matplotlib-users

[Matplotlib-users] Omitting curves from a legend

[Bill B. <wb...@gm...>]

There are a couple things about legend that I'm finding a little
irksome.  Is there some better way to do this?

1) if you have a contour, legend() wants to add all the contours to
the list.  calling contour(...,label='_nolegend_') doesn't seem to
help.

I'm trying to plot a bunch of different types of markers on top of a
contour plot, and put just the markers in the legend.  The workaround
seems to be
    contour(. . . )
    p1 = plot( . . .)
    p2 = plot(. . .)
    . . .
    pN = plot(. . .)

    legend((p1,p2,...,pN), ('Legend1', 'Legend2', ... , 'LegendN'))
    Basically I have to save the plot data and repeat myself later.

2) The '_nolegend_' trick doesn't work for things in the simple call
style for legend()

     legend('_nolegend_', 'Data 1', 'Data 2')

That makes a legend containing the string "_nolegend_".

My first hunch to do that before looking at the docs at all was to put
in a None for a legend I wanted to leave off.  That's raises an error
though.  Might be nice if it was made a legal way to omit something
from the legend.

--bb

Re: [Matplotlib-users] Omitting curves from a legend

[John H. <jd...@gm...>]

On 4/13/07, Bill Baxter <wb...@gm...> wrote:
> There are a couple things about legend that I'm finding a little
> irksome.  Is there some better way to do this?
>
> 1) if you have a contour, legend() wants to add all the contours to
> the list.  calling contour(...,label='_nolegend_') doesn't seem to
> help.

You should be able to set the "_nolegend_" label property on the
contour set line collections like so:

  >>> cs = contour(...blah...)

  >>> for coll in cs.collections:
            coll.set_label('_nolegend_')

of use "setp" for the same purpose.

  >>> setp(cs.collections, label='_nolegend_')

contour doesn't use the kwargs to set the line collection properties,
which is why it is not working in the contour commands.  some plot
commands use the kwargs to update the artist properties that the plot
command creates, some do not, and the only way to know is the read the
individual docstrings of the commands.

Let me know if this works because it is untested.

JDH

Re: [Matplotlib-users] Omitting curves from a legend

[Eric F. <ef...@ha...>]

Maybe I should make _nolegend_ the default for contour and contourf 
collections?

Eric

John Hunter wrote:
> On 4/13/07, Bill Baxter <wb...@gm...> wrote:
>> There are a couple things about legend that I'm finding a little
>> irksome.  Is there some better way to do this?
>>
>> 1) if you have a contour, legend() wants to add all the contours to
>> the list.  calling contour(...,label='_nolegend_') doesn't seem to
>> help.
> 
> You should be able to set the "_nolegend_" label property on the
> contour set line collections like so:
> 
>   >>> cs = contour(...blah...)
> 
>   >>> for coll in cs.collections:
>             coll.set_label('_nolegend_')
> 
> of use "setp" for the same purpose.
> 
>   >>> setp(cs.collections, label='_nolegend_')
> 
> contour doesn't use the kwargs to set the line collection properties,
> which is why it is not working in the contour commands.  some plot
> commands use the kwargs to update the artist properties that the plot
> command creates, some do not, and the only way to know is the read the
> individual docstrings of the commands.
> 
> Let me know if this works because it is untested.

Re: [Matplotlib-users] Omitting curves from a legend

[Bill B. <wb...@gm...>]

Ok.   Thanks.  I'll give the setp on the ContourSet thing a try.

Documentation issue/question:  I figured there was probably some way
to set attributes individually using the return value from contour
since contour's docstring helpfully tells me that countour returns a
ContourSet object.  However,  'ContourSet?' in ipython gives me
nothing.  Similarly with plot(), it says it returns a 'list of lines'
but that is not so useful since I can't look up the docstring for
'list of lines'.

--bb

On 4/17/07, Eric Firing <ef...@ha...> wrote:
> Maybe I should make _nolegend_ the default for contour and contourf
> collections?
>
> Eric
>
> John Hunter wrote:
> > On 4/13/07, Bill Baxter <wb...@gm...> wrote:
> >> There are a couple things about legend that I'm finding a little
> >> irksome.  Is there some better way to do this?
> >>
> >> 1) if you have a contour, legend() wants to add all the contours to
> >> the list.  calling contour(...,label='_nolegend_') doesn't seem to
> >> help.
> >
> > You should be able to set the "_nolegend_" label property on the
> > contour set line collections like so:
> >
> >   >>> cs = contour(...blah...)
> >
> >   >>> for coll in cs.collections:
> >             coll.set_label('_nolegend_')
> >
> > of use "setp" for the same purpose.
> >
> >   >>> setp(cs.collections, label='_nolegend_')
> >
> > contour doesn't use the kwargs to set the line collection properties,
> > which is why it is not working in the contour commands.  some plot
> > commands use the kwargs to update the artist properties that the plot
> > command creates, some do not, and the only way to know is the read the
> > individual docstrings of the commands.
> >
> > Let me know if this works because it is untested.
>

Re: [Matplotlib-users] Omitting curves from a legend

[<jk...@ik...>]

"Bill Baxter" <wb...@gm...> writes:

> Documentation issue/question:  I figured there was probably some way
> to set attributes individually using the return value from contour
> since contour's docstring helpfully tells me that countour returns a
> ContourSet object.  However,  'ContourSet?' in ipython gives me
> nothing.  Similarly with plot(), it says it returns a 'list of lines'
> but that is not so useful since I can't look up the docstring for
> 'list of lines'.

Call setp on the returned object:

In [8]:y = plot(sin(arange(10)))

In [9]:setp(y)
  alpha: float
  animated: [True | False]
  antialiased or aa: [True | False]
  axes: unknown
...
  xdata: array
  ydata: array
  zorder: any number

And similarly:

In [11]:z = contour(outerproduct(arange(10), arange(10)))

In [12]:setp(z)
  alpha: unknown
  array: unknown
  clim: a length 2 sequence of floats
  cmap: a colormap
  colorbar: unknown
  label_props: unknown
  norm: unknown

Or, in ipython you can type z.set_<TAB> to see the completions. With
the list returned from plot you cannot simply type y[0].set_<TAB> but
have to assign the value of y[0] to a variable first.

-- 
Jouni K. Seppänen
http://www.iki.fi/jks

Re: [Matplotlib-users] Omitting curves from a legend

[Eric F. <ef...@ha...>]

Bill Baxter wrote:
> There are a couple things about legend that I'm finding a little
> irksome.  Is there some better way to do this?
> 
> 1) if you have a contour, legend() wants to add all the contours to
> the list.  calling contour(...,label='_nolegend_') doesn't seem to
> help.

I think it would be quite unusual that someone would want contour lines 
to show up in a legend, so I made the change I suggested in an earlier 
response to this thread: the LineCollections in the ContourSet now have 
their labels set to _nolegend_.  If someone really does want contour 
lines in a legend, these labels still can be changed manually, as 
described earlier in this thread.

Eric

[Matplotlib-users] Omitting curves from a legend

[Bill B. <wb...@gm...>]

On 4/18/07, Eric Firing <ef...@ha...> wrote:
> Bill Baxter wrote:
> > There are a couple things about legend that I'm finding a little
> > irksome.  Is there some better way to do this?
> >
> > 1) if you have a contour, legend() wants to add all the contours to
> > the list.  calling contour(...,label='_nolegend_') doesn't seem to
> > help.
>
> I think it would be quite unusual that someone would want contour lines
> to show up in a legend, so I made the change I suggested in an earlier
> response to this thread: the LineCollections in the ContourSet now have
> their labels set to _nolegend_.  If someone really does want contour
> lines in a legend, these labels still can be changed manually, as
> described earlier in this thread.

You can also use colorbar() to show the levels if you want that
information to be visible, and that's probably more appropriate than a
legend anyway.  I think it's a good change.  Thanks!

--bb