matplotlib-users Mailing List for matplotlib

On Mon, Jul 8, 2013 at 4:36 PM, Benjamin Root <ben...@ou...> wrote:
> This looks very similar to something Tony Yu made once.  Check out this
> thread:
>
> http://matplotlib.1069221.n5.nabble.com/Combination-of-a-box-plot-and-a-histogram-td25313.html#a25319
>
> Note that Tony now has some sort of toolbox available that may have some of
> this implemented easily.

and some variations on the violins made it into statsmodels

http://statsmodels.sourceforge.net/devel/generated/statsmodels.graphics.boxplots.violinplot.html
http://statsmodels.sourceforge.net/devel/generated/statsmodels.graphics.boxplots.beanplot.html

Josef

>
> Cheers!
> Ben Root
>
>
>
> On Mon, Jul 8, 2013 at 4:19 PM, Carlos Becker <car...@gm...>
> wrote:
>>
>> Hi all,
>>
>> I was wondering if there is something similar to this
>> http://reference.wolfram.com/mathematica/ref/DistributionChart.html in
>> matplotlib, or if someone can give me a hint on how to implement it, since
>> it is very interesting for density plots.
>>
>> As far as I saw, there are histogram plots already available in
>> matplotlib, but I haven't seen something close to what is shown in the link
>> above.
>>
>> Thanks,
>>
>> ------------------------------------------
>> Carlos
>>
>>
>> ------------------------------------------------------------------------------
>> This SF.net email is sponsored by Windows:
>>
>> Build for Windows Store.
>>
>> http://p.sf.net/sfu/windows-dev2dev
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
>
> Build for Windows Store.
>
> http://p.sf.net/sfu/windows-dev2dev
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

This looks very similar to something Tony Yu made once.  Check out this
thread:

http://matplotlib.1069221.n5.nabble.com/Combination-of-a-box-plot-and-a-histogram-td25313.html#a25319

Note that Tony now has some sort of toolbox available that may have some of
this implemented easily.

Cheers!
Ben Root



On Mon, Jul 8, 2013 at 4:19 PM, Carlos Becker <car...@gm...>wrote:

> Hi all,
>
> I was wondering if there is something similar to this
> http://reference.wolfram.com/mathematica/ref/DistributionChart.html in
> matplotlib, or if someone can give me a hint on how to implement it, since
> it is very interesting for density plots.
>
> As far as I saw, there are histogram plots already available in
> matplotlib, but I haven't seen something close to what is shown in the link
> above.
>
> Thanks,
>
> ------------------------------------------
> Carlos
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Windows:
>
> Build for Windows Store.
>
> http://p.sf.net/sfu/windows-dev2dev
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>

Hi all,

I was wondering if there is something similar to this
http://reference.wolfram.com/mathematica/ref/DistributionChart.html in
matplotlib, or if someone can give me a hint on how to implement it, since
it is very interesting for density plots.

As far as I saw, there are histogram plots already available in matplotlib,
but I haven't seen something close to what is shown in the link above.

Thanks,

------------------------------------------
Carlos

Nicolas Mailhot wrote:
> Hi,
> 
> In matplotlib 1.2.1, the  get_name function is not garding against none
> self (unlike other functions); Unfortunately it seems I have a workload
> that makes matplotlib call get_name with None (wasn't the case in 1.2.0).
> I couldn't isolate the exact trigger, when I reduce the volume of data
> processed the problem goes away so I have to simple shareable reproducer.
> 
> Anyway, the following patch makes it all work for me, could it (or
> something similar) be merged?
> 
> diff -uNr matplotlib-1.2.1.orig/lib/matplotlib/font_manager.py
> matplotlib-1.2.1/lib/matplotlib/font_manager.py
> --- matplotlib-1.2.1.orig/lib/matplotlib/font_manager.py        2013-03-26
> 14:04:37.000000000 +0100
> +++ matplotlib-1.2.1/lib/matplotlib/font_manager.py     2013-07-08
> 14:49:37.791845661 +0200
> @@ -721,6 +721,8 @@
>          Return the name of the font that best matches the font
>          properties.
>          """
> +        if self._family is None:
> +          return rcParams['font.family']
>          return ft2font.FT2Font(str(findfont(self))).family_name
> 
>      def get_style(self):
> 
> Regards,
> 

Could you instead just test for "if not self._family"? Tests for equality
are more expensive (that means self._family == 0 or self._family == False will
also trigger your return).

Trying to install matplotlib-1.2.1 from source, i.e.,

[macpflm:~/matplotlib-1.2.1] pflm% python3 setup.py install

Everything seems to run ok, but I cannot import macplotlib:

[macpflm:~] pflm% python3
Python 3.3.1 (v3.3.1:d9893d13c628, Apr  6 2013, 11:07:11) 
[GCC 4.2.1 (Apple Inc. build 5666) (dot 3)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import matplotlib.pyplot
Python(16295) malloc: *** error for object 0x101c86820: pointer being freed
was not allocated
*** set a breakpoint in malloc_error_break to debug
Abort


 I tried updating libpng and freetype using macports and re-building, it did
not solve the problem.

There are some warnings printed during the build, e.g., 

ld: warning: in /opt/local/lib/libfreetype.dylib, file was built for
unsupported file format which is not the architecture being linked (i386)
ld: warning: in /opt/local/lib/libz.dylib, file was built for unsupported
file format which is not the architecture being linked (i386)
ld: warning: in
/Developer/SDKs/MacOSX10.6.sdk/usr/local/lib/libstdc++.dylib, file was built
for unsupported file format which is not the architecture being linked
(i386)

..and..

/Library/Frameworks/Python.framework/Versions/3.3/lib/python3.3/site-packages/numpy/core/include/numpy/npy_deprecated_api.h:11:2:
warning: #warning "Using deprecated NumPy API, disable it by #defining
NPY_NO_DEPRECATED_API NPY_1_7_API_VERSION"

(both warning several times).

Some details from the start of the screen output during the build.
basedirlist is: ['/usr/local/', '/usr', '/usr/X11', '/opt/local']
============================================================================
BUILDING MATPLOTLIB
            matplotlib: 1.2.1
                python: 3.3.1 (v3.3.1:d9893d13c628, Apr  6 2013, 11:07:11)
                        [GCC 4.2.1 (Apple Inc. build 5666) (dot 3)]
              platform: darwin

REQUIRED DEPENDENCIES
                 numpy: 1.7.0
             freetype2: 16.2.10

OPTIONAL BACKEND DEPENDENCIES
                libpng: 1.5.16
               Tkinter: Tkinter: version not identified, Tk: 8.5, Tcl: 8.5
                  Gtk+: no
                        * Building for Gtk+ requires pygtk; you must be able
                        * to "import gtk" in your build/install environment
       Mac OS X native: yes
                    Qt: no
                   Qt4: no
                PySide: no
                 Cairo: no

OPTIONAL DATE/TIMEZONE DEPENDENCIES
              dateutil: matplotlib will provide
                  pytz: matplotlib will provide
                   six: matplotlib will provide

OPTIONAL USETEX DEPENDENCIES
                dvipng: 1.5
           ghostscript: 6.01
                 latex: 3.141592

[Edit setup.cfg to suppress the above messages]
============================================================================

 Help!





--
View this message in context: http://matplotlib.1069221.n5.nabble.com/Install-problems-OSX-10-6-8-Python-3-3-1-tp41430.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Hi,

In matplotlib 1.2.1, the  get_name function is not garding against none
self (unlike other functions); Unfortunately it seems I have a workload
that makes matplotlib call get_name with None (wasn't the case in 1.2.0).
I couldn't isolate the exact trigger, when I reduce the volume of data
processed the problem goes away so I have to simple shareable reproducer.

Anyway, the following patch makes it all work for me, could it (or
something similar) be merged?

diff -uNr matplotlib-1.2.1.orig/lib/matplotlib/font_manager.py
matplotlib-1.2.1/lib/matplotlib/font_manager.py
--- matplotlib-1.2.1.orig/lib/matplotlib/font_manager.py        2013-03-26
14:04:37.000000000 +0100
+++ matplotlib-1.2.1/lib/matplotlib/font_manager.py     2013-07-08
14:49:37.791845661 +0200
@@ -721,6 +721,8 @@
         Return the name of the font that best matches the font
         properties.
         """
+        if self._family is None:
+          return rcParams['font.family']
         return ft2font.FT2Font(str(findfont(self))).family_name

     def get_style(self):

Regards,

-- 
Nicolas Mailhot

Hi,

Following an hardware failure I had to move some matplotlib scripts from
an ancient (but updated) Debian box to a clean RHEL 6 VM. As a result my
matplotlib version changed from 1.2.0 (as present in Debian instable) to
1.2.1 (some of the code wanted matplotlib ≥ 1.0, so I rebuilt an 1.2.1
packaged based on the Fedora 1.2.0 one. 1.3.0-rc2 / 1.3.0git gave me so
many dependencies griefs I gave up on it)

Anyway the result seems to work fine so far, except for a few scores of
time graphs where some tick labels gone missing. Those graphs are supposed
to show weekly data, with:
– ticks on month ends,
– subticks on week ends,
– week numbers in the middle of each week (top) and
– month numbers on the middle of each month (bottom).

With the new installation months numbers got missing.

Here is an extract of the relevant code :

%<---
  subplot.set_xticks([dd.monthstart(m) for m in months])
  subplot.set_xticks([w.gmticks() for w in weeks], minor=True)
  subplot.set_xticklabels([])
  plt.xlim(dd.monthstart(min(months)),dd.monthend(max(months)))

  subplot1 = subplot.twiny()
  mmonths = [m for i,m in enumerate(months)
                if operator.mod(i,1+(len(months)-1)/12) == 0]
  subplot1.set_xticks([dd.monthmiddle(m) for m in mmonths])
  subplot1.set_xticklabels(mmonths)
  subplot1.tick_params(labelsize='6.5',labelbottom=True, labeltop=False)
  plt.xlim(dd.monthstart(min(months)),dd.monthend(max(months)))

  subplot2 = subplot.twiny()
  mweeks = [w for i,w in enumerate(weeks)
                if operator.mod(i,1+(len(weeks)-1)/53) == 0]
  subplot2.set_xticks([dd.weekmiddle(w)   for w in mweeks])
  subplot2.set_xticklabels([w.iso_week[1] for w in mweeks])
  subplot2.tick_params(labelsize='5',labelbottom=False, labeltop=True,pad=3)
  plt.xlim(dd.monthstart(min(months)),dd.monthend(max(months)))

  subplot2.xaxis.set_ticks_position('none')
  subplot1.xaxis.set_ticks_position('none')
  subplot.xaxis.set_ticks_position('both')
%<---

I've tried to move the
subplot1.tick_params(labelsize='6.5',labelbottom=True, labeltop=False)
line to the end of the block, but it seems matplotlib is confusing
subplot1 and subplot2, and that results in the displaying of week numbers
(not month numbers) to the bottom of the plot (and the week numbers
disappear from the top of the graph).

Any idea how to restore the previously working behaviour ?

Regards,

-- 
Nicolas Mailhot