matplotlib-users Mailing List for matplotlib (Page 5)

Hi, I'm also new to matplotlib, but this sounds like a general type of
problem you might in find in matlab too.  I think you are trying to plot two
separate lines, one for the first set of time and one for the second set of
time.  If you don't want them connected then you must issue two plot()
commands, which means you have to split the data into the sections that you
want connected.

 

Michael

 

  _____  

From: mat...@li...
[mailto:mat...@li...] On Behalf Of
Rodrigues, Joseph
Sent: Friday, July 21, 2006 11:25 AM
To: mat...@li...
Subject: [Matplotlib-users] iregular data set

 

I have a question (new user here): I have datasets of irregular time spaced
data and am having a problem with Matplot lib.

 

I am plotting two datasets: one is continuous time series data (simple line
chart) and the other irregular time series data.

 

The problem I have is: MatPlotLib is interpolating data for the irregular
data where there is no data.

 

So if I have data from 1985 - 1989 then no data for 1990 - 1994, matplotlib
is drawing a line connecting the missing data.

 

How does one plot irregular spaced time series data?

Thanks

Joseph

 

p.s.: if this is the wrong forum, could you let me know where to post?

I have a question (new user here): I have datasets of irregular time
spaced data and am having a problem with Matplot lib.

=20

I am plotting two datasets: one is continuous time series data (simple
line chart) and the other irregular time series data.

=20

The problem I have is: MatPlotLib is interpolating data for the
irregular data where there is no data.

=20

So if I have data from 1985 - 1989 then no data for 1990 - 1994,
matplotlib is drawing a line connecting the missing data.

=20

How does one plot irregular spaced time series data?

Thanks

Joseph

=20

p.s.: if this is the wrong forum, could you let me know where to post?

Hi all,

I don't like having to specify the marker on each call to plot(). So I
have something like this in my code:

import pylab as _p
_p.rc('lines', marker = 'o')

This works pretty well in the sense that each call to plot by default
uses the 'o' marker. However, now my graphs have markers on the axes
as well. I'm attaching a sample graph as an example.

So:

a) is this intended behavior? If so, whats the reasoning?
b) how can I turn it off? (btw, I just discovered that with
grid(True), the markers appear on each major tick. With grid(False),
they only appear at the corners. Doesn't solve the problem entirely
though).

Thanks,
Diwaker
-- 
Web/Blog/Gallery: http://floatingsun.net/blog

Please submit these kinds of bug reports to the sourceforge bug tracker. Minor 
problems like broken links do not need to be broadcast to over 480 people.

Thanks,
Darren


On Thursday 20 July 2006 11:54 pm, John Pye wrote:
> Hi,
>
> Just wanted to let you know that the link at to the Numeric Python
> tutorial at
>
>     http://matplotlib.sourceforge.net/tutorial.html
>
> is broken.
>
> Cheers
> JP

-- 
Darren S. Dale, Ph.D.
dd...@co...

Note: I just verified that this was introduced into 0.87.4.
0.87.3 doesn't exhibit the problem. See attachment.

Gary R.

gr...@bi... wrote:
> The following minimal script reveals a rendering problem with displaying a histogram on a log vertical axis.
> I'm using matplotlib0.87.4 in WinXP with python 2.3.5 Enthon.
> 
> from pylab import *
> hist(rand(100), 20, bottom=1)
> setp(gca(), yscale="log")
> show()
> 
> 
> Gary R.

The following minimal script reveals a rendering problem with displaying a histogram on a log vertical axis.
I'm using matplotlib0.87.4 in WinXP with python 2.3.5 Enthon.

from pylab import *
hist(rand(100), 20, bottom=1)
setp(gca(), yscale="log")
show()


Gary R.

Hi,

Just wanted to let you know that the link at to the Numeric Python
tutorial at

    http://matplotlib.sourceforge.net/tutorial.html

is broken.

Cheers
JP

-- 
John Pye
School of Mechanical and Manufacturing Engineering
The University of New South Wales
Sydney  NSW 2052  Australia
t +61 2 9385 5127
f +61 2 9663 1222
mailto:john.pye_AT_student_DOT_unsw.edu.au
http://pye.dyndns.org/

Hi, I've been using matplotlib for about a week now and am exceedingly
confused by all the class hierarchies and stuff.

 

I'm trying to draw an arc and some other polygons and render them to either
GTK or postscript.  I'm running Python 2.4 with Idle and XP

 

I can successfully do the following:

1.  Using the matlab-ish commands, I can create a figure, polygons, circles,
and use add_patch() to add these shapes to the figure.  

2.  I can direct it to GTK or PS with the matplotlib.use command.  I can
print out the resulting PS file just fine, and the plot shows up on the
screen with GTK just fine.

 

The problem is I want to draw arcs.  I don't see an Arc() command at the
same level as Polygon and Circle.  The only place I see an arc is with the
renderer classes.

 

I tried extracting the current renderer from the figure and axis (via the
_cachedRenderer attribute), but those don't seem to be assigned until I do a
savefig() or a show(), but obviously I want the renderer before any
show()ing or savefig()ing.

 

So then I tried creating a RendererPS class directly (instead of RendererAgg
as in the User's Guide).  It seems to want a pswriter argument.  I looked
around and saw that pswriter is expected to be some type of cString.StringIO
or something, which seemed too low-level, and I figured I didn't want to
mess around any more.  I just want to draw a stupid arc.

 

Also, if I have a GTK figure, and then do show(), it plots a nice figure on
the screen.  If I then close it, type show() again, my python window freezes
up and I have to kill it with the process manager.

 

If anyone can shed some light on these problems I'd be much happier :-)

 

Thanks

Michael

Hi,

I would like to visualize an image after some rotation: this means that
I would like to have to view each squared pixels as "rotated" (hence an
inclined square). I have in fact several images which I need to plot on
the same figure (with subplot), each of these having different "rotation
angles".

The rough solution would be to rotate the data itself (x and y) and use
imshow after some rebinning on a squared grid. But I would really like
to see the "true" pixels, hence the rotated squares...

Is it possible to do this in mpl?

thanks!
Eric

Hi,

since openoffice can use python as macro language, does
anyone know or has an idea, if is possible to use matplotlib
for plotting inside it. Would be nice to create diagrams
just be marking columns!?

Greetings!
 Fabian

>>>>> "Michael" == Michael Fitzgerald <mp...@ca...> writes:

    Michael> Hi all,

    Michael> I would like a way to set the marker face color to 'None'
    Michael> so that I may see overlapping markers.  I am aware of the
    Michael> workaround using scatter() to set alpha, but alas that's
    Michael> no good for PS output:

    Michael> 	http://sourceforge.net/mailarchive/message.php?msg_id=12261949

    Michael> Is this currently possible, or on the dev. radar?  If
    Michael> not, I'll file an official feature request.

You should be able to set the markerfacecolor property to 'None' the
string, as opposed to None, the python null object.  This is a bit
ugly, but None is used in the matplotlib properties to mean "do the
default, ie the rc setting".  If 

  plot(x, y, 'o', markerfacecolor='None')

fails, please post a complete example and we'll look into it.

JDH

Hi all,

I would like a way to set the marker face color to 'None' so that I  
may see overlapping markers.  I am aware of the workaround using  
scatter() to set alpha, but alas that's no good for PS output:

	http://sourceforge.net/mailarchive/message.php?msg_id=12261949

Is this currently possible, or on the dev. radar?  If not, I'll file  
an official feature request.

Mike

Is there a keyword or a function to lock the figures so that they can  
not be resized?

Cheers
Tommy
tg...@ma...
http://homepage.mac.com/tgrav/

"Any intelligent fool can make things bigger,
more complex, and more violent. It takes a
touch of genious -- and a lot of courage --
to move in the opposite direction"
                          -- Albert Einstein

Lionel,

Assuming you have the current version of mpl, use the "extend" kwarg:

     colorbar properties:

         extend=3D'neither', 'both', 'min', 'max'
                 If not 'neither', make pointed end(s) for out-of-ran=
ge
                 values.  These are set for a given colormap using th=
e
                 colormap set_under and set_over methods.

If the colorbar is attached to a filled contour plot, however, you do=
=20
not need to specify this kwarg in the colorbar call because it will b=
e=20
grabbed from the ContourSet object.

Eric


Lionel Roubeyrie wrote:
> HI,
> how do you set the arrows on the bottom and the top of your colorba=
r? I don't=20
> find any option for that?
> Thanks
>=20
> Le Vendredi 07 Juillet 2006 10:11, dd...@ja... a =C3=A9cr=
it :
>=20
>>On Friday 07 July 2006 4:11 pm, Lionel Roubeyrie wrote:
>>
>>>Hi all,
>>>Is it possible to import colormap file from GMT (.cpt) to basemap?
>>>If yes how can we do that?
>>>Thanks
>>
>>It's possible. Here's how I've been doing it.
>>
>>(1) I use the gmtcolormapPylab function defined in:
>>http://sourceforge.net/mailarchive/message.php?msg_id=3D11255878
>>
>>(2) Using the above function and a GMT cpt file (e.g.
>>GMT_no_green.cpt), I set a color dictionary as:
>>cdict =3D gmtcolormapPylab("GMT_no_green")
>>
>>(3) I define a colormap using:
>>mycmap =3D cm.colors.LinearSegmentedColormap("GMT_no_green",cdict,2=
56)
>>
>>(4) I apply the above using cmap=3Dmycolormap in contourf or pcolor=
mesh.
>>
>>By the way, in case you don't know of the following site, it has an
>>exhaustive selection of color palettes: http://cpt-city.org.uk/
>>
>>I've also attached an example of a basemap image using a cpt from t=
he
>>above site. Masked values not defined by the colormap are shown in
>>gray.
>=20
>=20

John,

You are correct about the way contour handles masked data, and I agree 
that it is ugly; unfortunately, it is inherent in the routine we are 
using.  You could reduce but not eliminate the problem by interpolating 
to a finer grid.

The routine we are using, which came from the gist package, has the 
virtues of an acceptable license, production of regular and filled 
contours, and handling of missing data (but with the problem you note, 
plus one bug that I have not been able to fix yet.)  I think it is also 
very efficient, although I have not done any benchmarking. If a better 
routine with the essential virtues could be found, we would probably use 
it; but I don't know of any.  In particular, open source routines that 
handle filled contours and masking seem to be very rare.

Eric

John Pye wrote:
> Hi all
> 
> I have an application requiring that I do contour plots for masked data.
> At the edge of my contour plots, I'm getting fragmented contours due to
> missing points on the corner of a cell in my grid.
> 
> Is there any way that the contour drawing code can be set up so that it
> plots contours even if a cell in the grid only has data on two sides?
> 
> I suspect that the rule for contour plots is that no contours are shown
> if *any* of the point on the four corners are missing.
> 
> There is an example with the jagged edge shown in the attached image.
> 
> Cheers
> JP
> 
> 
> 
> ------------------------------------------------------------------------
> 
> 
> ------------------------------------------------------------------------
> 
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys -- and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

>>>>> "Vinj" == Vinj Vinj <vin...@ya...> writes:

    Vinj> Some of the y axis labels end up having different fonts:
    Vinj> http://s3.amazonaws.com/oylb/wlsn_qm.png?20060717232235

    Vinj> In the above image, 7 and 7.5 have a larger font. I'm using
    Vinj> matplotlib 0.87.3 with ubuntu

Bizarre.  Are you using usetex?  It doesn't look like it from the
fonts.  Does it help to flush the ttf cache in
~/.matplotlib/ttffont.cache?  Is it possible you are making a call to
set the fontsize somewhere and are just picking up selected elements?

I'll need to see some code to help debug this one.

JDH

>>>>> "Michael" == Michael Fitzgerald <mp...@be...> writes:

    Michael> Hi all,

    Michael> I'm seeing some unexpected behavior when trying to hide a
    Michael> tick label, which I like to do with ganged plots.  When I
    Michael> attempt this, however, it seems the axis label is no
    Michael> longer drawn.  For example,

    Michael> x = arange(5) ax = subplot(111) ax.plot(x, x) xtl =
    Michael> ax.get_xticklabels() xtl[0].set_visible(False)
    Michael> ax.set_xlabel('foo')

    Michael> will correctly not draw the leftmost x tick label, but it
    Michael> won't draw the axis label ('foo') either.  Subsequently
    Michael> issuing ax.get_xaxis ().get_label().set_visible(True)
    Michael> doesn't help.  Any thoughts?

I just fixed this in svn -- the xlabel gets it's position by
examining the bounding boxes of all the tick labels, and wasn't
handling the invisible tick bounding box properly.  This is fixed in
svn.

Thanks for the report and example.
JDH

Hi all,

I'm seeing some unexpected behavior when trying to hide a tick label,  
which I like to do with ganged plots.  When I attempt this, however,  
it seems the axis label is no longer drawn.  For example,

x = arange(5)
ax = subplot(111)
ax.plot(x, x)
xtl = ax.get_xticklabels()
xtl[0].set_visible(False)
ax.set_xlabel('foo')

will correctly not draw the leftmost x tick label, but it won't draw  
the axis label ('foo') either.  Subsequently issuing ax.get_xaxis 
().get_label().set_visible(True) doesn't help.  Any thoughts?

Mike

>>>>> "Tommy" == Tommy Grav <tg...@ma...> writes:

    Tommy> I have a program that uses matplotlib to plot two images.
    Tommy> figim24 = figure(figsize=(5,5)) figim70 =
    Tommy> figure(figsize=(5,5))

    Tommy> I want an event loop that will be able to register which of
    Tommy> the two figures I mouse click in. When I only had one
    Tommy> figure I used

    Tommy> figim24.canvas.mpl_connect("button_press_event",clicker)

    Tommy> but that obviously only register events in figim24. Can
    Tommy> anyone guide me in the right direction for doing this?

You can either register two different callbacks (func1 and func2
below) or you can register both figures to the same callback (funcboth
below) and check which figure generated the event by inspecting the
figure.inaxes.figure attribute.  The latter approach only works if you
click over an axes,

from pylab import figure, show

fig1 = figure()
fig2 = figure()

ax1 = fig1.add_subplot(111)
ax2 = fig2.add_subplot(111)

def func1(event):
    print 'func1'

def func2(event):
    print 'func2'

def funcboth(event):
    if event.inaxes is None: return
    if event.inaxes.figure==fig1:
        print 'funcboth: 1'
    elif event.inaxes.figure==fig2:
        print 'funcboth: 2'

fig1.canvas.mpl_connect('button_press_event', func1)
fig2.canvas.mpl_connect('button_press_event', func2)

fig1.canvas.mpl_connect('button_press_event', funcboth)
fig2.canvas.mpl_connect('button_press_event', funcboth)

show()

> but when i try to plot the same at (0,2), i get an error:
> 
>     plot([0], [2],'o-'); show()

Well, this works perfectly fine for me (with mpl 0.87.4, rev 2558).
Maybe if you upgrade your mpl?

Cheers.

HI,
how do you set the arrows on the bottom and the top of your colorbar? I don=
't=20
find any option for that?
Thanks

Le Vendredi 07 Juillet 2006 10:11, dd...@ja... a =C3=A9crit=C2=A0:
> On Friday 07 July 2006 4:11 pm, Lionel Roubeyrie wrote:
> > Hi all,
> > Is it possible to import colormap file from GMT (.cpt) to basemap?
> > If yes how can we do that?
> > Thanks
>
> It's possible. Here's how I've been doing it.
>
> (1) I use the gmtcolormapPylab function defined in:
> http://sourceforge.net/mailarchive/message.php?msg_id=3D11255878
>
> (2) Using the above function and a GMT cpt file (e.g.
> GMT_no_green.cpt), I set a color dictionary as:
> cdict =3D gmtcolormapPylab("GMT_no_green")
>
> (3) I define a colormap using:
> mycmap =3D cm.colors.LinearSegmentedColormap("GMT_no_green",cdict,256)
>
> (4) I apply the above using cmap=3Dmycolormap in contourf or pcolormesh.
>
> By the way, in case you don't know of the following site, it has an
> exhaustive selection of color palettes: http://cpt-city.org.uk/
>
> I've also attached an example of a basemap image using a cpt from the
> above site. Masked values not defined by the colormap are shown in
> gray.

=2D-=20
Lionel Roubeyrie - lro...@li...
LIMAIR
http://www.limair.asso.fr

Hi Experts,

i just stumbled upon something i can't quite explain (or rather i think i can
explain it, but i cannot understand it):

Consider the following minimal example:

    plot([1], [2],'o-'); show()

draws a single point at coords (1,2) - perfectly fine.

but when i try to plot the same at (0,2), i get an error:

    plot([0], [2],'o-'); show()

--------------------------------------------------------------------------------
/usr/lib/python2.4/site-packages/matplotlib/pylab.py in plot(*args, **kwargs)
   2016         hold(h)
   2017     try:
-> 2018         ret =  gca().plot(*args, **kwargs)
   2019         draw_if_interactive()
   2020     except:

/usr/lib/python2.4/site-packages/matplotlib/axes.py in plot(self, *args, **kwargs)
   2788             lines.append(line)
   2789         lines = [line for line in lines] # consume the generator
-> 2790         self.autoscale_view()
   2791         return lines
   2792

/usr/lib/python2.4/site-packages/matplotlib/axes.py in autoscale_view(self)
    813
    814         locator = self.xaxis.get_major_locator()
--> 815         self.set_xlim(locator.autoscale())
    816         locator = self.yaxis.get_major_locator()
    817         self.set_ylim(locator.autoscale())

/usr/lib/python2.4/site-packages/matplotlib/ticker.py in autoscale(self)
    796         dmin, dmax = self.dataInterval.get_bounds()
    797         dmin, dmax = self.nonsingular(dmin, dmax, expander = 0.05)
--> 798         return take(self.bin_boundaries(dmin, dmax), [0,-1])
    799
    800

/usr/lib/python2.4/site-packages/matplotlib/ticker.py in bin_boundaries(self, vmin, vmax)
    766     def bin_boundaries(self, vmin, vmax):
    767         nbins = self._nbins
--> 768         scale, offset = scale_range(vmin, vmax, nbins)
    769         vmin -= offset
    770         vmax -= offset

/usr/lib/python2.4/site-packages/matplotlib/ticker.py in scale_range(vmin, vmax, n, threshold)
    731     dv = abs(vmax - vmin)
    732     meanv = 0.5*(vmax+vmin)
--> 733     var = dv/max(abs(vmin), abs(vmax))
    734     if var < 1e-12:
    735         return 1.0, 0.0

ZeroDivisionError: float division
--------------------------------------------------------------------------------

I assume this is becaus matplotlib is unable to figure out the width to which it should scale the plot...


Any idea how to fix this?

thanks in advance,


Holger

Hi all

I have an application requiring that I do contour plots for masked data.
At the edge of my contour plots, I'm getting fragmented contours due to
missing points on the corner of a cell in my grid.

Is there any way that the contour drawing code can be set up so that it
plots contours even if a cell in the grid only has data on two sides?

I suspect that the rule for contour plots is that no contours are shown
if *any* of the point on the four corners are missing.

There is an example with the jagged edge shown in the attached image.

Cheers
JP

-- 
John Pye
School of Mechanical and Manufacturing Engineering
The University of New South Wales
Sydney  NSW 2052  Australia
t +61 2 9385 5127
f +61 2 9663 1222
mailto:john.pye_AT_student_DOT_unsw.edu.au
http://pye.dyndns.org/

Some of the y axis labels end up having different fonts:

http://s3.amazonaws.com/oylb/wlsn_qm.png?20060717232235

In the above image, 7 and 7.5 have a larger font. I'm using matplotlib 0.87.3 with ubuntu

Vineet

>>sy = (ax.bbox.ur().y() - ax.bbox.ll().y()) / (ax.viewLim.ur().y() - 
>>ax.viewLim.ll().y())
I solved the problem by using .get() on the BinOp's and then converting it back to a BinOP using Value()

Vineet