matplotlib-users Mailing List for matplotlib

Mike Bauer wrote:
> Eric,
> 
> Thanks for the reply.  I'm trying to show the relative 2d distribuion 
> between 2 sets of data. I thought the normalization would ease the 
> comparison. Fixing the ' doesn't help.
> 
> So are you saying I need an instance of something.normalize rather than 
> just passing norm='normalize'?

It sounds like you are misunderstanding the norm kwarg; it is for 
controlling the mapping of an arbitrary range of numbers to the 0-1 
range that is used in color mapping.  The default is a linear mapping; 
one can use a log mapping instead ("norm=colors.LogNorm()"), or make 
your own mapping function, etc.  The norm kwarg takes an instance of a 
Normalize class or subclass.  See colors.py to find out what Normalize 
subclasses are available.  But, you may not need to specify one at all, 
depending on what it is you are trying to do.

I still don't understand what it is that you wanted to "normalize". 
What was the undesirable characteristic of the plot you had before you 
put in the norm kwarg?

Eric

> 
> Mike
> 
> Sent from my iPhone
> 
> On Mar 20, 2009, at 5:15 PM, Eric Firing <ef...@ha...> wrote:
> 
>> Mike Bauer wrote:
>>> Hello,
>>> Quick note. I'm making plots with hexbin and everything works  
>>> correctly until I try to use the norm='Normalize' option at which  
>>> point I get:
>>> Traceback (most recent call last):
>>>   File "diff_engine_v2tmp.py", line 731, in <module>
>>>     kept_and_discards)
>>>   File "diff_engine_v2tmp.py", line 605, in main
>>>     plt.hexbin(xdat,ydat,cmap=cm.jet,gridsize=25,norm=Normalize' )
>>
>> What is that single quote mark doing after Normalize?  If we ignore 
>> it, then it looks like you are passing a class, not a class instance 
>> as the kwarg needs.
>>
>>>   File "/opt/local/Library/Frameworks/Python.framework/Versions/2.5/ 
>>> lib/python2.5/site-packages/matplotlib/pyplot.py", line 1920, in hexbin
>>>     ret =  gca().hexbin(*args, **kwargs)
>>>   File "/opt/local/Library/Frameworks/Python.framework/Versions/2.5/ 
>>> lib/python2.5/site-packages/matplotlib/axes.py", line 5452, in hexbin
>>>     collection.autoscale_None()
>>>   File "/opt/local/Library/Frameworks/Python.framework/Versions/2.5/ 
>>> lib/python2.5/site-packages/matplotlib/cm.py", line 148, in  
>>> autoscale_None
>>>     self.norm.autoscale_None(self._A)
>>> AttributeError: 'int' object has no attribute 'autoscale_None'
>>
>> This part of the traceback is also a little puzzling; I'm not sure why 
>> self.norm is an int at this point.
>>
>>> I assume this a bug of some sort.
>>
>> No, I think the problem is that you are passing a class instead of an 
>> instance of a class as the norm kwarg to hexbin.  (It is not 
>> completely clear to me from the traceback, however--there is that 
>> strange single quote mark.) What kind of normalization are you trying 
>> to to?  In other words, what are you trying to accomplish by 
>> specifying the norm kwarg?
>>
>> Eric
>>
>>
>>> Thanks for any ideas.
>>> Mike
>>> Using:
>>>    os-x 10.5.6
>>>    python 2.5.4 from macports
>>>    matplotlib  0.98.5.2 from macports
>>> ------------------------------------------------------------------------------ 
>>>
>>> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
>>> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
>>> easily build your RIAs with Flex Builder, the Eclipse(TM)based 
>>> development
>>> software that enables intelligent coding and step-through debugging.
>>> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
>>> _______________________________________________
>>> Matplotlib-users mailing list
>>> Mat...@li...
>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>

Mike Bauer wrote:
> Hello,
> 
> Quick note. I'm making plots with hexbin and everything works  
> correctly until I try to use the norm='Normalize' option at which  
> point I get:
> 
> Traceback (most recent call last):
>    File "diff_engine_v2tmp.py", line 731, in <module>
>      kept_and_discards)
>    File "diff_engine_v2tmp.py", line 605, in main
>      plt.hexbin(xdat,ydat,cmap=cm.jet,gridsize=25,norm=Normalize' )

What is that single quote mark doing after Normalize?  If we ignore it, 
then it looks like you are passing a class, not a class instance as the 
kwarg needs.

>    File "/opt/local/Library/Frameworks/Python.framework/Versions/2.5/ 
> lib/python2.5/site-packages/matplotlib/pyplot.py", line 1920, in hexbin
>      ret =  gca().hexbin(*args, **kwargs)
>    File "/opt/local/Library/Frameworks/Python.framework/Versions/2.5/ 
> lib/python2.5/site-packages/matplotlib/axes.py", line 5452, in hexbin
>      collection.autoscale_None()
>    File "/opt/local/Library/Frameworks/Python.framework/Versions/2.5/ 
> lib/python2.5/site-packages/matplotlib/cm.py", line 148, in  
> autoscale_None
>      self.norm.autoscale_None(self._A)
> AttributeError: 'int' object has no attribute 'autoscale_None'

This part of the traceback is also a little puzzling; I'm not sure why 
self.norm is an int at this point.

> 
> I assume this a bug of some sort.

No, I think the problem is that you are passing a class instead of an 
instance of a class as the norm kwarg to hexbin.  (It is not completely 
clear to me from the traceback, however--there is that strange single 
quote mark.) What kind of normalization are you trying to to?  In other 
words, what are you trying to accomplish by specifying the norm kwarg?

Eric


> 
> Thanks for any ideas.
> 
> Mike
> 
> Using:
> 	os-x 10.5.6
> 	python 2.5.4 from macports
> 	matplotlib  0.98.5.2 from macports
> 
> 
> 
> ------------------------------------------------------------------------------
> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
> software that enables intelligent coding and step-through debugging.
> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Hello,

Quick note. I'm making plots with hexbin and everything works  
correctly until I try to use the norm='Normalize' option at which  
point I get:

Traceback (most recent call last):
   File "diff_engine_v2tmp.py", line 731, in <module>
     kept_and_discards)
   File "diff_engine_v2tmp.py", line 605, in main
     plt.hexbin(xdat,ydat,cmap=cm.jet,gridsize=25,norm=Normalize' )
   File "/opt/local/Library/Frameworks/Python.framework/Versions/2.5/ 
lib/python2.5/site-packages/matplotlib/pyplot.py", line 1920, in hexbin
     ret =  gca().hexbin(*args, **kwargs)
   File "/opt/local/Library/Frameworks/Python.framework/Versions/2.5/ 
lib/python2.5/site-packages/matplotlib/axes.py", line 5452, in hexbin
     collection.autoscale_None()
   File "/opt/local/Library/Frameworks/Python.framework/Versions/2.5/ 
lib/python2.5/site-packages/matplotlib/cm.py", line 148, in  
autoscale_None
     self.norm.autoscale_None(self._A)
AttributeError: 'int' object has no attribute 'autoscale_None'

I assume this a bug of some sort.

Thanks for any ideas.

Mike

Using:
	os-x 10.5.6
	python 2.5.4 from macports
	matplotlib  0.98.5.2 from macports

Now it works like charm. Thank you for quick answer.

Stano.

On Fri, Mar 20, 2009 at 7:10 PM, Eric Firing <ef...@ha...> wrote:
> Stano Paška wrote:
>>
>> Hi,
>> I am using matplotlib first time and I am wonder how to free resources
>>
>> I have this script
>>
>> # coding: utf-8
>>
>> import matplotlib
>> matplotlib.use('Agg')
>>
>> import pylab
>>
>> class Graph:
>>
>>    def __init__(self):
>>        # sirka stlpcov
>>        self.width = 0.5
>>        self.font = {'fontname':'Tahoma', 'fontsize':10}
>>
>>    def setTitle(self, title):
>>        """Nastavi nadpis grafu
>>        """
>>        self.title = title
>>
>>    def setXLabel(self, label):
>>        self.x_label = label
>>
>>    def setYLabel(self, label):
>>        self.y_label = label
>>
>>    def setYValues(self, values):
>>        """Nastavi y-ove hodnoty
>>        """
>>        self.y_values = values
>>        # vypocitame lave spodne rohy stlpcov
>>        self.x_values = [x + self.width / 2 for x in
>> range(len(self.y_values))]
>>        # vypocitame stred stlpcov (znacky na x-ovej osi)
>>        self.x_ticks = [x + self.width for x in range(len(self.y_values))]
>>
>>    def setXTickLabels(self, labels):
>>        """Nastavi popisky pre ciarky na x-ovej osi
>>        """
>>        self.x_ticksLabels = labels
>>
>>    def makeGraph(self):
>>        self.fig = pylab.figure()
>>        self.fig.set_dpi(72)
>>        self.fig.set_figheight(3)
>>        self.fig.set_figwidth(5)
>>
>>        self.fig.subplots_adjust(bottom=0.15)
>>
>>        self.ax = self.fig.add_subplot(111)
>>
>>        self.ax.bar(left=self.x_values, height=self.y_values,
>> width=self.width, color='#f00000')
>>        self.ax.set_xticks(self.x_ticks)
>>        self.ax.set_xticklabels(self.x_ticksLabels)
>>
>>        self.ax.set_xlabel(self.x_label, **self.font)
>>        self.ax.set_ylabel(self.y_label, **self.font)
>>        self.ax.set_title(self.title, **self.font)
>>
>>    def saveGraph(self, path):
>>        self.fig.savefig(path, format='png')
>>
>>    def __del__(self):
>>        self.fig.delaxes(self.ax)
>>        del self.fig
>>        del self.ax
>>
>> if __name__ == '__main__':
>>    import random
>>    import gc
>>    for i in range(5000):
>>        g = Graph()
>>        g.setTitle(u'Spotreba za rok 2008 podľa ATC skupín')
>>        g.setXLabel(u'ATC skupiny')
>>        g.setYLabel(u'Spotreba v EUR')
>>        g.setYValues([random.randint(0, x) for x in range(20)])
>>        g.setXTickLabels([chr(random.randint(65, 90)) for x in range(20)])
>>        g.makeGraph()
>>        g.saveGraph('grafy/' + str(i) + '.png')
>>        del g
>>        print gc.get_count()
>>        gc.collect()
>>        print gc.get_count()
>>
>> When I used TkAgg, it crashes after 190 images (Fail to create pixmap
>> with Tk_GetPixmap in ImgPhotoInstanceSetSize. tried to delete photo
>> image when instances still exist).
>> When I switch to WX, it crashes after 400 images.
>> And Agg crashes when pagefile reaches 1.5x ram (cca 2000 images).
>>
>> It is possible somehow free resources in matplotlib?
>
> You need to close the figure.  Try putting
> pylab.close(self.fig) at the start of your __del__ method.  In fact, I doubt
> you need to explicitly del anything in that method--you just need to close
> the figure so that pylab will release its references related to that figure.
>
> There may be more things you need to change, but closing the figure is
> certainly going to be a big one.
>
> Eric
>
>>
>> I use python 2.5.4, wx 2.8.9.2, numpy 1.3.0b1 on win XP
>>
>> Thanks for answers.
>>
>> Stano.
>>
>>
>> ------------------------------------------------------------------------------
>> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
>> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
>> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
>> software that enables intelligent coding and step-through debugging.
>> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>

Thanks for sharing this -- I'm curious about how you've dealt with some 
of these issues and see if any of them can be brought into the core.  
Overlapping text has long been something I've wanted to address, but 
it's difficult to solve and maintain as much flexibility as we currently 
have.

Running your script, I get this traceback:

Traceback (most recent call last):
  File "AutoscaledText.py", line 236, in <module>
    1024, 768, 'UseOffset_ON', 'ScientificNotation_ON')
  File "AutoscaledText.py", line 171, in HistogramPlot
    fig, ax = CommonPlottingCode(in_WidthInPixels, in_HeightInPixels, 
in_DataName, 'Frequency', useOffsetIfNeeded, scientificNotation, False, 
0.0, 0.0, 1.0, 1.0)
  File "AutoscaledText.py", line 75, in CommonPlottingCode
    x_label._fontproperties._size = x_label._fontproperties._size * 
heightRatioForTextSize
TypeError: can't multiply sequence by non-int of type 'float'
 > /wonkabar/data1/scraps/AutoscaledText.py(75)CommonPlottingCode()
-> x_label._fontproperties._size = x_label._fontproperties._size * 
heightRatioForTextSize


_fontproperties._size can be a CSS size name, such as "medium" or 
"large", so this line of code won't work.  I replaced this with:

    x_label.set_size(x_label.get_size() * heightRatioForTextSize)
    y_label.set_size(y_label.get_size() * widthRatioForTextSize)

which also has the advantage of avoiding private APIs that may change in 
the future.  This seems to work for me, but I don't know if it matches 
your results.

Mike


James Phillips wrote:
> All,
>
>      Attached, and below, is public domain code for making 
> variable-sized plots with autoscaled text that exactly fits the 
> available visual plot space, useful for web sites where users choose 
> output files with different sizes.  Examples are at the bottom of the 
> file.
>
>           James R. Phillips
>           2548 Vera Cruz Drive
>           Birmingham, AL 35235 USA
>           email: zu...@zu... <mailto:zu...@zu...>
>           http://zunzun.com
>
>
>
> #    Entered into the public domain 20 March 2009
> #    James R. Phillips
> #    2548 Vera Cruz Drive
> #    Birmingham, AL 35235 USA
> #    email: zu...@zu... <mailto:zu...@zu...>
> #    http://zunzun.com
>
> import numpy as np
> import math, matplotlib
> matplotlib.use('Agg') # must be used prior to the next two statements
> import matplotlib.pyplot as plt
> import matplotlib.mlab as mlab
>
>
> def DetermineOnOrOffFromString(in_String):
>     tempString = in_String.split('_')[-1:][0].upper() # allows any 
> amount of prefacing text
>     if tempString == 'ON':
>         return True
>     return False
>
>
> def DetermineScientificNotationFromString(inData, in_String):
>     tempString = in_String.split('_')[-1:][0].upper() # allows any 
> amount of prefacing text
>     if tempString == 'ON':
>         return True
>     elif tempString == 'OFF':
>         return False
>     else: # must be AUTO
>         minVal = np.abs(np.min(inData))
>         maxVal = np.abs(np.max(inData))
>         deltaVal = np.abs(maxVal - minVal)
>        
>         scientificNotation = False
>         if (maxVal > 100.0) or (minVal < -100.0) or (deltaVal < .05):
>             scientificNotation = True
>     return scientificNotation
>
>
> def CommonPlottingCode(in_WidthInPixels, in_HeightInPixels, in_XName, 
> in_YName, in_UseOffsetIfNeeded, in_X_UseScientificNotationIfNeeded, 
> in_Y_UseScientificNotationIfNeeded, in_Left, in_Bottom, in_Right, 
> in_Top): # default to lots of room around graph
>
>     # a litle more room between x axis and tick mark labels, so not 
> text overlap at the bottom left corner - set this before other calls
>     matplotlib.rcParams['xtick.major.pad'] = 5+ 
> (float(in_HeightInPixels) / 100.0) # minimum + some scaled
>
>     fig = plt.figure(figsize=(float(in_WidthInPixels ) / 100.0, 
> float(in_HeightInPixels ) / 100.0), dpi=100)
>     fig.subplotpars.update(in_Left, in_Bottom, in_Right, in_Top)
>     ax = fig.add_subplot(111, frameon=True)
>
>     # white background, almost no border space
>     fig.set_facecolor('w')
>
>     xFormatter = fig.gca().xaxis.get_major_formatter()
>     xFormatter._useOffset = in_UseOffsetIfNeeded
>     xFormatter.set_scientific(in_X_UseScientificNotationIfNeeded)
>     fig.gca().xaxis.set_major_formatter(xFormatter)
>
>     yFormatter = fig.gca().yaxis.get_major_formatter()
>     yFormatter._useOffset = in_UseOffsetIfNeeded
>     yFormatter.set_scientific(in_Y_UseScientificNotationIfNeeded)
>     fig.gca().yaxis.set_major_formatter(yFormatter)
>
>     # Scale text to imagesize.  Text sizes originally determined at 
> image size of 500 x 400
>     widthRatioForTextSize = float(in_WidthInPixels) / 500.0
>     heightRatioForTextSize = float(in_HeightInPixels) / 400.0
>     for xlabel_i in ax.get_xticklabels():
>         xlabel_i.set_fontsize(xlabel_i.get_fontsize() * 
> heightRatioForTextSize)
>     xOffsetText = fig.gca().xaxis.get_offset_text()
>     xOffsetText.set_fontsize(xOffsetText.get_fontsize() * 
> heightRatioForTextSize * 0.9)
>     for ylabel_i in ax.get_yticklabels():
>         ylabel_i.set_fontsize(ylabel_i.get_fontsize() * 
> widthRatioForTextSize)
>     yOffsetText = fig.gca().yaxis.get_offset_text()
>     yOffsetText.set_fontsize(yOffsetText.get_fontsize() * 
> heightRatioForTextSize * 0.9)
>    
>     x_label = ax.set_xlabel(in_XName)
>     y_label = ax.set_ylabel(in_YName)
>     x_label._fontproperties._size = x_label._fontproperties._size * 
> heightRatioForTextSize
>     y_label._fontproperties._size = y_label._fontproperties._size * 
> widthRatioForTextSize
>
>     plt.grid(True) # call this just before returning
>
>     return fig, ax
>
>
> def YieldNewExtentsAndNumberOfMajor_X_TickMarks(fig, ax, 
> in_WidthInPixels, in_HeightInPixels, in_OffsetUsed):
>     # draw everything so items can be measured for size
>     canvas = plt.get_current_fig_manager().canvas
>     canvas.draw()
>    
>     # some preliminary info
>     xLabelPoints = 
> ax.set_xlabel(ax.get_xlabel()).get_window_extent().get_points() # [ 
> [x,y], [x,y] ]
>     yLabelPoints = 
> ax.set_ylabel(ax.get_ylabel()).get_window_extent().get_points() # [ 
> [x,y], [x,y] ], rotated 90 degrees
>     xTickZeroPoints = 
> ax.get_xticklabels()[0].get_window_extent().get_points()
>     yTickZeroPoints = 
> ax.get_yticklabels()[0].get_window_extent().get_points()
>     xTickIndexPoints = 
> ax.get_xticklabels()[len(ax.get_xticklabels())-1].get_window_extent().get_points()
>     yTickIndexPoints = 
> ax.get_yticklabels()[len(ax.get_yticklabels())-1].get_window_extent().get_points()
>     currentPoints = ax.bbox.get_points()
>     maxLeft = currentPoints[0][0]
>     maxBottom = currentPoints[0][1]
>     maxRight = currentPoints[1][0]
>     maxTop = currentPoints[1][1]
>    
>     # find the most left-ward location
>     if xTickZeroPoints[0][0] < maxLeft:
>         maxLeft = xTickZeroPoints[0][0]
>     if yTickZeroPoints[0][0] < maxLeft:
>         maxLeft = yTickZeroPoints[0][0]
>     if yTickIndexPoints[0][0] < maxLeft:
>         maxLeft = yTickIndexPoints[0][0]
>     if xLabelPoints[0][0] < maxLeft:
>         maxLeft = xLabelPoints[0][0]
>     if yLabelPoints[0][0] < maxLeft: # 90 degrees
>         maxLeft = yLabelPoints[0][0]
>
>     # find the most right-ward location
>     if xTickIndexPoints[1][0] > maxRight:
>         maxRight = xTickIndexPoints[1][0]
>     if xLabelPoints[1][0] > maxRight:
>         maxRight = xLabelPoints[1][0]
>
>     # find the most bottom-ward location
>     if xTickZeroPoints[0][1] < maxBottom:
>         maxBottom = xTickZeroPoints[0][1]
>     if xLabelPoints[0][1] < maxBottom:
>         maxBottom = xLabelPoints[0][1]
>     if yLabelPoints[0][1] < maxBottom:
>         maxBottom = yLabelPoints[0][1]
>
>     # find the most top-ward location
>     if yTickIndexPoints[1][1] > maxTop:
>         maxTop = yTickIndexPoints[1][1]
>     if True == in_OffsetUsed: # could not find a better way to get this
>         yp = ax.get_yticklabels()[0].get_window_extent().get_points()
>         maxTop += yp[1][1] - yp[0][1]
>
>     newLeft = ax.bbox._bbox.get_points()[0][0] - (float(maxLeft) / 
> float(in_WidthInPixels)) + 0.01
>     newBottom = ax.bbox._bbox.get_points()[0][1] - (float(maxBottom) / 
> float(in_HeightInPixels)) + 0.01
>     newRight = ax.bbox._bbox.get_points()[1][0] + (1.0 - 
> (float(maxRight) / float(in_WidthInPixels))) - 0.01
>     newTop = ax.bbox._bbox.get_points()[1][1] + (1.0 - (float(maxTop) 
> / float(in_HeightInPixels))) - 0.01
>
>     # now redraw and check number of X tick marks
>     canvas.draw()
>
>     # Calculate major number of X tick marks based on label size
>     totalWidth = 0.0
>     maxWidth = 0.0
>     numberOfMajor_X_TickMarks = len(ax.get_xticklabels())
>     for xlabel_i in ax.get_xticklabels():
>         w = xlabel_i.get_window_extent().get_points() # the drawn text 
> bounding box corners as numpy array of [x,y], [x,y]
>         width = w[1][0] - w[0][0]
>         totalWidth += width
>         if width > maxWidth:
>             maxWidth = width
>     if totalWidth > (0.95 * ((newRight - newLeft) * 
> float(in_WidthInPixels))): # 0.95 for some spacing between tick labels
>         numberOfMajor_X_TickMarks = int(math.floor((0.95 * ((newRight 
> - newLeft) * float(in_WidthInPixels))) / maxWidth))
>
>     return (newLeft, newBottom, newRight, newTop, 
> numberOfMajor_X_TickMarks,)
>
>
> def HistogramPlot(in_DataToPlot, in_FileNameAndPath, in_DataName, 
> in_FillColor, in_WidthInPixels, in_HeightInPixels, 
> in_UseOffsetIfNeeded, in_UseScientificNotationIfNeeded):
>
>     # decode ends of strings ('XYZ_ON', 'XYZ_OFF', 'XYZ_AUTO', etc.) 
> to boolean values
>     scientificNotation = 
> DetermineScientificNotationFromString(in_DataToPlot, 
> in_UseScientificNotationIfNeeded)
>     useOffsetIfNeeded = DetermineOnOrOffFromString(in_UseOffsetIfNeeded)
>
>     numberOfBins = len(in_DataToPlot) / 2
>     if numberOfBins > 25:
>         numberOfBins = 25
>     if numberOfBins < 5:
>         numberOfBins = 5
>
>     # first with 0, 0, 1, 1
>     fig, ax = CommonPlottingCode(in_WidthInPixels, in_HeightInPixels, 
> in_DataName, 'Frequency', useOffsetIfNeeded, scientificNotation, 
> False, 0.0, 0.0, 1.0, 1.0)
>    
>     # histogram of data
>     n, bins, patches = ax.hist(in_DataToPlot, numberOfBins, 
> facecolor=in_FillColor)
>    
>     # some axis space at the top of the graph
>     ylim = ax.get_ylim()
>     if ylim[1] == max(n):
>         ax.set_ylim(0.0, ylim[1] + 1)
>
>     newLeft, newBottom, newRight, newTop, numberOfMajor_X_TickMarks = 
> YieldNewExtentsAndNumberOfMajor_X_TickMarks(fig, ax, in_WidthInPixels, 
> in_HeightInPixels, False)
>
>     # now with scaled
>     fig, ax = CommonPlottingCode(in_WidthInPixels, in_HeightInPixels, 
> in_DataName, 'Frequency', useOffsetIfNeeded, scientificNotation, 
> False, newLeft, newBottom, newRight, newTop)
>    
>     # histogram of data
>     n, bins, patches = ax.hist(in_DataToPlot, numberOfBins, 
> facecolor=in_FillColor)
>
>     # some axis space at the top of the graph
>     ylim = ax.get_ylim()
>     if ylim[1] == max(n):
>         ax.set_ylim(0.0, ylim[1] + 1)
>
>     if  len(ax.get_xticklabels()) > numberOfMajor_X_TickMarks:
>         
> ax.xaxis.set_major_locator(matplotlib.ticker.MaxNLocator(numberOfMajor_X_TickMarks))
>
>     fig.savefig(in_FileNameAndPath, format = 'png', dpi=100)
>
>
> def ScatterPlot(in_DataToPlot, in_FileNameAndPath, in_DataNameX, 
> in_DataNameY, in_WidthInPixels, in_HeightInPixels,
>                 in_UseOffsetIfNeeded, in_ReverseXY, 
> in_X_UseScientificNotationIfNeeded, in_Y_UseScientificNotationIfNeeded):
>
>     # decode ends of strings ('XYZ_ON', 'XYZ_OFF', 'XYZ_AUTO', etc.) 
> to boolean values
>     scientificNotationX = 
> DetermineScientificNotationFromString(in_DataToPlot[0], 
> in_X_UseScientificNotationIfNeeded)
>     scientificNotationY = 
> DetermineScientificNotationFromString(in_DataToPlot[1], 
> in_Y_UseScientificNotationIfNeeded)
>     useOffsetIfNeeded = DetermineOnOrOffFromString(in_UseOffsetIfNeeded)
>     reverseXY = DetermineOnOrOffFromString(in_ReverseXY)
>
>     if reverseXY:
>         fig, ax = CommonPlottingCode(in_WidthInPixels, 
> in_HeightInPixels, in_DataNameY, in_DataNameX, useOffsetIfNeeded, 
> scientificNotationX, scientificNotationY, 0.0, 0.0, 1.0, 1.0)
>         ax.plot(np.array([min(in_DataToPlot[1]), 
> max(in_DataToPlot[1])]), np.array([min(in_DataToPlot[0]), 
> max(in_DataToPlot[0])]), 'o') # first ax.plot() is only with extents
>         newLeft, newBottom, newRight, newTop, 
> numberOfMajor_X_TickMarks = 
> YieldNewExtentsAndNumberOfMajor_X_TickMarks(fig, ax, in_WidthInPixels, 
> in_HeightInPixels, scientificNotationY or useOffsetIfNeeded)
>         fig, ax = CommonPlottingCode(in_WidthInPixels, 
> in_HeightInPixels, in_DataNameY, in_DataNameX, useOffsetIfNeeded, 
> scientificNotationX, scientificNotationY, newLeft, newBottom, 
> newRight, newTop)
>         if  len(ax.get_xticklabels()) > numberOfMajor_X_TickMarks:
>             
> ax.xaxis.set_major_locator(matplotlib.ticker.MaxNLocator(numberOfMajor_X_TickMarks))
>         ax.plot(in_DataToPlot[1], in_DataToPlot[0], 'o') # now that 
> autoscaling is done, use all data for second ax.plot()
>     else:
>         fig, ax = CommonPlottingCode(in_WidthInPixels, 
> in_HeightInPixels, in_DataNameX, in_DataNameY, useOffsetIfNeeded, 
> scientificNotationY, scientificNotationX, 0.0, 0.0, 1.0, 1.0)
>         ax.plot(np.array([min(in_DataToPlot[0]), 
> max(in_DataToPlot[0])]), np.array([min(in_DataToPlot[1]), 
> max(in_DataToPlot[1])]), 'o') # first ax.plot() is only with extents
>         newLeft, newBottom, newRight, newTop, 
> numberOfMajor_X_TickMarks = 
> YieldNewExtentsAndNumberOfMajor_X_TickMarks(fig, ax, in_WidthInPixels, 
> in_HeightInPixels, scientificNotationY or useOffsetIfNeeded)
>         fig, ax = CommonPlottingCode(in_WidthInPixels, 
> in_HeightInPixels, in_DataNameX, in_DataNameY, useOffsetIfNeeded, 
> scientificNotationY, scientificNotationX, newLeft, newBottom, 
> newRight, newTop)
>         if  len(ax.get_xticklabels()) > numberOfMajor_X_TickMarks:
>             
> ax.xaxis.set_major_locator(matplotlib.ticker.MaxNLocator(numberOfMajor_X_TickMarks))
>         ax.plot(in_DataToPlot[0], in_DataToPlot[1], 'o') # now that 
> autoscaling is done, use all data for second ax.plot()
>    
>     fig.savefig(in_FileNameAndPath, format = 'png', dpi=100)
>
>
> if __name__ in ('main', '__main__'):
>
>     testData1D = 12345678901.5 + np.random.randn(100)
>     testData2D = [testData1D, 1000.0 * testData1D + 1500 + 200.0 * 
> np.random.randn(100)]
>    
>     # note file names
>     HistogramPlot(testData1D, 'test_histogram_large.png', 'Test Data 
> Name', 'lightgrey',
>                   1024, 768, 'UseOffset_ON', 'ScientificNotation_ON')
>                  
>     HistogramPlot(testData1D, 'test_histogram_small.png', 'Test Data 
> Name', 'lightgrey',
>                   320, 240, 'UseOffset_ON', 'ScientificNotation_ON')
>                  
>     ScatterPlot(testData2D, 'test_scatterplot_small.png', 'Test Data X 
> Name', 'Test Data Y Name',
>                 320, 240, 'UseOffset_ON', 'ReverseXY_OFF', 
> 'ScientificNotation_X_OFF', 'ScientificNotation_Y_OFF')
>                
>     ScatterPlot(testData2D, 'test_scatterplot_large.png', 'Test Data X 
> Name', 'Test Data Y Name',
>                 1024, 768, 'UseOffset_ON', 'ReverseXY_ON', 
> 'ScientificNotation_X_OFF', 'ScientificNotation_Y_ON')
>
> ------------------------------------------------------------------------
>
> ------------------------------------------------------------------------------
> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
> software that enables intelligent coding and step-through debugging.
> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
> ------------------------------------------------------------------------
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

Trying your suggestion did not work.  However, doing this:

export CPP=/usr/bin/g++
python setup.py build

did work.  Go figure.

--Mike

Michael Droettboom wrote:
> Hmm... the selection of the compiler command is actually deep within 
> distutils -- matplotlib doesn't address it.  I'm surprised the CXX 
> environment variable isn't getting picked up.  I use that all the time 
> to test different compilers.
>
> Did you try:
>
> export CXX=/usr/bin/g++
> python setup.py build
>
> If that doesn't work, I'm at a loss -- you could start investigating 
> on distutils and/or gcc lists.
>
> Cheers,
> Mike
>
>
>
> Michael Hearne wrote:
>> I tried setting CXX=/usr/bin/g++ in my .bashrc, but I still get the 
>> same error.  Is there some file in the matplotlib distribution that I 
>> need to edit with this information?
>>
>> --Mike
>>
>> Michael Hearne wrote:
>>> I can't using gcc, but I can using g++, which is installed.  I guess 
>>> I'll just set the CXX variable to the path for g++.
>>>
>>> I don't know enough about gcc to understand why the two aren't 
>>> linked together...
>>>
>>> Thanks for the help,
>>>
>>> Mike
>>> Michael Droettboom wrote:
>>>  
>>>> Michael Hearne wrote:
>>>>   
>>>>> Two questions:
>>>>> 1) I'm trying to upgrade an installation of matplotlib I have on a 
>>>>> RHEL5 system.  When trying:
>>>>>
>>>>> /usr/local/bin/python setup.py build
>>>>>
>>>>> I get the error message:
>>>>> "gcc: src/ft2font.cpp: C++ compiler not installed on this system
>>>>> error: command 'gcc' failed with exit status 1"
>>>>>
>>>>> I do in fact have a C++ compiler on the system, in the form of 
>>>>> g++.  Is there a place where I can configure the C++ compiler to use?
>>>>>         
>>>> distutils will use the CXX environment variable if it is set.  
>>>> Though you should be able to compile C++ with gcc as well, if the 
>>>> C++ backend is installed -- that's why the error message is 
>>>> surprising to me if you're certain you have g++ installed.  Can you 
>>>> compile a simple C++ file with gcc directly from the commandline?
>>>>
>>>> Cheers,
>>>> Mike
>>>>
>>>>     
>>>
>>>
>>> ------------------------------------------------------------------------------ 
>>>
>>> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
>>> powering Web 2.0 with engaging, cross-platform capabilities. Quickly 
>>> and
>>> easily build your RIAs with Flex Builder, the Eclipse(TM)based 
>>> development
>>> software that enables intelligent coding and step-through debugging.
>>> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
>>> _______________________________________________
>>> Matplotlib-users mailing list
>>> Mat...@li...
>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>>   
>>
>

Hmm... the selection of the compiler command is actually deep within 
distutils -- matplotlib doesn't address it.  I'm surprised the CXX 
environment variable isn't getting picked up.  I use that all the time 
to test different compilers.

Did you try:

export CXX=/usr/bin/g++
python setup.py build

If that doesn't work, I'm at a loss -- you could start investigating on 
distutils and/or gcc lists.

Cheers,
Mike



Michael Hearne wrote:
> I tried setting CXX=/usr/bin/g++ in my .bashrc, but I still get the 
> same error.  Is there some file in the matplotlib distribution that I 
> need to edit with this information?
>
> --Mike
>
> Michael Hearne wrote:
>> I can't using gcc, but I can using g++, which is installed.  I guess 
>> I'll just set the CXX variable to the path for g++.
>>
>> I don't know enough about gcc to understand why the two aren't linked 
>> together...
>>
>> Thanks for the help,
>>
>> Mike
>> Michael Droettboom wrote:
>>  
>>> Michael Hearne wrote:
>>>    
>>>> Two questions:
>>>> 1) I'm trying to upgrade an installation of matplotlib I have on a 
>>>> RHEL5 system.  When trying:
>>>>
>>>> /usr/local/bin/python setup.py build
>>>>
>>>> I get the error message:
>>>> "gcc: src/ft2font.cpp: C++ compiler not installed on this system
>>>> error: command 'gcc' failed with exit status 1"
>>>>
>>>> I do in fact have a C++ compiler on the system, in the form of 
>>>> g++.  Is there a place where I can configure the C++ compiler to use?
>>>>         
>>> distutils will use the CXX environment variable if it is set.  
>>> Though you should be able to compile C++ with gcc as well, if the 
>>> C++ backend is installed -- that's why the error message is 
>>> surprising to me if you're certain you have g++ installed.  Can you 
>>> compile a simple C++ file with gcc directly from the commandline?
>>>
>>> Cheers,
>>> Mike
>>>
>>>     
>>
>>
>> ------------------------------------------------------------------------------ 
>>
>> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
>> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
>> easily build your RIAs with Flex Builder, the Eclipse(TM)based 
>> development
>> software that enables intelligent coding and step-through debugging.
>> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>   
>

-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

Stano Paška wrote:
> Hi,
> I am using matplotlib first time and I am wonder how to free resources
> 
> I have this script
> 
> # coding: utf-8
> 
> import matplotlib
> matplotlib.use('Agg')
> 
> import pylab
> 
> class Graph:
> 
>     def __init__(self):
>         # sirka stlpcov
>         self.width = 0.5
>         self.font = {'fontname':'Tahoma', 'fontsize':10}
> 
>     def setTitle(self, title):
>         """Nastavi nadpis grafu
>         """
>         self.title = title
> 
>     def setXLabel(self, label):
>         self.x_label = label
> 
>     def setYLabel(self, label):
>         self.y_label = label
> 
>     def setYValues(self, values):
>         """Nastavi y-ove hodnoty
>         """
>         self.y_values = values
>         # vypocitame lave spodne rohy stlpcov
>         self.x_values = [x + self.width / 2 for x in range(len(self.y_values))]
>         # vypocitame stred stlpcov (znacky na x-ovej osi)
>         self.x_ticks = [x + self.width for x in range(len(self.y_values))]
> 
>     def setXTickLabels(self, labels):
>         """Nastavi popisky pre ciarky na x-ovej osi
>         """
>         self.x_ticksLabels = labels
> 
>     def makeGraph(self):
>         self.fig = pylab.figure()
>         self.fig.set_dpi(72)
>         self.fig.set_figheight(3)
>         self.fig.set_figwidth(5)
> 
>         self.fig.subplots_adjust(bottom=0.15)
> 
>         self.ax = self.fig.add_subplot(111)
> 
>         self.ax.bar(left=self.x_values, height=self.y_values,
> width=self.width, color='#f00000')
>         self.ax.set_xticks(self.x_ticks)
>         self.ax.set_xticklabels(self.x_ticksLabels)
> 
>         self.ax.set_xlabel(self.x_label, **self.font)
>         self.ax.set_ylabel(self.y_label, **self.font)
>         self.ax.set_title(self.title, **self.font)
> 
>     def saveGraph(self, path):
>         self.fig.savefig(path, format='png')
> 
>     def __del__(self):
>         self.fig.delaxes(self.ax)
>         del self.fig
>         del self.ax
> 
> if __name__ == '__main__':
>     import random
>     import gc
>     for i in range(5000):
>         g = Graph()
>         g.setTitle(u'Spotreba za rok 2008 podľa ATC skupín')
>         g.setXLabel(u'ATC skupiny')
>         g.setYLabel(u'Spotreba v EUR')
>         g.setYValues([random.randint(0, x) for x in range(20)])
>         g.setXTickLabels([chr(random.randint(65, 90)) for x in range(20)])
>         g.makeGraph()
>         g.saveGraph('grafy/' + str(i) + '.png')
>         del g
>         print gc.get_count()
>         gc.collect()
>         print gc.get_count()
> 
> When I used TkAgg, it crashes after 190 images (Fail to create pixmap
> with Tk_GetPixmap in ImgPhotoInstanceSetSize. tried to delete photo
> image when instances still exist).
> When I switch to WX, it crashes after 400 images.
> And Agg crashes when pagefile reaches 1.5x ram (cca 2000 images).
> 
> It is possible somehow free resources in matplotlib?

You need to close the figure.  Try putting
pylab.close(self.fig) at the start of your __del__ method.  In fact, I 
doubt you need to explicitly del anything in that method--you just need 
to close the figure so that pylab will release its references related to 
that figure.

There may be more things you need to change, but closing the figure is 
certainly going to be a big one.

Eric

> 
> I use python 2.5.4, wx 2.8.9.2, numpy 1.3.0b1 on win XP
> 
> Thanks for answers.
> 
> Stano.
> 
> ------------------------------------------------------------------------------
> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
> software that enables intelligent coding and step-through debugging.
> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

I tried setting CXX=/usr/bin/g++ in my .bashrc, but I still get the same 
error.  Is there some file in the matplotlib distribution that I need to 
edit with this information?

--Mike

Michael Hearne wrote:
> I can't using gcc, but I can using g++, which is installed.  I guess 
> I'll just set the CXX variable to the path for g++.
>
> I don't know enough about gcc to understand why the two aren't linked 
> together...
>
> Thanks for the help,
>
> Mike
> Michael Droettboom wrote:
>   
>> Michael Hearne wrote:
>>     
>>> Two questions:
>>> 1) I'm trying to upgrade an installation of matplotlib I have on a 
>>> RHEL5 system.  When trying:
>>>
>>> /usr/local/bin/python setup.py build
>>>
>>> I get the error message:
>>> "gcc: src/ft2font.cpp: C++ compiler not installed on this system
>>> error: command 'gcc' failed with exit status 1"
>>>
>>> I do in fact have a C++ compiler on the system, in the form of g++.  
>>> Is there a place where I can configure the C++ compiler to use?
>>>   
>>>       
>> distutils will use the CXX environment variable if it is set.  Though 
>> you should be able to compile C++ with gcc as well, if the C++ backend 
>> is installed -- that's why the error message is surprising to me if 
>> you're certain you have g++ installed.  Can you compile a simple C++ 
>> file with gcc directly from the commandline?
>>
>> Cheers,
>> Mike
>>
>>     
>
>
> ------------------------------------------------------------------------------
> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
> software that enables intelligent coding and step-through debugging.
> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>   

I can't using gcc, but I can using g++, which is installed.  I guess 
I'll just set the CXX variable to the path for g++.

I don't know enough about gcc to understand why the two aren't linked 
together...

Thanks for the help,

Mike
Michael Droettboom wrote:
> Michael Hearne wrote:
>> Two questions:
>> 1) I'm trying to upgrade an installation of matplotlib I have on a 
>> RHEL5 system.  When trying:
>>
>> /usr/local/bin/python setup.py build
>>
>> I get the error message:
>> "gcc: src/ft2font.cpp: C++ compiler not installed on this system
>> error: command 'gcc' failed with exit status 1"
>>
>> I do in fact have a C++ compiler on the system, in the form of g++.  
>> Is there a place where I can configure the C++ compiler to use?
>>   
> distutils will use the CXX environment variable if it is set.  Though 
> you should be able to compile C++ with gcc as well, if the C++ backend 
> is installed -- that's why the error message is surprising to me if 
> you're certain you have g++ installed.  Can you compile a simple C++ 
> file with gcc directly from the commandline?
>
> Cheers,
> Mike
>

Michael Droettboom wrote:
> Unfortunately, the EMF backend is no longer maintained.  We had 
> discussed removing it well over a year ago, and I actually thought it 
> had been removed, but obviously it wasn't.
> 
> It needs someone to port it to the new backend API introduced in 0.98.x 
> and to maintain that code.  When quite some time ago, there really 
> wasn't much interest in doing that.  I would be happy to help someone 
> with the details involved, but don't have the time myself at the moment.
> 
> Sorry for the confusion -- what do other developers think?  Should we 
> remove the backend and document the change, at least for the time being?

Mike,

Yes, like you I thought we had included it on the to-be-removed list, 
but evidently I remembered incorrectly.

Here is the history of the file:

10 months  	jdh2358  	removed pre py2.4 compatability code and some 
other dead code
10 months 	mdboom 	Merged revisions 5106-5108 via svnmerge from
18 months 	mdboom 	Refactoring of FigureCanvas*.print_figure so that 
each backend does
19 months 	efiring 	Use pickle to cache entire fontManager instance.
21 months 	nnemec 	remove trailing spaces
2 years 	cmoad 	added rotation arg to draw_arc method everywhere it was 
missing
2 years 	jdh2358 	added custom figure class hook
2 years 	jdh2358 	fixed figimage bug
3 years 	jdh2358 	added Aarre linestyle fix for emf backend
3 years 	dsdale 	added **kwargs to all backend_*.print_figure
3 years 	jdh2358 	made the rc warnings a bit more readable
3 years 	jdh2358 	added pyemf

I think it should be removed now.  If someone comes forward with a 
strong commitment to update and maintain it, it can be reinstated 
easily--after that person has done the updating required to make it work 
in the trunk.

Eric


> 
> Mike
> 
> Torsten Bronger wrote:
>> Hallöchen!
>>
>> I use pyemf 2.0 and Matplotlib 0.98.5.  If I try to save a trivial
>> plot into an EMF file, I get
>>
>> Traceback (most recent call last):
>>   File "./make_plots.py", line 23, in <module>
>>     figure.savefig(sample_name + "_pds.emf")
>>   File "/usr/lib/python2.5/[...]/matplotlib/figure.py", line 990, in savefig
>>     self.canvas.print_figure(*args, **kwargs)
>>   File "/usr/lib/python2.5/[...]/matplotlib/backend_bases.py", line 1419, in print_figure
>>     **kwargs)
>>   File "/usr/lib/python2.5/[...]/matplotlib/backends/backend_emf.py", line 604, in print_emf
>>     self.figure.draw(renderer)
>>   File "/usr/lib/python2.5/[...]/matplotlib/figure.py", line 738, in draw
>>     if self.frameon: self.patch.draw(renderer)
>>   File "/usr/lib/python2.5/[...]/matplotlib/patches.py", line 301, in draw
>>     renderer.draw_path(gc, tpath, affine, rgbFace)
>>   File "/usr/lib/python2.5/[...]/matplotlib/backend_bases.py", line 74, in draw_path
>>     raise NotImplementedError
>> NotImplementedError
>>
>> Is this a bug or a mal-configuration?  The Python code was
>>
>>             figure = pyplot.figure()
>>             axes = figure.add_subplot(111)
>>             x, y = numpy.loadtxt("data.dat", comments="#", usecols=(0, 1), unpack=True)
>>             axes.semilogy(x, y)
>>             figure.savefig("test.emf")
>>
>> Tschö,
>> Torsten.
>>
>>   
> 

Hi,
I am using matplotlib first time and I am wonder how to free resources

I have this script

# coding: utf-8

import matplotlib
matplotlib.use('Agg')

import pylab

class Graph:

    def __init__(self):
        # sirka stlpcov
        self.width = 0.5
        self.font = {'fontname':'Tahoma', 'fontsize':10}

    def setTitle(self, title):
        """Nastavi nadpis grafu
        """
        self.title = title

    def setXLabel(self, label):
        self.x_label = label

    def setYLabel(self, label):
        self.y_label = label

    def setYValues(self, values):
        """Nastavi y-ove hodnoty
        """
        self.y_values = values
        # vypocitame lave spodne rohy stlpcov
        self.x_values = [x + self.width / 2 for x in range(len(self.y_values))]
        # vypocitame stred stlpcov (znacky na x-ovej osi)
        self.x_ticks = [x + self.width for x in range(len(self.y_values))]

    def setXTickLabels(self, labels):
        """Nastavi popisky pre ciarky na x-ovej osi
        """
        self.x_ticksLabels = labels

    def makeGraph(self):
        self.fig = pylab.figure()
        self.fig.set_dpi(72)
        self.fig.set_figheight(3)
        self.fig.set_figwidth(5)

        self.fig.subplots_adjust(bottom=0.15)

        self.ax = self.fig.add_subplot(111)

        self.ax.bar(left=self.x_values, height=self.y_values,
width=self.width, color='#f00000')
        self.ax.set_xticks(self.x_ticks)
        self.ax.set_xticklabels(self.x_ticksLabels)

        self.ax.set_xlabel(self.x_label, **self.font)
        self.ax.set_ylabel(self.y_label, **self.font)
        self.ax.set_title(self.title, **self.font)

    def saveGraph(self, path):
        self.fig.savefig(path, format='png')

    def __del__(self):
        self.fig.delaxes(self.ax)
        del self.fig
        del self.ax

if __name__ == '__main__':
    import random
    import gc
    for i in range(5000):
        g = Graph()
        g.setTitle(u'Spotreba za rok 2008 podľa ATC skupín')
        g.setXLabel(u'ATC skupiny')
        g.setYLabel(u'Spotreba v EUR')
        g.setYValues([random.randint(0, x) for x in range(20)])
        g.setXTickLabels([chr(random.randint(65, 90)) for x in range(20)])
        g.makeGraph()
        g.saveGraph('grafy/' + str(i) + '.png')
        del g
        print gc.get_count()
        gc.collect()
        print gc.get_count()

When I used TkAgg, it crashes after 190 images (Fail to create pixmap
with Tk_GetPixmap in ImgPhotoInstanceSetSize. tried to delete photo
image when instances still exist).
When I switch to WX, it crashes after 400 images.
And Agg crashes when pagefile reaches 1.5x ram (cca 2000 images).

It is possible somehow free resources in matplotlib?

I use python 2.5.4, wx 2.8.9.2, numpy 1.3.0b1 on win XP

Thanks for answers.

Stano.

Michael Hearne wrote:
> Two questions:
> 1) I'm trying to upgrade an installation of matplotlib I have on a RHEL5 
> system.  When trying:
>
> /usr/local/bin/python setup.py build
>
> I get the error message:
> "gcc: src/ft2font.cpp: C++ compiler not installed on this system
> error: command 'gcc' failed with exit status 1"
>
> I do in fact have a C++ compiler on the system, in the form of g++.  Is 
> there a place where I can configure the C++ compiler to use?
>   
distutils will use the CXX environment variable if it is set.  Though 
you should be able to compile C++ with gcc as well, if the C++ backend 
is installed -- that's why the error message is surprising to me if 
you're certain you have g++ installed.  Can you compile a simple C++ 
file with gcc directly from the commandline?

Cheers,
Mike

-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

Michael Hearne wrote:
> 2)  So that I don't have to bother the list with things like this, how 
> can I _search_ the mailing list for keywords?  

I usually search mailing lists through gmane.org, for this particular 
list the link is:

http://dir.gmane.org/gmane.comp.python.matplotlib.general

JLS

Two questions:
1) I'm trying to upgrade an installation of matplotlib I have on a RHEL5 
system.  When trying:

/usr/local/bin/python setup.py build

I get the error message:
"gcc: src/ft2font.cpp: C++ compiler not installed on this system
error: command 'gcc' failed with exit status 1"

I do in fact have a C++ compiler on the system, in the form of g++.  Is 
there a place where I can configure the C++ compiler to use?

2)  So that I don't have to bother the list with things like this, how 
can I _search_ the mailing list for keywords?  If I go here:

http://sourceforge.net/mailarchive/forum.php?forum_name=matplotlib-users

I don't see any way to search the archives, other than manually paging 
through them.

 

Unfortunately, the EMF backend is no longer maintained.  We had 
discussed removing it well over a year ago, and I actually thought it 
had been removed, but obviously it wasn't.

It needs someone to port it to the new backend API introduced in 0.98.x 
and to maintain that code.  When quite some time ago, there really 
wasn't much interest in doing that.  I would be happy to help someone 
with the details involved, but don't have the time myself at the moment.

Sorry for the confusion -- what do other developers think?  Should we 
remove the backend and document the change, at least for the time being?

Mike

Torsten Bronger wrote:
> Hallöchen!
>
> I use pyemf 2.0 and Matplotlib 0.98.5.  If I try to save a trivial
> plot into an EMF file, I get
>
> Traceback (most recent call last):
>   File "./make_plots.py", line 23, in <module>
>     figure.savefig(sample_name + "_pds.emf")
>   File "/usr/lib/python2.5/[...]/matplotlib/figure.py", line 990, in savefig
>     self.canvas.print_figure(*args, **kwargs)
>   File "/usr/lib/python2.5/[...]/matplotlib/backend_bases.py", line 1419, in print_figure
>     **kwargs)
>   File "/usr/lib/python2.5/[...]/matplotlib/backends/backend_emf.py", line 604, in print_emf
>     self.figure.draw(renderer)
>   File "/usr/lib/python2.5/[...]/matplotlib/figure.py", line 738, in draw
>     if self.frameon: self.patch.draw(renderer)
>   File "/usr/lib/python2.5/[...]/matplotlib/patches.py", line 301, in draw
>     renderer.draw_path(gc, tpath, affine, rgbFace)
>   File "/usr/lib/python2.5/[...]/matplotlib/backend_bases.py", line 74, in draw_path
>     raise NotImplementedError
> NotImplementedError
>
> Is this a bug or a mal-configuration?  The Python code was
>
>             figure = pyplot.figure()
>             axes = figure.add_subplot(111)
>             x, y = numpy.loadtxt("data.dat", comments="#", usecols=(0, 1), unpack=True)
>             axes.semilogy(x, y)
>             figure.savefig("test.emf")
>
> Tschö,
> Torsten.
>
>   

-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

All,

     Attached, and below, is public domain code for making variable-sized
plots with autoscaled text that exactly fits the available visual plot
space, useful for web sites where users choose output files with different
sizes.  Examples are at the bottom of the file.

          James R. Phillips
          2548 Vera Cruz Drive
          Birmingham, AL 35235 USA
          email: zu...@zu...
          http://zunzun.com



#    Entered into the public domain 20 March 2009
#    James R. Phillips
#    2548 Vera Cruz Drive
#    Birmingham, AL 35235 USA
#    email: zu...@zu...
#    http://zunzun.com

import numpy as np
import math, matplotlib
matplotlib.use('Agg') # must be used prior to the next two statements
import matplotlib.pyplot as plt
import matplotlib.mlab as mlab


def DetermineOnOrOffFromString(in_String):
    tempString = in_String.split('_')[-1:][0].upper() # allows any amount of
prefacing text
    if tempString == 'ON':
        return True
    return False


def DetermineScientificNotationFromString(inData, in_String):
    tempString = in_String.split('_')[-1:][0].upper() # allows any amount of
prefacing text
    if tempString == 'ON':
        return True
    elif tempString == 'OFF':
        return False
    else: # must be AUTO
        minVal = np.abs(np.min(inData))
        maxVal = np.abs(np.max(inData))
        deltaVal = np.abs(maxVal - minVal)

        scientificNotation = False
        if (maxVal > 100.0) or (minVal < -100.0) or (deltaVal < .05):
            scientificNotation = True
    return scientificNotation


def CommonPlottingCode(in_WidthInPixels, in_HeightInPixels, in_XName,
in_YName, in_UseOffsetIfNeeded, in_X_UseScientificNotationIfNeeded,
in_Y_UseScientificNotationIfNeeded, in_Left, in_Bottom, in_Right, in_Top): #
default to lots of room around graph

    # a litle more room between x axis and tick mark labels, so not text
overlap at the bottom left corner - set this before other calls
    matplotlib.rcParams['xtick.major.pad'] = 5+ (float(in_HeightInPixels) /
100.0) # minimum + some scaled

    fig = plt.figure(figsize=(float(in_WidthInPixels ) / 100.0,
float(in_HeightInPixels ) / 100.0), dpi=100)
    fig.subplotpars.update(in_Left, in_Bottom, in_Right, in_Top)
    ax = fig.add_subplot(111, frameon=True)

    # white background, almost no border space
    fig.set_facecolor('w')

    xFormatter = fig.gca().xaxis.get_major_formatter()
    xFormatter._useOffset = in_UseOffsetIfNeeded
    xFormatter.set_scientific(in_X_UseScientificNotationIfNeeded)
    fig.gca().xaxis.set_major_formatter(xFormatter)

    yFormatter = fig.gca().yaxis.get_major_formatter()
    yFormatter._useOffset = in_UseOffsetIfNeeded
    yFormatter.set_scientific(in_Y_UseScientificNotationIfNeeded)
    fig.gca().yaxis.set_major_formatter(yFormatter)

    # Scale text to imagesize.  Text sizes originally determined at image
size of 500 x 400
    widthRatioForTextSize = float(in_WidthInPixels) / 500.0
    heightRatioForTextSize = float(in_HeightInPixels) / 400.0
    for xlabel_i in ax.get_xticklabels():
        xlabel_i.set_fontsize(xlabel_i.get_fontsize() *
heightRatioForTextSize)
    xOffsetText = fig.gca().xaxis.get_offset_text()
    xOffsetText.set_fontsize(xOffsetText.get_fontsize() *
heightRatioForTextSize * 0.9)
    for ylabel_i in ax.get_yticklabels():
        ylabel_i.set_fontsize(ylabel_i.get_fontsize() *
widthRatioForTextSize)
    yOffsetText = fig.gca().yaxis.get_offset_text()
    yOffsetText.set_fontsize(yOffsetText.get_fontsize() *
heightRatioForTextSize * 0.9)

    x_label = ax.set_xlabel(in_XName)
    y_label = ax.set_ylabel(in_YName)
    x_label._fontproperties._size = x_label._fontproperties._size *
heightRatioForTextSize
    y_label._fontproperties._size = y_label._fontproperties._size *
widthRatioForTextSize

    plt.grid(True) # call this just before returning

    return fig, ax


def YieldNewExtentsAndNumberOfMajor_X_TickMarks(fig, ax, in_WidthInPixels,
in_HeightInPixels, in_OffsetUsed):
    # draw everything so items can be measured for size
    canvas = plt.get_current_fig_manager().canvas
    canvas.draw()

    # some preliminary info
    xLabelPoints =
ax.set_xlabel(ax.get_xlabel()).get_window_extent().get_points() # [ [x,y],
[x,y] ]
    yLabelPoints =
ax.set_ylabel(ax.get_ylabel()).get_window_extent().get_points() # [ [x,y],
[x,y] ], rotated 90 degrees
    xTickZeroPoints =
ax.get_xticklabels()[0].get_window_extent().get_points()
    yTickZeroPoints =
ax.get_yticklabels()[0].get_window_extent().get_points()
    xTickIndexPoints =
ax.get_xticklabels()[len(ax.get_xticklabels())-1].get_window_extent().get_points()
    yTickIndexPoints =
ax.get_yticklabels()[len(ax.get_yticklabels())-1].get_window_extent().get_points()
    currentPoints = ax.bbox.get_points()
    maxLeft = currentPoints[0][0]
    maxBottom = currentPoints[0][1]
    maxRight = currentPoints[1][0]
    maxTop = currentPoints[1][1]

    # find the most left-ward location
    if xTickZeroPoints[0][0] < maxLeft:
        maxLeft = xTickZeroPoints[0][0]
    if yTickZeroPoints[0][0] < maxLeft:
        maxLeft = yTickZeroPoints[0][0]
    if yTickIndexPoints[0][0] < maxLeft:
        maxLeft = yTickIndexPoints[0][0]
    if xLabelPoints[0][0] < maxLeft:
        maxLeft = xLabelPoints[0][0]
    if yLabelPoints[0][0] < maxLeft: # 90 degrees
        maxLeft = yLabelPoints[0][0]

    # find the most right-ward location
    if xTickIndexPoints[1][0] > maxRight:
        maxRight = xTickIndexPoints[1][0]
    if xLabelPoints[1][0] > maxRight:
        maxRight = xLabelPoints[1][0]

    # find the most bottom-ward location
    if xTickZeroPoints[0][1] < maxBottom:
        maxBottom = xTickZeroPoints[0][1]
    if xLabelPoints[0][1] < maxBottom:
        maxBottom = xLabelPoints[0][1]
    if yLabelPoints[0][1] < maxBottom:
        maxBottom = yLabelPoints[0][1]

    # find the most top-ward location
    if yTickIndexPoints[1][1] > maxTop:
        maxTop = yTickIndexPoints[1][1]
    if True == in_OffsetUsed: # could not find a better way to get this
        yp = ax.get_yticklabels()[0].get_window_extent().get_points()
        maxTop += yp[1][1] - yp[0][1]

    newLeft = ax.bbox._bbox.get_points()[0][0] - (float(maxLeft) /
float(in_WidthInPixels)) + 0.01
    newBottom = ax.bbox._bbox.get_points()[0][1] - (float(maxBottom) /
float(in_HeightInPixels)) + 0.01
    newRight = ax.bbox._bbox.get_points()[1][0] + (1.0 - (float(maxRight) /
float(in_WidthInPixels))) - 0.01
    newTop = ax.bbox._bbox.get_points()[1][1] + (1.0 - (float(maxTop) /
float(in_HeightInPixels))) - 0.01

    # now redraw and check number of X tick marks
    canvas.draw()

    # Calculate major number of X tick marks based on label size
    totalWidth = 0.0
    maxWidth = 0.0
    numberOfMajor_X_TickMarks = len(ax.get_xticklabels())
    for xlabel_i in ax.get_xticklabels():
        w = xlabel_i.get_window_extent().get_points() # the drawn text
bounding box corners as numpy array of [x,y], [x,y]
        width = w[1][0] - w[0][0]
        totalWidth += width
        if width > maxWidth:
            maxWidth = width
    if totalWidth > (0.95 * ((newRight - newLeft) *
float(in_WidthInPixels))): # 0.95 for some spacing between tick labels
        numberOfMajor_X_TickMarks = int(math.floor((0.95 * ((newRight -
newLeft) * float(in_WidthInPixels))) / maxWidth))

    return (newLeft, newBottom, newRight, newTop,
numberOfMajor_X_TickMarks,)


def HistogramPlot(in_DataToPlot, in_FileNameAndPath, in_DataName,
in_FillColor, in_WidthInPixels, in_HeightInPixels, in_UseOffsetIfNeeded,
in_UseScientificNotationIfNeeded):

    # decode ends of strings ('XYZ_ON', 'XYZ_OFF', 'XYZ_AUTO', etc.) to
boolean values
    scientificNotation =
DetermineScientificNotationFromString(in_DataToPlot,
in_UseScientificNotationIfNeeded)
    useOffsetIfNeeded = DetermineOnOrOffFromString(in_UseOffsetIfNeeded)

    numberOfBins = len(in_DataToPlot) / 2
    if numberOfBins > 25:
        numberOfBins = 25
    if numberOfBins < 5:
        numberOfBins = 5

    # first with 0, 0, 1, 1
    fig, ax = CommonPlottingCode(in_WidthInPixels, in_HeightInPixels,
in_DataName, 'Frequency', useOffsetIfNeeded, scientificNotation, False, 0.0,
0.0, 1.0, 1.0)

    # histogram of data
    n, bins, patches = ax.hist(in_DataToPlot, numberOfBins,
facecolor=in_FillColor)

    # some axis space at the top of the graph
    ylim = ax.get_ylim()
    if ylim[1] == max(n):
        ax.set_ylim(0.0, ylim[1] + 1)

    newLeft, newBottom, newRight, newTop, numberOfMajor_X_TickMarks =
YieldNewExtentsAndNumberOfMajor_X_TickMarks(fig, ax, in_WidthInPixels,
in_HeightInPixels, False)

    # now with scaled
    fig, ax = CommonPlottingCode(in_WidthInPixels, in_HeightInPixels,
in_DataName, 'Frequency', useOffsetIfNeeded, scientificNotation, False,
newLeft, newBottom, newRight, newTop)

    # histogram of data
    n, bins, patches = ax.hist(in_DataToPlot, numberOfBins,
facecolor=in_FillColor)

    # some axis space at the top of the graph
    ylim = ax.get_ylim()
    if ylim[1] == max(n):
        ax.set_ylim(0.0, ylim[1] + 1)

    if  len(ax.get_xticklabels()) > numberOfMajor_X_TickMarks:

ax.xaxis.set_major_locator(matplotlib.ticker.MaxNLocator(numberOfMajor_X_TickMarks))

    fig.savefig(in_FileNameAndPath, format = 'png', dpi=100)


def ScatterPlot(in_DataToPlot, in_FileNameAndPath, in_DataNameX,
in_DataNameY, in_WidthInPixels, in_HeightInPixels,
                in_UseOffsetIfNeeded, in_ReverseXY,
in_X_UseScientificNotationIfNeeded, in_Y_UseScientificNotationIfNeeded):

    # decode ends of strings ('XYZ_ON', 'XYZ_OFF', 'XYZ_AUTO', etc.) to
boolean values
    scientificNotationX =
DetermineScientificNotationFromString(in_DataToPlot[0],
in_X_UseScientificNotationIfNeeded)
    scientificNotationY =
DetermineScientificNotationFromString(in_DataToPlot[1],
in_Y_UseScientificNotationIfNeeded)
    useOffsetIfNeeded = DetermineOnOrOffFromString(in_UseOffsetIfNeeded)
    reverseXY = DetermineOnOrOffFromString(in_ReverseXY)

    if reverseXY:
        fig, ax = CommonPlottingCode(in_WidthInPixels, in_HeightInPixels,
in_DataNameY, in_DataNameX, useOffsetIfNeeded, scientificNotationX,
scientificNotationY, 0.0, 0.0, 1.0, 1.0)
        ax.plot(np.array([min(in_DataToPlot[1]), max(in_DataToPlot[1])]),
np.array([min(in_DataToPlot[0]), max(in_DataToPlot[0])]), 'o') # first
ax.plot() is only with extents
        newLeft, newBottom, newRight, newTop, numberOfMajor_X_TickMarks =
YieldNewExtentsAndNumberOfMajor_X_TickMarks(fig, ax, in_WidthInPixels,
in_HeightInPixels, scientificNotationY or useOffsetIfNeeded)
        fig, ax = CommonPlottingCode(in_WidthInPixels, in_HeightInPixels,
in_DataNameY, in_DataNameX, useOffsetIfNeeded, scientificNotationX,
scientificNotationY, newLeft, newBottom, newRight, newTop)
        if  len(ax.get_xticklabels()) > numberOfMajor_X_TickMarks:

ax.xaxis.set_major_locator(matplotlib.ticker.MaxNLocator(numberOfMajor_X_TickMarks))
        ax.plot(in_DataToPlot[1], in_DataToPlot[0], 'o') # now that
autoscaling is done, use all data for second ax.plot()
    else:
        fig, ax = CommonPlottingCode(in_WidthInPixels, in_HeightInPixels,
in_DataNameX, in_DataNameY, useOffsetIfNeeded, scientificNotationY,
scientificNotationX, 0.0, 0.0, 1.0, 1.0)
        ax.plot(np.array([min(in_DataToPlot[0]), max(in_DataToPlot[0])]),
np.array([min(in_DataToPlot[1]), max(in_DataToPlot[1])]), 'o') # first
ax.plot() is only with extents
        newLeft, newBottom, newRight, newTop, numberOfMajor_X_TickMarks =
YieldNewExtentsAndNumberOfMajor_X_TickMarks(fig, ax, in_WidthInPixels,
in_HeightInPixels, scientificNotationY or useOffsetIfNeeded)
        fig, ax = CommonPlottingCode(in_WidthInPixels, in_HeightInPixels,
in_DataNameX, in_DataNameY, useOffsetIfNeeded, scientificNotationY,
scientificNotationX, newLeft, newBottom, newRight, newTop)
        if  len(ax.get_xticklabels()) > numberOfMajor_X_TickMarks:

ax.xaxis.set_major_locator(matplotlib.ticker.MaxNLocator(numberOfMajor_X_TickMarks))
        ax.plot(in_DataToPlot[0], in_DataToPlot[1], 'o') # now that
autoscaling is done, use all data for second ax.plot()

    fig.savefig(in_FileNameAndPath, format = 'png', dpi=100)


if __name__ in ('main', '__main__'):

    testData1D = 12345678901.5 + np.random.randn(100)
    testData2D = [testData1D, 1000.0 * testData1D + 1500 + 200.0 *
np.random.randn(100)]

    # note file names
    HistogramPlot(testData1D, 'test_histogram_large.png', 'Test Data Name',
'lightgrey',
                  1024, 768, 'UseOffset_ON', 'ScientificNotation_ON')

    HistogramPlot(testData1D, 'test_histogram_small.png', 'Test Data Name',
'lightgrey',
                  320, 240, 'UseOffset_ON', 'ScientificNotation_ON')

    ScatterPlot(testData2D, 'test_scatterplot_small.png', 'Test Data X
Name', 'Test Data Y Name',
                320, 240, 'UseOffset_ON', 'ReverseXY_OFF',
'ScientificNotation_X_OFF', 'ScientificNotation_Y_OFF')

    ScatterPlot(testData2D, 'test_scatterplot_large.png', 'Test Data X
Name', 'Test Data Y Name',
                1024, 768, 'UseOffset_ON', 'ReverseXY_ON',
'ScientificNotation_X_OFF', 'ScientificNotation_Y_ON')

Hallöchen!

I use pyemf 2.0 and Matplotlib 0.98.5.  If I try to save a trivial
plot into an EMF file, I get

Traceback (most recent call last):
  File "./make_plots.py", line 23, in <module>
    figure.savefig(sample_name + "_pds.emf")
  File "/usr/lib/python2.5/[...]/matplotlib/figure.py", line 990, in savefig
    self.canvas.print_figure(*args, **kwargs)
  File "/usr/lib/python2.5/[...]/matplotlib/backend_bases.py", line 1419, in print_figure
    **kwargs)
  File "/usr/lib/python2.5/[...]/matplotlib/backends/backend_emf.py", line 604, in print_emf
    self.figure.draw(renderer)
  File "/usr/lib/python2.5/[...]/matplotlib/figure.py", line 738, in draw
    if self.frameon: self.patch.draw(renderer)
  File "/usr/lib/python2.5/[...]/matplotlib/patches.py", line 301, in draw
    renderer.draw_path(gc, tpath, affine, rgbFace)
  File "/usr/lib/python2.5/[...]/matplotlib/backend_bases.py", line 74, in draw_path
    raise NotImplementedError
NotImplementedError

Is this a bug or a mal-configuration?  The Python code was

            figure = pyplot.figure()
            axes = figure.add_subplot(111)
            x, y = numpy.loadtxt("data.dat", comments="#", usecols=(0, 1), unpack=True)
            axes.semilogy(x, y)
            figure.savefig("test.emf")

Tschö,
Torsten.

-- 
Torsten Bronger, aquisgrana, europa vetus
                   Jabber ID: tor...@ja...

Hi!
I'm a newbie of matplotlib, and I'm trying to plot a set of data... but I
got blocked...

$ cat matplotliberr.py
#!/usr/bin/env python

# dummy data to plot
from datetime import date, timedelta
from random import randint
x = [ date.today() + timedelta(i) for i in range(10) ]
y = [ randint(0, i) for i in range(10) ]

from matplotlib import pyplot
pyplot.fill(x, y) # no problem using pyplot.plot(x, y)
pyplot.show()
$ ./matplotliberr.py
/usr/lib/python2.6/site-packages/pytz/tzinfo.py:5: DeprecationWarning: the
sets module is deprecated
  from sets import Set
Traceback (most recent call last):
  File "./matplotliberr.py", line 10, in <module>
    pyplot.fill(x, y)
  File "/usr/lib/python2.6/site-packages/matplotlib/pyplot.py", line 1876,
in fill
    ret =  gca().fill(*args, **kwargs)
  File "/usr/lib/python2.6/site-packages/matplotlib/axes.py", line 5558, in
fill
    for poly in self._get_patches_for_fill(*args, **kwargs):
  File "/usr/lib/python2.6/site-packages/matplotlib/axes.py", line 394, in
_grab_next_args
    for seg in self._plot_2_args(remaining, **kwargs):
  File "/usr/lib/python2.6/site-packages/matplotlib/axes.py", line 331, in
_plot_2_args
    func(x, y)
  File "/usr/lib/python2.6/site-packages/matplotlib/axes.py", line 314, in
makefill
    (x[:,np.newaxis],y[:,np.newaxis])),
TypeError: list indices must be integers, not tuple
$ rpm -q python-matplotlib
python-matplotlib-0.98.5.2-1.3

-- 
By ZeD