legend(loc='best') a bit off

In the attached, the legend box slightly covers the graph, but there is no
need to, it could just be moved a bit lower and then would cover nothing.

My questions:
1) can I fix it?
2) is this a bug?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: saturated_ascma_paper_rot.pdf
Type: application/pdf
Size: 30071 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20170109/2d15af29/attachment-0001.pdf>

I am pretty sure that 'best' is picking from one of the 9 positions [(top,
middle, bottom)X(left, center, right)] not picking a continuous position.

Fixing this would be useful and a pretty major new feature. It is not
clear to me what the best algorithm for this would be. The obvious thing
is to run N more refinements around each point, but hopefully there is some
computational geometry that I do not know.

Tom

···

On Mon, Jan 9, 2017, 09:07 Neal Becker <ndbecker2 at gmail.com> wrote:

In the attached, the legend box slightly covers the graph, but there is no
need to, it could just be moved a bit lower and then would cover nothing.

My questions:
1) can I fix it?
2) is this a bug?_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
https://mail.python.org/mailman/listinfo/matplotlib-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20170109/64c1a7ad/attachment.html>

Would an approach similar to adjustText be something you could consider for
mpl? I was already thinking of adding this to the library.

9 ??? 2017 ?. 15:15 ??? "Thomas Caswell" <tcaswell at gmail.com>
???:

I am pretty sure that 'best' is picking from one of the 9 positions [(top,
middle, bottom)X(left, center, right)] not picking a continuous position.

Fixing this would be useful and a pretty major new feature. It is not
clear to me what the best algorithm for this would be. The obvious thing
is to run N more refinements around each point, but hopefully there is some
computational geometry that I do not know.

Tom

In the attached, the legend box slightly covers the graph, but there is no
need to, it could just be moved a bit lower and then would cover nothing.

My questions:
1) can I fix it?
2) is this a bug?_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
https://mail.python.org/mailman/listinfo/matplotlib-users

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
https://mail.python.org/mailman/listinfo/matplotlib-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20170109/da0305bd/attachment.html>

···

On Mon, Jan 9, 2017, 09:07 Neal Becker <ndbecker2 at gmail.com> wrote:

In principle yes, but I think it would depend on how general / reliable the
algorithm is vs how complex and invasive the algorithm ends up being.

Tom

···

On Mon, Jan 9, 2017 at 2:40 PM Ilya Flyamer <flyamer at gmail.com> wrote:

Would an approach similar to adjustText be something you could consider
for mpl? I was already thinking of adding this to the library.

9 ??? 2017 ?. 15:15 ??? "Thomas Caswell" <tcaswell at gmail.com>
???:

I am pretty sure that 'best' is picking from one of the 9 positions [(top,
middle, bottom)X(left, center, right)] not picking a continuous position.

Fixing this would be useful and a pretty major new feature. It is not
clear to me what the best algorithm for this would be. The obvious thing
is to run N more refinements around each point, but hopefully there is some
computational geometry that I do not know.

Tom

On Mon, Jan 9, 2017, 09:07 Neal Becker <ndbecker2 at gmail.com> wrote:

In the attached, the legend box slightly covers the graph, but there is no
need to, it could just be moved a bit lower and then would cover nothing.

My questions:
1) can I fix it?
2) is this a bug?_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
https://mail.python.org/mailman/listinfo/matplotlib-users

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
https://mail.python.org/mailman/listinfo/matplotlib-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20170110/d87679f7/attachment-0001.html>

OK, when I find the time to add it to adjustText, I'll get back to you to
discuss this.

2017-01-10 18:22 GMT+00:00 Thomas Caswell <tcaswell at gmail.com>:

In principle yes, but I think it would depend on how general / reliable
the algorithm is vs how complex and invasive the algorithm ends up being.

Tom

Would an approach similar to adjustText be something you could consider
for mpl? I was already thinking of adding this to the library.

9 ??? 2017 ?. 15:15 ??? "Thomas Caswell" <tcaswell at gmail.com>
???:

I am pretty sure that 'best' is picking from one of the 9 positions
[(top, middle, bottom)X(left, center, right)] not picking a continuous
position.

Fixing this would be useful and a pretty major new feature. It is not
clear to me what the best algorithm for this would be. The obvious thing
is to run N more refinements around each point, but hopefully there is some
computational geometry that I do not know.

Tom

In the attached, the legend box slightly covers the graph, but there is no
need to, it could just be moved a bit lower and then would cover nothing.

My questions:
1) can I fix it?
2) is this a bug?_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
https://mail.python.org/mailman/listinfo/matplotlib-users

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
https://mail.python.org/mailman/listinfo/matplotlib-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20170110/d87ac81b/attachment.html>

···

On Mon, Jan 9, 2017 at 2:40 PM Ilya Flyamer <flyamer at gmail.com> wrote:

On Mon, Jan 9, 2017, 09:07 Neal Becker <ndbecker2 at gmail.com> wrote: