed, yes.

mkborregaard · 2020-03-02T08:16:46Z

Whether the bounding box is present depends on the recipe type. With this being a type recipe, I'm pretty sure it isn't - if it were a series recipe it would be present as a key in a captured plotattributes object (as they are processed later in the pipeline where the bounding box is known), but then series recipes won't let you make another window.
With regards to the keywords - try pop!ing them from the plotattributes Dict and store them locally, them passing them in the inset.

BeastyBlacksmith · 2020-03-03T22:24:02Z

Ok, that worked. Could use some testing before merge.
An example is

pl = plot((1:5).^2)
lens!(pl, [1.2, 2.8], [3.4, 7.8], inset=(1,bbox(0.5,0.2,0.5,0.5)))

mkborregaard · 2020-03-03T22:29:00Z

You could add a testing image to examples maybe? That way it would become part of the testing suite - and be documented...

BeastyBlacksmith · 2020-03-04T14:22:11Z

I added the example and made sure, I can easily revert the formatting changes, if they don't belong here.

mkborregaard · 2020-03-04T14:28:56Z

src/examples.jl

@@ -311,6 +311,21 @@ const _examples = PlotExample[
        ],
    ),
    PlotExample(
+        "Lens",
+        "A lens lets you easyli magnify a region of a plot. x and y coordinates refer to the to be magnified region and the via the `inset` keyword the subplot index and the bounding box (in relative coordinates) of the inset plot with the magnified plot can be specified. Additional attributes count for the inset plot.",


mkborregaard · 2020-03-04T14:32:20Z

src/examples.jl

+            end
+        ],
+    ),
+        * PlotExample(


mkborregaard · 2020-03-04T14:32:47Z

Nice - would you mind splitting the formatting changes to a separate PR?
Are you familiar with how to add a figure to PlotReferenceImages to make tests pass? (examples double as tests)

BeastyBlacksmith · 2020-03-04T14:36:15Z

Are you familiar with how to add a figure to PlotReferenceImages to make tests pass? (examples double as tests)

No, never did this.

daschw · 2020-03-04T14:42:09Z

Firstly, I'd suggest putting new test_examples at the end of Plots._examples, because the test examples in PlotReferenceImages are named after their index in that Array. This way the proper images are compared to each other.

To update the PlotReferenceImages, you basically have to

]dev PlotReferenceImages, fork it, create a new branch, ...
Run ]test Plots
If images differ a GUI opens show ing the old and new image and asking you if you want to save the changes.
After tests finished, add new images in your PlotReferenceImages branch, push to fork and PR

After the PlotReferenceImages PR is merged, the tests of this PR will pass

mkborregaard reviewed Jan 29, 2020

View reviewed changes

BeastyBlacksmith marked this pull request as ready for review January 31, 2020 22:27

BeastyBlacksmith requested a review from mkborregaard January 31, 2020 22:27

BeastyBlacksmith force-pushed the bbs/lens-recipe branch from ce81c96 to 6f80c0e Compare March 3, 2020 22:22

mkborregaard reviewed Mar 4, 2020

View reviewed changes

src/examples.jl Outdated

end

],

),

* PlotExample(

Copy link

Member

mkborregaard Mar 4, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

stray *

BeastyBlacksmith force-pushed the bbs/lens-recipe branch from cd2b060 to 696ece1 Compare March 4, 2020 15:09

BeastyBlacksmith added 9 commits March 4, 2020 21:57

create recipe, add box and line

d231618

add subplot

fc4683c

improve linking

00f1fed

add ghost subplot

b886207

draw line only if inset is in subplot

96978cf

sp_bbox -> inset_bbox

bf619ee

better linking

53fc342

add lens! shorthand

fa4db31

apply keywords only to inset

4b9cfa4

BeastyBlacksmith force-pushed the bbs/lens-recipe branch from fe265ca to 3de69e3 Compare March 4, 2020 20:58

BeastyBlacksmith added 4 commits March 4, 2020 22:20

check bbox type

e816895

add lens example

54aaf58

fix typo and stray *

4de93dd

move legend out of the plot

ae065c9

BeastyBlacksmith force-pushed the bbs/lens-recipe branch from 3de69e3 to ae065c9 Compare March 4, 2020 21:21

move lens example down

4e2a60d

BeastyBlacksmith merged commit 55ea8b4 into master Mar 5, 2020

BeastyBlacksmith deleted the bbs/lens-recipe branch March 6, 2020 12:01

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

New recipe: lens #2372

New recipe: lens #2372

BeastyBlacksmith commented Jan 28, 2020 •

edited

Loading

mkborregaard Jan 29, 2020

BeastyBlacksmith Jan 29, 2020 •

edited

Loading

BeastyBlacksmith Jan 29, 2020

mkborregaard Jan 29, 2020

BeastyBlacksmith Jan 29, 2020

mkborregaard Jan 29, 2020 •

edited

Loading

BeastyBlacksmith Jan 29, 2020

BeastyBlacksmith commented Jan 30, 2020

BeastyBlacksmith commented Jan 30, 2020

BeastyBlacksmith commented Jan 31, 2020

mkborregaard commented Feb 29, 2020

BeastyBlacksmith commented Mar 1, 2020

mkborregaard commented Mar 2, 2020

BeastyBlacksmith commented Mar 3, 2020

mkborregaard commented Mar 3, 2020

BeastyBlacksmith commented Mar 4, 2020

mkborregaard Mar 4, 2020

mkborregaard Mar 4, 2020

mkborregaard commented Mar 4, 2020

BeastyBlacksmith commented Mar 4, 2020 •

edited

Loading

daschw commented Mar 4, 2020

New recipe: lens #2372

New recipe: lens #2372

Conversation

BeastyBlacksmith commented Jan 28, 2020 • edited Loading

mkborregaard Jan 29, 2020

Choose a reason for hiding this comment

BeastyBlacksmith Jan 29, 2020 • edited Loading

Choose a reason for hiding this comment

BeastyBlacksmith Jan 29, 2020

Choose a reason for hiding this comment

mkborregaard Jan 29, 2020

Choose a reason for hiding this comment

BeastyBlacksmith Jan 29, 2020

Choose a reason for hiding this comment

mkborregaard Jan 29, 2020 • edited Loading

Choose a reason for hiding this comment

BeastyBlacksmith Jan 29, 2020

Choose a reason for hiding this comment

BeastyBlacksmith commented Jan 30, 2020

BeastyBlacksmith commented Jan 30, 2020

BeastyBlacksmith commented Jan 31, 2020

mkborregaard commented Feb 29, 2020

BeastyBlacksmith commented Mar 1, 2020

mkborregaard commented Mar 2, 2020

BeastyBlacksmith commented Mar 3, 2020

mkborregaard commented Mar 3, 2020

BeastyBlacksmith commented Mar 4, 2020

mkborregaard Mar 4, 2020

Choose a reason for hiding this comment

mkborregaard Mar 4, 2020

Choose a reason for hiding this comment

mkborregaard commented Mar 4, 2020

BeastyBlacksmith commented Mar 4, 2020 • edited Loading

daschw commented Mar 4, 2020

BeastyBlacksmith commented Jan 28, 2020 •

edited

Loading

BeastyBlacksmith Jan 29, 2020 •

edited

Loading

mkborregaard Jan 29, 2020 •

edited

Loading

BeastyBlacksmith commented Mar 4, 2020 •

edited

Loading