matplotlib-users Mailing List for matplotlib (Page 5)

On Mon, Feb 21, 2011 at 12:54 PM, Waléria Antunes David <
wal...@gm...> wrote:

> I did what you sai but the yscale was: 0, 0, 0, 0, 0, 0, 0, 1, 100 . Would
> be the correct way: 1, 10, 100, 1000.
>
> See the wrong image attached.  And the correct image - correct_graph
> attached.
>
> Thanks,
>
>
The graph isn't wrong.

You are plotting really small values.  If you don't want those values
plotted, then don't plot them, or set the lower limits.

Ben Root

On Mon, Feb 21, 2011 at 12:37 PM, Waléria Antunes David <
wal...@gm...> wrote:

> I don't know how to do this. Can you help me?
>
>
>
In your case, I would do something like this:

import matplotlib.ticker as mtick

# Other code here


# Now getting ready to plot
fig = plt.figure()
ax = fig.add_subplot(111)

ax.set_yscale('log')
ax.yaxis.set_major_formatter(mtick.FormatStrFormatter('%d'))

# Then plot as usual, be sure to set the y-limits.

Ben Root





> On Mon, Feb 21, 2011 at 3:32 PM, Benjamin Root <ben...@ou...> wrote:
>
>>
>>
>> On Mon, Feb 21, 2011 at 12:24 PM, Waléria Antunes David <
>> wal...@gm...> wrote:
>>
>>> I need the yscale so: 1 - 10 - 100 - 1000 . See my image attached. Can
>>> you help me,  please.
>>>
>>> Thanks,
>>>
>>>
>>> On Mon, Feb 21, 2011 at 1:51 PM, Waléria Antunes David <
>>> wal...@gm...> wrote:
>>>
>>>> So...now my xscale is correct but the yscale ...i need that in the
>>>> yscale should have a scale like this: 1 - 10 - 100 - 1000 ... without using
>>>> power rating . Can you help me??
>>>> See my image attached..
>>>>
>>>> Thanks,
>>>>
>>>>
>> Waleria,
>>
>> The data you are plotting contains points that have y-values as low as
>> 10^-17.  Plotting so that the y-*axis* has ticks of 1 - 10 - 100 - 1000
>> would result in those points to be missing from the plot.  Is that what you
>> want?
>>
>> To address the specific formatting of your tick labels, you can change the
>> formatter being used for that axis:
>>
>>
>> http://matplotlib.sourceforge.net/api/ticker_api.html?highlight=tick%20formatter
>>
>> By default, using ax.set_yscale('log') will automatically set the
>> LogFormatter object for formatting the y-axis.  If you don't want this
>> formatter, you can replace it with another formatter object.
>>
>> Ben Root
>>
>>
>

I don't know how to do this. Can you help me?


On Mon, Feb 21, 2011 at 3:32 PM, Benjamin Root <ben...@ou...> wrote:

>
>
> On Mon, Feb 21, 2011 at 12:24 PM, Waléria Antunes David <
> wal...@gm...> wrote:
>
>> I need the yscale so: 1 - 10 - 100 - 1000 . See my image attached. Can you
>> help me,  please.
>>
>> Thanks,
>>
>>
>> On Mon, Feb 21, 2011 at 1:51 PM, Waléria Antunes David <
>> wal...@gm...> wrote:
>>
>>> So...now my xscale is correct but the yscale ...i need that in the yscale
>>> should have a scale like this: 1 - 10 - 100 - 1000 ... without using power
>>> rating . Can you help me??
>>> See my image attached..
>>>
>>> Thanks,
>>>
>>>
> Waleria,
>
> The data you are plotting contains points that have y-values as low as
> 10^-17.  Plotting so that the y-*axis* has ticks of 1 - 10 - 100 - 1000
> would result in those points to be missing from the plot.  Is that what you
> want?
>
> To address the specific formatting of your tick labels, you can change the
> formatter being used for that axis:
>
>
> http://matplotlib.sourceforge.net/api/ticker_api.html?highlight=tick%20formatter
>
> By default, using ax.set_yscale('log') will automatically set the
> LogFormatter object for formatting the y-axis.  If you don't want this
> formatter, you can replace it with another formatter object.
>
> Ben Root
>
>

On Mon, Feb 21, 2011 at 12:24 PM, Waléria Antunes David <
wal...@gm...> wrote:

> I need the yscale so: 1 - 10 - 100 - 1000 . See my image attached. Can you
> help me,  please.
>
> Thanks,
>
>
> On Mon, Feb 21, 2011 at 1:51 PM, Waléria Antunes David <
> wal...@gm...> wrote:
>
>> So...now my xscale is correct but the yscale ...i need that in the yscale
>> should have a scale like this: 1 - 10 - 100 - 1000 ... without using power
>> rating . Can you help me??
>> See my image attached..
>>
>> Thanks,
>>
>>
Waleria,

The data you are plotting contains points that have y-values as low as
10^-17.  Plotting so that the y-*axis* has ticks of 1 - 10 - 100 - 1000
would result in those points to be missing from the plot.  Is that what you
want?

To address the specific formatting of your tick labels, you can change the
formatter being used for that axis:

http://matplotlib.sourceforge.net/api/ticker_api.html?highlight=tick%20formatter

By default, using ax.set_yscale('log') will automatically set the
LogFormatter object for formatting the y-axis.  If you don't want this
formatter, you can replace it with another formatter object.

Ben Root

On Mon, Feb 21, 2011 at 10:23 AM, Waléria Antunes David <
wal...@gm...> wrote:

> See my code: http://pastebin.com/xtgKteuW
>
> I need that tn the x-axis should have a scale like this: 0 - 2 - 4 - 6 - 8
> - 10 - 12 - 14 and in the y-axis a scale: 1 - 10 - 100 -1000 . See my
> images, the image correct_graph is correct.
>
> Thanks,
>
>
>
One problem at a time.  You are doing an errorbar plot and then creating a
subplot.  The subplot axes will not be the same thing as the axes you just
made the errorbar plot on.

See the correction here: http://pastebin.com/aZEJuGs2

Note, this does not address your tick label problem yet, just your plotting
problem.

Ben Root

Thanks Eric!

On Sun, Feb 20, 2011 at 9:02 PM, Eric Firing <ef...@ha...> wrote:
> On 02/20/2011 06:17 AM, Daniel Hyams wrote:
>>
>> The function "clabel" does not allow the use of formatters for the
>> label; this small patch enables their use, so that one can pass in any
>> subclass of ticker.Formatter in the 'fmt' argument.
>>
>> Replace the following in contour.py:
>>
>>     def get_text(self, lev, fmt):
>> "get the text of the label"
>>         if cbook.is_string_like(lev):
>>             return lev
>>         else:
>>             if isinstance(fmt,dict):
>>                 return fmt[lev]
>>             else:
>>                 return fmt%lev
>>
>> With the following:
>>
>>     def get_text(self, lev, fmt):
>> "get the text of the label"
>>         if cbook.is_string_like(lev):
>>             return lev
>>         else:
>>             if isinstance(fmt,dict):
>>                 return fmt[lev]
>>           elif isinstance(fmt,ticker.Formatter):
>>                 return fmt(lev)
>>             else:
>>                 return fmt%lev
>>
>>
>> As you can see, there is only a two line change, but it does enable
>> formatters.  Perhaps one of the devs can put this in?
>
> Done.
>
> https://github.com/matplotlib/matplotlib/commit/9114598093984f39f40e50db9a8c15cfe1dc0d74
>
> Eric
>>
>>
>> --
>> Daniel Hyams
>> dh...@gm... <mailto:dh...@gm...>
>
> ------------------------------------------------------------------------------
> The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
> Pinpoint memory and threading errors before they happen.
> Find and fix more than 250 security defects in the development cycle.
> Locate bottlenecks in serial and parallel code that limit performance.
> http://p.sf.net/sfu/intel-dev2devfeb
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>



-- 
Daniel Hyams
dh...@gm...

On Mon, Feb 21, 2011 at 10:07 AM, Waléria Antunes David <
wal...@gm...> wrote:

> Hello all,
>
> I have this graph but the scale is wrong. See my images. In the image
> my_graph the xscale an yscale is wrong. I need to let the scale of x and y
> my_graph as correct_graph. Can you help me??
>
> Thanks
>

Have you tried ax.set_yscale('log')?  I don't see anything wrong with the
scale of the xaxis. While the range of values are different, that has
nothing to do with the "scale" but the limits and the axis tickers.

Ben Root

Fri, 18 Feb 2011 20:24:31 +0100, Nils Wagner wrote:
> what is the reason for the white areas in the corners of the
> interpolation domain?
> Any idea ?

Griddata does not do any extrapolation, and the corners are outside the 
convex hull of the point set.

> import numpy as np
> from scipy.interpolate import griddata

BTW, if you're using Scipy's griddata, the Scipy lists might be a better 
place to ask :)

I am using a workaround now. But that is a hackery solution.

Before plotting my data I convert it to dBs and limit it to the lowest value
I want to display. Then I plot it using a regular polar plot with a custom
formatting function that sets the tick labels with respect to the data
offset.
Since I use a custom Navigation-Toolbar anyways it was no big deal to add
the few necessery lines of codes to handle the offset there, too.

The plot now looks exactly as I want it. But: I'd still prefer using a scale
that does all the work in the background.


-- 
View this message in context: http://old.nabble.com/Polar-plot---problem-with-negative-values-for-radius-tp30936638p30975519.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

On 02/20/2011 06:17 AM, Daniel Hyams wrote:
>
> The function "clabel" does not allow the use of formatters for the
> label; this small patch enables their use, so that one can pass in any
> subclass of ticker.Formatter in the 'fmt' argument.
>
> Replace the following in contour.py:
>
>     def get_text(self, lev, fmt):
> "get the text of the label"
>         if cbook.is_string_like(lev):
>             return lev
>         else:
>             if isinstance(fmt,dict):
>                 return fmt[lev]
>             else:
>                 return fmt%lev
>
> With the following:
>
>     def get_text(self, lev, fmt):
> "get the text of the label"
>         if cbook.is_string_like(lev):
>             return lev
>         else:
>             if isinstance(fmt,dict):
>                 return fmt[lev]
>           elif isinstance(fmt,ticker.Formatter):
>                 return fmt(lev)
>             else:
>                 return fmt%lev
>
>
> As you can see, there is only a two line change, but it does enable
> formatters.  Perhaps one of the devs can put this in?

Done.

https://github.com/matplotlib/matplotlib/commit/9114598093984f39f40e50db9a8c15cfe1dc0d74

Eric
>
>
> --
> Daniel Hyams
> dh...@gm... <mailto:dh...@gm...>

On 02/20/2011 06:17 AM, Daniel Hyams wrote:
>
> The function "clabel" does not allow the use of formatters for the
> label; this small patch enables their use, so that one can pass in any
> subclass of ticker.Formatter in the 'fmt' argument.
>
> Replace the following in contour.py:
>
>     def get_text(self, lev, fmt):
> "get the text of the label"
>         if cbook.is_string_like(lev):
>             return lev
>         else:
>             if isinstance(fmt,dict):
>                 return fmt[lev]
>             else:
>                 return fmt%lev
>
> With the following:
>
>     def get_text(self, lev, fmt):
> "get the text of the label"
>         if cbook.is_string_like(lev):
>             return lev
>         else:
>             if isinstance(fmt,dict):
>                 return fmt[lev]
>           elif isinstance(fmt,ticker.Formatter):
>                 return fmt(lev)
>             else:
>                 return fmt%lev
>
>
> As you can see, there is only a two line change, but it does enable
> formatters.  Perhaps one of the devs can put this in?

Yes, I have it done at home, and when I am a little more confident I 
know how to interact with github, I will push it up.

Eric

>
>
> --
> Daniel Hyams
> dh...@gm... <mailto:dh...@gm...>

On 2011-02-18 07:32:48 +0100, Robert Abiad said:

> Dear Folks,
> 
> I'm finding that hist has problems computing on 2d arrays.
> 
>     import  numpy
>     import  pylab
>     mu,  sigma  =  2,  0.5
>     v  =  numpy.random.normal(mu,sigma,160000)
>     pylab.hist(v,  bins=1000,  normed=1)
> 
> This works without any problems.  But if you try this:
> 
>     w=v.reshape(400,400)
>     pylab.hist(w,  bins=1000,  normed=1)
> 
> it doesn't come back on my machine until all of memory is used up.  However:
> 
>     n,bins = numpy.histogram(w,bins=1000,normed=1)
> 
> works just fine.

That's by design. For a n x m array, pylab.hist is doing m times a 
histogram of the n items subarray and then tries to plot them somehow 
on top of each other. The detail I don't know, but most likely you want 
to do
pylab.hist(w.flatten(), ....)
to get a 1-dim array that you want to fill in ONE histogram. (for 
example for image arrays).

HTH,
Michael

> 
> 
> 
> ------------------------------------------------------------------------------
> The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
> Pinpoint memory and threading errors before they happen.
> Find and fix more than 250 security defects in the development cycle.
> Locate bottlenecks in serial and parallel code that limit performance.
> http://p.sf.net/sfu/intel-dev2devfeb

The function "clabel" does not allow the use of formatters for the label;
this small patch enables their use, so that one can pass in any subclass of
ticker.Formatter in the 'fmt' argument.

Replace the following in contour.py:

   def get_text(self, lev, fmt):
       "get the text of the label"
       if cbook.is_string_like(lev):
           return lev
       else:
           if isinstance(fmt,dict):
               return fmt[lev]
           else:
               return fmt%lev

With the following:

   def get_text(self, lev, fmt):
       "get the text of the label"
       if cbook.is_string_like(lev):
           return lev
       else:
           if isinstance(fmt,dict):
               return fmt[lev]
           elif isinstance(fmt,ticker.Formatter):
               return fmt(lev)
           else:
               return fmt%lev


As you can see, there is only a two line change, but it does enable
formatters.  Perhaps one of the devs can put this in?


-- 
Daniel Hyams
dh...@gm...

Jeremy Brown-9 wrote:
> 
> I ran into a nasty problem and was stuck until I realized that my numpy
> version was 1.3 when it should of been 1.5. But the errors never gave me
> an indication that this was the issue. 
> 
> ... 
> It is working correctly just hope to save someone the headache. 
> 

No, it was too late to save me from a headache :-(

But this indeed fixed my segmentation fault too, so thanks lot!

(Running OpenSuse 11.2; Python segfaulted on a mere "import
matplotlib.pyplot as plt" statement)

Can someone please update the system requirements / dependencies for
matplotlib, or fix this in the source code?

Regards,

Diederik
-- 
View this message in context: http://old.nabble.com/matplotlib-gcc-segmentation-fault---exit-status-1-on-SLES-and-SUSE-linux-tp30516102p30971237.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Hello,

I am trying to change the default font by adding the following lines
to the matplotlibrc file:

## FONT
font.family : serif
font.serif : Palatino

However, I get the following error.

/Library/Frameworks/Python.framework/Versions/6.3/lib/python2.6/site-packages/matplotlib/font_manager.py:1242:
UserWarning: findfont: Font family ['serif'] not found. Falling back
to Bitstream Vera Sans
  (prop.get_family(), self.defaultFamily[fontext]))

If I just use:

font.family : serif

I get a serif font (that is not Palatino). I certainly have Palatino
on my machine (Mac OS X 10.5.8).
Am I doing something wrong? If I want all text to be some font, is
there a way to change it in matplotlibrc file?

Thanks for your help.

On 2/20/2011 12:25 AM, Richard D. Moores wrote:
> Windows Vista
> Python 2.6
>
> I downloaded matplotlib-1.0.1.win-amd64-py2.6.exe from
> http://sourceforge.net/projects/matplotlib/files/matplotlib/matplotlib-1.0.1/,
> but I get "Python version 2.6 required, which was not found in the
> registry".
>
> I do have Python 2.6, but its folder is C:\Python26. What to do?
>
> Thanks,
>
> Dick Moores
>

You likely have the 32 bit version of Python 2.6 installed. You need to 
download and install matplotlib-1.0.1.win32-py2.6.exe.

Christoph

Windows Vista
Python 2.6

I downloaded matplotlib-1.0.1.win-amd64-py2.6.exe from
http://sourceforge.net/projects/matplotlib/files/matplotlib/matplotlib-1.0.1/,
but I get "Python version 2.6 required, which was not found in the
registry".

I do have Python 2.6, but its folder is C:\Python26. What to do?

Thanks,

Dick Moores

Hi,

If I want to specify the color_cycle for plot lines, I can use the
following line in a matplotlibrc file:

axes.color_cycle : b, g, r, c

Is there a similar specification that can be used for cycling through
markers? Something like:

lines.marker_cycle : s, o, x, +

Thank you.

Thomas and Goyo,

Thanks for the answers. I am fine with them being coverted to ints. In
fact, my ax.set_ylim(bottom=0, top=6000) contains ints and I don't
understand why they are converted to floats. I was thinking that there
is perhaps another (more correct) way to set tick values. I will
either use Thomas' approach or Goyo's.

Thanks again.

On Sat, Feb 19, 2011 at 4:52 PM, Goyo <goy...@gm...> wrote:
> 2011/2/19 Thomas Lecocq <thl...@ms...>:
>
>> To avoid having the warning, you can explicitly provide ints :
>>
>>  ax.set_yticks(range(int(ax.get_ylim()[0]), int(ax.get_ylim()[1]), 400))
>>
>> I don't know any easier method of setting the ticks... Let's wait for the
>> user community input !
>
> If you don't want your floats converted to ints then use numpy arrange
> or linespace.
>
> Goyo
>

On Sat, Feb 19, 2011 at 4:21 PM, K.-Michael Aye <kmi...@gm...>wrote:

> Dear all,
>
> I wondered if there's any principle reason one should not try to use
> Basemap with other planetary data?
> I am assuming, as long as projections are used that follow certain
> standards there should not be any problems, right?
> Of course any of the methods like drawcountries() and so on don't make
> sense, but I guess it would be cool to have these or similar things (no
> countries on Mars yet) reimplemented for some useful things for Mars?
> Something like drawcanyons()
>
> Please tell me if one could see an upcoming problem with this idea!
>
> Best regards,
> Michael
>
>
>
Others can probably comment further on this, but at the very least, I know
that a user can utilize basemap for plotting other pieces of geographical
data by supplying their own shapefiles.  I do this all the time for my own
plots where I want to include local counties and municipal roads.  If there
are shapefiles for Martian geographical features, it is possible to use
basemap for that.

Ben Root

Dear all,

I wondered if there's any principle reason one should not try to use 
Basemap with other planetary data?
I am assuming, as long as projections are used that follow certain 
standards there should not be any problems, right?
Of course any of the methods like drawcountries() and so on don't make 
sense, but I guess it would be cool to have these or similar things (no 
countries on Mars yet) reimplemented for some useful things for Mars? 
Something like drawcanyons()

Please tell me if one could see an upcoming problem with this idea!

Best regards,
Michael 

2011/2/19 Thomas Lecocq <thl...@ms...>:

> To avoid having the warning, you can explicitly provide ints :
>
>  ax.set_yticks(range(int(ax.get_ylim()[0]), int(ax.get_ylim()[1]), 400))
>
> I don't know any easier method of setting the ticks... Let's wait for the
> user community input !

If you don't want your floats converted to ints then use numpy arrange
or linespace.

Goyo

Good evening,

> 
> Hello,
> 
> I was trying to change the tick values that show up on the y-axis. I
> wanted a more finer range than the default. So I added:
> 
>         ax.set_yticks(range(ax.get_ylim()[0], ax.get_ylim()[1], 400))
> 
> to my plot. And before this line, I have
> 
>         ax.set_ylim(bottom=0, top=6000)
> 
> I get the results I want, but I get the following warning:
> 
> :91: DeprecationWarning: integer argument expected, got float
>   ax.set_yticks(range(ax.get_ylim()[0], ax.get_ylim()[1], 400))
> 

To avoid having the warning, you can explicitly provide ints :

 ax.set_yticks(range(int(ax.get_ylim()[0]), int(ax.get_ylim()[1]), 400))

I don't know any easier method of setting the ticks... Let's wait for the user community input !

Cheers,

Thomas
 		 	   		  

Hello,

I was trying to change the tick values that show up on the y-axis. I
wanted a more finer range than the default. So I added:

        ax.set_yticks(range(ax.get_ylim()[0], ax.get_ylim()[1], 400))

to my plot. And before this line, I have

        ax.set_ylim(bottom=0, top=6000)

I get the results I want, but I get the following warning:

:91: DeprecationWarning: integer argument expected, got float
  ax.set_yticks(range(ax.get_ylim()[0], ax.get_ylim()[1], 400))


Is there a better way of doing this, other than explicitly typing the
lower and upper limits in ax.set_yticks() function.

Thank you.

On Fri, Feb 18, 2011 at 3:57 PM, Jose Guzman <sjm...@go...>wrote:

>  Hi everybody
>
> is there anybody to set the distance between the x/y axis numbers and the
> corresponding axis labels. I found the default distance is rather small, and
> I would like to enlarge it. I was playing around in .matplolibrc but I did
> not find anything useful.
>
> Any suggestion??
>
> Thanks in advance
>
> Jose
>
>
Jose,

When setting the axes label, you can pass a keyword argument to labelpad
which is the number of points that should be between the label and the
axis.  While this isn't the distance between the numbers and the label, this
should help.  There should be a corresponding rcParam as well.

http://matplotlib.sourceforge.net/api/axes_api.html#matplotlib.axes.Axes.set_xlabel

I hope this helps!
Ben Root