Patch for scatter plot legend enhancement

Attached is a diff against revision 6115 that contains a patch to
improve the behavior of the legend function when showing legends for
Collections (e.g. pyplot.scatter plots). The implemented behavior
shows three sample scatter points, adapting the properties and showing
three different colors or sizes in the even that either vary across
the points.

However, this patch is not entirely correct - I've attached a testing
script and an example of the output - as you can see, with a number of
scatter plots, the alignment between the scatter symbols and the text
is off in the legend. I suspect this has to do with the
get_window_extent(self,render) function I had to add to the Collection
class... I'm really not entirely sure what that function is supposed
to do (or what coordinates it's supposed to be transformed into). Can
anyone provide some guidance as to what should be done to fix this
vertical offset problem?

scatleg.diff (2.54 KB)

scatter_test.py (641 Bytes)

scatleg.png

Erik,

I haven't had a chance to get to this yet. Could you please also post
it on the sf patch tracker so it doesn't get dropped, and ping us with
a reminder in a few days if nothing has happened....

JDH

···

On Tue, Sep 23, 2008 at 12:20 AM, Erik Tollerud <erik.tollerud@...149...> wrote:

Attached is a diff against revision 6115 that contains a patch to
improve the behavior of the legend function when showing legends for

Sorry for the dealyed reply - I've been out of town... I posted to the
patch tracker, and am dutifully pinging :slight_smile:

···

On Tue, Sep 23, 2008 at 11:41 AM, John Hunter <jdh2358@...149...> wrote:

On Tue, Sep 23, 2008 at 12:20 AM, Erik Tollerud <erik.tollerud@...149...> wrote:

Attached is a diff against revision 6115 that contains a patch to
improve the behavior of the legend function when showing legends for

Erik,

I haven't had a chance to get to this yet. Could you please also post
it on the sf patch tracker so it doesn't get dropped, and ping us with
a reminder in a few days if nothing has happened....

JDH

Does anyone have anything new here? I'm perfectly willing to
experiment, but I'm really at a loss as to what
get_window_extent(self,render) is supposed to do (clearly get some
window extent, but exactly what window and what coordinates the extent
is in is what is confusing me).

···

On Tue, Sep 23, 2008 at 11:41 AM, John Hunter <jdh2358@...149...> wrote:

On Tue, Sep 23, 2008 at 12:20 AM, Erik Tollerud <erik.tollerud@...149...> wrote:

Attached is a diff against revision 6115 that contains a patch to
improve the behavior of the legend function when showing legends for

Erik,

I haven't had a chance to get to this yet. Could you please also post
it on the sf patch tracker so it doesn't get dropped, and ping us with
a reminder in a few days if nothing has happened....

JDH

Hi Eric,

As far as I know, get_window_extent is meant to return the extent of
the object in the display coordinate. And, as you may have noticed,
often this is used to calculate the relative position of objects.

I quickly went through your patch and my guess is your implementation
of get_window_extent is correct in this regard, but I haven't
considered this seriously so I may be wrong.

On the other hand, I guess the original problem you had is not related
with the get_window_extents() method.
The legend class has a _update_positions() method which is called
before the legends are drawn. And you have to update positions of your
handles within this method.

In the simple patch below. I tried to implement some basic update code
for the polycollections ( I also slightly adjusted the y-offsets. This
is just my personal preference). See if this patch works for you.

Regards,

-JJ

Index: lib/matplotlib/legend.py

···

===================================================================
--- lib/matplotlib/legend.py (revision 6163)
+++ lib/matplotlib/legend.py (working copy)
@@ -532,6 +540,12 @@
             elif isinstance(handle, Rectangle):
                 handle.set_y(y+1/4*h)
                 handle.set_height(h/2)
+ elif isinstance(handle, RegularPolyCollection):
+ offsets = handle.get_offsets()
+ xvals = [x for (x, _) in offsets]
+ yy = y + h
+ yvals=[yy-4./8*h,yy-3./8*h,yy-4./8*h]
+ handle.set_offsets(zip(xvals, yvals))

         # Set the data for the legend patch
         bbox = self._get_handle_text_bbox(renderer)

On Tue, Oct 7, 2008 at 12:15 AM, Erik Tollerud <erik.tollerud@...149...> wrote:

Does anyone have anything new here? I'm perfectly willing to
experiment, but I'm really at a loss as to what
get_window_extent(self,render) is supposed to do (clearly get some
window extent, but exactly what window and what coordinates the extent
is in is what is confusing me).

On Tue, Sep 23, 2008 at 11:41 AM, John Hunter <jdh2358@...149...> wrote:

On Tue, Sep 23, 2008 at 12:20 AM, Erik Tollerud <erik.tollerud@...149...> wrote:

Attached is a diff against revision 6115 that contains a patch to
improve the behavior of the legend function when showing legends for

Erik,

I haven't had a chance to get to this yet. Could you please also post
it on the sf patch tracker so it doesn't get dropped, and ping us with
a reminder in a few days if nothing has happened....

JDH

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Matplotlib-devel mailing list
Matplotlib-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

Ah, that makes more sense Jae-Joon - thanks!

With this addition it all looks fine to me - I've attached a patch
against 6174 that does everything I'd like to see it do... note that
I've left in the TODO in collections.py simply because I still am not
certain that what I'm doing there is correct in all use-cases...

scatleg.diff (3.16 KB)

···

On Mon, Oct 6, 2008 at 11:22 PM, Jae-Joon Lee <lee.j.joon@...149...> wrote:

Hi Eric,

As far as I know, get_window_extent is meant to return the extent of
the object in the display coordinate. And, as you may have noticed,
often this is used to calculate the relative position of objects.

I quickly went through your patch and my guess is your implementation
of get_window_extent is correct in this regard, but I haven't
considered this seriously so I may be wrong.

On the other hand, I guess the original problem you had is not related
with the get_window_extents() method.
The legend class has a _update_positions() method which is called
before the legends are drawn. And you have to update positions of your
handles within this method.

In the simple patch below. I tried to implement some basic update code
for the polycollections ( I also slightly adjusted the y-offsets. This
is just my personal preference). See if this patch works for you.

Regards,

-JJ

Index: lib/matplotlib/legend.py

--- lib/matplotlib/legend.py (revision 6163)
+++ lib/matplotlib/legend.py (working copy)
@@ -532,6 +540,12 @@
            elif isinstance(handle, Rectangle):
                handle.set_y(y+1/4*h)
                handle.set_height(h/2)
+ elif isinstance(handle, RegularPolyCollection):
+ offsets = handle.get_offsets()
+ xvals = [x for (x, _) in offsets]
+ yy = y + h
+ yvals=[yy-4./8*h,yy-3./8*h,yy-4./8*h]
+ handle.set_offsets(zip(xvals, yvals))

        # Set the data for the legend patch
        bbox = self._get_handle_text_bbox(renderer)

On Tue, Oct 7, 2008 at 12:15 AM, Erik Tollerud <erik.tollerud@...149...> wrote:

Does anyone have anything new here? I'm perfectly willing to
experiment, but I'm really at a loss as to what
get_window_extent(self,render) is supposed to do (clearly get some
window extent, but exactly what window and what coordinates the extent
is in is what is confusing me).

On Tue, Sep 23, 2008 at 11:41 AM, John Hunter <jdh2358@...149...> wrote:

On Tue, Sep 23, 2008 at 12:20 AM, Erik Tollerud <erik.tollerud@...149...> wrote:

Attached is a diff against revision 6115 that contains a patch to
improve the behavior of the legend function when showing legends for

Erik,

I haven't had a chance to get to this yet. Could you please also post
it on the sf patch tracker so it doesn't get dropped, and ping us with
a reminder in a few days if nothing has happened....

JDH

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Matplotlib-devel mailing list
Matplotlib-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

--
Erik Tollerud
Graduate Student
Center For Cosmology
Department of Physics and Astronomy
2142 Frederick Reines Hall
University of California, Irvine
Office Phone: (949)824-2587
Cell: (651)307-9409
etolleru@...244...

Jae-Joon -- could you handle this patch submission?

Thanks,
JDH

···

On Wed, Oct 8, 2008 at 10:27 PM, Erik Tollerud <erik.tollerud@...149...> wrote:

Ah, that makes more sense Jae-Joon - thanks!

Okay.
I may work on it sometime tomorrow or during the weekends.

Regards,

-JJ

···

On Thu, Oct 9, 2008 at 12:54 PM, John Hunter <jdh2358@...149...> wrote:

On Wed, Oct 8, 2008 at 10:27 PM, Erik Tollerud <erik.tollerud@...149...> wrote:

Ah, that makes more sense Jae-Joon - thanks!

Jae-Joon -- could you handle this patch submission?

Thanks,
JDH

John Hunter wrote:

Ah, that makes more sense Jae-Joon - thanks!

Jae-Joon -- could you handle this patch submission?

Thanks,
JDH

Hi,
  I also had a look at this patch -- as that's a feature I was
interested in, too. There are a few things that I think need to be
changed; just want to share my thoughts:

- the parameter numpoints should be used (it's ignored right now)

- Some private variables are accessed and a new RegularPolycollection is
created (does this work eg. with a StarPolygonCollection? I haven't
checked, but I don't think so !). Instead of creating a new
RegularPolyCollection it might be more useful to make a copy of the
existing object... I was thinking about a update_from() method for the
Collection class(es) similar to update_from() for lines.

mm

···

On Wed, Oct 8, 2008 at 10:27 PM, Erik Tollerud <erik.tollerud@...149...> wrote:

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Matplotlib-devel mailing list
Matplotlib-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

- the parameter numpoints should be used (it's ignored right now)

Thanks Manuel. I guess we can simply reuse xdata_marker for this purpose.

- Some private variables are accessed and a new RegularPolycollection is
created (does this work eg. with a StarPolygonCollection? I haven't
checked, but I don't think so !). Instead of creating a new
RegularPolyCollection it might be more useful to make a copy of the
existing object... I was thinking about a update_from() method for the
Collection class(es) similar to update_from() for lines.

By changing "RegularPolyCoolection" to "type(handles)", it works for
StarPolygonCollection.

In Erik's current implementation, the markers in the legend have
varying colors, sizes, and y offsets.
The color variation seems fine. But do we need to vary the sizes and
y-offsets? My inclination is to use a fixed size (median?) and a fixed
y offset. How does Erik and others think?

Regards,

-JJ

Jae-Joon Lee wrote:

- the parameter numpoints should be used (it's ignored right now)

Thanks Manuel. I guess we can simply reuse xdata_marker for this purpose.

- Some private variables are accessed and a new RegularPolycollection is
created (does this work eg. with a StarPolygonCollection? I haven't
checked, but I don't think so !). Instead of creating a new
RegularPolyCollection it might be more useful to make a copy of the
existing object... I was thinking about a update_from() method for the
Collection class(es) similar to update_from() for lines.

By changing "RegularPolyCoolection" to "type(handles)", it works for
StarPolygonCollection.
In Erik's current implementation, the markers in the legend have
varying colors, sizes, and y offsets.
The color variation seems fine.

+1

But do we need to vary the sizes and
y-offsets? My inclination is to use a fixed size (median?) and a fixed
y offset. How does Erik and others think?

I also would prefer fixed sizes for the legend, but maybe one should
ensure that the markers are not too large (larger than text height)?

mm

···

Regards,

-JJ

Jae-Joon Lee wrote:

- the parameter numpoints should be used (it's ignored right now)

Thanks Manuel. I guess we can simply reuse xdata_marker for this purpose.

- Some private variables are accessed and a new RegularPolycollection is
created (does this work eg. with a StarPolygonCollection? I haven't
checked, but I don't think so !). Instead of creating a new
RegularPolyCollection it might be more useful to make a copy of the
existing object... I was thinking about a update_from() method for the
Collection class(es) similar to update_from() for lines.

By changing "RegularPolyCoolection" to "type(handles)", it works for
StarPolygonCollection.

In Erik's current implementation, the markers in the legend have
varying colors, sizes, and y offsets.
The color variation seems fine. But do we need to vary the sizes and
y-offsets? My inclination is to use a fixed size (median?) and a fixed
y offset. How does Erik and others think?

Regards,

-JJ

Attached is my current version of the patch. I've moved all of the
properties-copying stuff to collections, which makes the changes
legend.py more clearer (but I'm not fully happy with the patch and
haven't commit anything yet)

mm

scatleg.patch (3.01 KB)