matplotlib-users Mailing List for matplotlib

Alan G Isaac wrote:
> On 11/28/2008 9:21 PM Eric Firing apparently wrote:
>> I suggest using twinx(); the scale for one line will be on the left, the 
>> scale for the other on the right.  You can make the scale colors match 
>> the line colors, if you want to.  I just updated 
>> examples/api/two_scales.py to show this.
> 
> 
> Hmmm, not here:
> http://matplotlib.sourceforge.net/examples/api/two_scales.html

The web page docs don't get updated automatically every time svn 
changes--which is probably a good thing.  The revised example is attached.

> And here the test on the y-axes does not display properly on the
> posted figures.

That's puzzling.  It looks like some odd rcParams settings must have 
been used when the figures were generated for the web page.

Eric
> 
> But I like the idea.
> 
> Alan Isaac
> 
> `
> 
> 
> 
> -------------------------------------------------------------------------
> 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-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

On 11/29/2008 7:48 AM TP apparently wrote:
> In my own application, I want to specialize matplotlib.lines.Line2D to be
> able to deal with units.
> Concerning this class, is there a direct means to plot instances?


There must be.  I have not done this, but looking at
http://matplotlib.sourceforge.net/api/artist_api.html#module-matplotlib.lines
it looks like you can set the ``axes`` property and looking at the base class
http://matplotlib.sourceforge.net/api/artist_api.html#matplotlib.artist.Artist
there is a ``draw`` method.

hth,
Alan Isaac

On Thu, Nov 27, 2008 at 11:21 AM, Mauro Cavalcanti <mau...@gm...> wrote:

> Then, in a checkbox event, I do the following:
>
> index = event.GetSelection()
> plot = self.plot_list[index]
> if self.FileList.IsChecked(index):
>       plot[0].set_visible(True)
> else:
>       plot[0].set_visible(False)
>
> This works quite well and points are properly toggled on/off the Basemap.
>
> But then I want to plot minimum spanning trees connecting the point
> datasets. For this I do:
>
> n = len(nodes)
> for i in range(n):
>       try:
>               t = edges[i,0]-1
>               u = edges[i,1]-1
>               x = [nodes[t,0], nodes[u,0]]
>               y = [nodes[t,1], nodes[u,1]]
>              self.plot_list.append(map.ax.plot(x,y,'-b'))
>       except:
>               continue
>
> where nodes is the array of point coordinates and edges are the
> from/to indexes of the coordinates, computed with Prim's algorithm.
> The lines between points (representing the minimum spanning tree) are
> displayed OK. But then my problem appears: I cannot find a way to turn
> the entire tree (which is composed of n-1 line segments) on and off of
> the map in the same way I do with the points (as shown above). What I
> would like is to store all line segments inside the for loop, and then
> show the tree at once, outside the loop; so I could use the same
> "plot[0].set_visible(True|False)" I use for the points.

Hi Mauro,

Yes, you should be able to use a collection for this quite easily.
One comment first.  You never want to try/except and catch all
exceptions w/o handling them in some way.  If you want to catch a
specific exception, fine, or of you want to catch all of them and log
them and then reraise, also fine, but there is not good use case for
catching them all and then continuing silently.

Now, on to collections.  In your example above, you would simply do::

    segments = []
    n = len(nodes)
    for i in range(n):
        t = edges[i,0]-1
        u = edges[i,1]-1
        xt, yt = nodes[t,0], nodes[t,1]   # assuming xt, yt are scalars here....
        xu, yu = nodes[u,0], nodes[u,1]
        segments.append( [ (xt,yt),  (xu,yu)  ] )

    collection = collections.LineCollection(segments)

and later::

    collection.set_visible(True|False)  # etc...

Hope this helps,

JDH

On Sat, Nov 29, 2008 at 6:48 AM, TP <par...@fr...> wrote:
> Hi everybody,
>
> In my own application, I want to specialize matplotlib.lines.Line2D to be
> able to deal with units.
> Concerning this class, is there a direct means to plot instances?
>
> For example, if I do:
>
>>>> a=matplotlib.lines.Line2D([5,6],[7,8],color='m')

You can add it directly to the Axes/Subplot instance with

    ax.add_line(a)

See also the http://matplotlib.sourceforge.net/examples/pylab_examples/clippedline.html
example which shows how to create a custom, derived line class.

JDH

Hi everybody,

In my own application, I want to specialize matplotlib.lines.Line2D to be
able to deal with units.
Concerning this class, is there a direct means to plot instances?

For example, if I do:

>>> a=matplotlib.lines.Line2D([5,6],[7,8],color='m')

To plot it, I have not found another way than doing:

>>> plot(a.get_xdata(), a.get_ydata())

But it is clumsy, because if I want to keep the line color:

>>> plot(a.get_xdata(), a.get_ydata(), a.get_color())

And so on for marker style, etc.

Is there a direct means to plot a Line2D instance?

Some time ago, a colleague wrote a "Curve_2D" class from scratch, to store a
2D curve, independently from matplotlib. Instead of that, I would like to be
able to specialize Line2D matplotlib class.

Am I compelled to write a .plot() method to Line2D?

In other words, when I do:

>>> a=plot((1,2),(2,3),'r--')

I obtain a Line2D instance:

>>> a[0]
<matplotlib.lines.Line2D instance at 0x954d08c>

I would like to do the contrary: define a Line2D instance, and then plot
this Line2D. The goal is to avoid re-inventing the wheel in my application,
as did my colleague.

Thanks a lot for your help

Julien

-- 
python -c "print ''.join([chr(154 - ord(c)) for c in '*9(9&(18%.9&1+,\'Z
(55l4('])"

"When a distinguished but elderly scientist states that something is
possible, he is almost certainly right. When he states that something is
impossible, he is very probably wrong." (first law of AC Clarke)

On 11/28/2008 9:21 PM Eric Firing apparently wrote:
> I suggest using twinx(); the scale for one line will be on the left, the 
> scale for the other on the right.  You can make the scale colors match 
> the line colors, if you want to.  I just updated 
> examples/api/two_scales.py to show this.


Hmmm, not here:
http://matplotlib.sourceforge.net/examples/api/two_scales.html
And here the test on the y-axes does not display properly on the
posted figures.

But I like the idea.

Alan Isaac

`

Eric Firing, on 2008-11-28 17:55, wrote:
> Paul Ivanov wrote:
>> Hi Jesper,
>>
>> confirming the problem over here, as well. both numpy and mpl from svn
>> (also on an Ubuntu 8.04 machine).
> 
> Just checking: did you do a clean build of numpy (deleting any old build
> directory) and then a clean build of mpl?  Distutils is not very smart
> about dependencies.
> 
> Eric

Yeah, just checked again against numpy 1.3.0.dev6118 and mpl r6456 after
clearing everything.

Paul

> 
>>
>> good luck,
>> Paul
>>
>>
>> Jesper Larsen, on 2008-11-27 02:10, wrote:
>>> Hi Eric and Mauro,
>>>
>>> Thanks for your answers.
>>>
>>> 2008/11/27 Eric Firing <ef...@ha...>:
>>>> It looks OK to me with mpl and numpy from svn.
>>> I tried upgrading to numpy from svn as well. Unfortunately the problem
>>> persists (I have attached a plot). I have seen the problem on two of
>>> my Ubuntu machines. Maybe it is caused by my specific setup and
>>> supporting libraries.
>>>
>>> Since I have a working solution and it does not seem to affect others
>>> (based on a survey of two:-) let us just leave the problem for now. If
>>> someone else encounter it please let me know and I will try to dive a
>>> bit into the issue. If the problem turns up again when I have a need
>>> to upgrade numpy (which is probably when matplotlib requires me to) I
>>> will also look into it.
>>>
>>> Best regards,
>>> Jesper
>>>
>>>
>>> ------------------------------------------------------------------------
>>>
>>>
>>> ------------------------------------------------------------------------
>>>
>>> -------------------------------------------------------------------------
>>>
>>> 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-users mailing list
>>> Mat...@li...
>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>
>>
>> -------------------------------------------------------------------------
>> 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-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 

Robin wrote:
> Hi,
> 
> I have a plot containing two lines that are quite far apart - ie one
> line oscillates around y=2, the other around y=10. The osciallations
> are small, but I would like to show the detail better (while having
> htem in a single plot).

I suggest using twinx(); the scale for one line will be on the left, the 
scale for the other on the right.  You can make the scale colors match 
the line colors, if you want to.  I just updated 
examples/api/two_scales.py to show this.

Eric

> 
> So I thought it would be nice if the y-axis scale went from 1-3 then
> had a break (denoted with some kind of cross mark), then went from
> 9-11. I tried googling but I'm not really sure what the official name
> for such a thing is, if there is one.
> 
> Is it possible to get this sort of effect with matplotlib? Or can
> people suggest an alternative (I guess I will look at doing 2 subplots
> one above the other very close together).
> 
> Cheers
> 
> Robin
> 
> -------------------------------------------------------------------------
> 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-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Paul Ivanov wrote:
> Hi Jesper,
> 
> confirming the problem over here, as well. both numpy and mpl from svn
> (also on an Ubuntu 8.04 machine).

Just checking: did you do a clean build of numpy (deleting any old build 
directory) and then a clean build of mpl?  Distutils is not very smart 
about dependencies.

Eric

> 
> good luck,
> Paul
> 
> 
> Jesper Larsen, on 2008-11-27 02:10, wrote:
>> Hi Eric and Mauro,
>>
>> Thanks for your answers.
>>
>> 2008/11/27 Eric Firing <ef...@ha...>:
>>> It looks OK to me with mpl and numpy from svn.
>> I tried upgrading to numpy from svn as well. Unfortunately the problem
>> persists (I have attached a plot). I have seen the problem on two of
>> my Ubuntu machines. Maybe it is caused by my specific setup and
>> supporting libraries.
>>
>> Since I have a working solution and it does not seem to affect others
>> (based on a survey of two:-) let us just leave the problem for now. If
>> someone else encounter it please let me know and I will try to dive a
>> bit into the issue. If the problem turns up again when I have a need
>> to upgrade numpy (which is probably when matplotlib requires me to) I
>> will also look into it.
>>
>> Best regards,
>> Jesper
>>
>>
>> ------------------------------------------------------------------------
>>
>>
>> ------------------------------------------------------------------------
>>
>> -------------------------------------------------------------------------
>> 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-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
> 
> -------------------------------------------------------------------------
> 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-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users