matplotlib-users Mailing List for matplotlib

On Thu, Jun 26, 2008 at 11:31 AM, Ben Axelrod <bax...@co...> wrote:
> I am new to MatPlotLib and I saw in the archives that 3d plotting is no
> longer supported in version 0.98. This seems like a major feature drop to
> me.
> I would really like to use this feature. Does anyone know what must be done
> to get this working again?
Michael did a significant rewrite of our transformation infrastructure
in 0.98. This was long overdue and helps us support 2D plotting
better -- eg user extensible projections -- and results in cleaner
code. None of the developers thus far have taken 3d plotting under
their wing to support it. The implementation, while extremely clever,
is slow and has some zorder problems, so we haven't fully embraced it,
though we have said for a long time we would like some minimum level
of 3D support that works out of the box.
Michael did a lot of work to make the transformation code mostly
compatible at the API level, and included some untested place holders
for higher dimensional data structures, but their are differences
which are detailed in the migration document in the source directory.
Since the 3d coding naturally uses a lot of transformations, someone
would have to do the work to port these calls to the new API, *and*
agree to support it going forward for us to reinclude it.
JDH

I am new to MatPlotLib and I saw in the archives that 3d plotting is no longer supported in version 0.98. This seems like a major feature drop to me.
I would really like to use this feature. Does anyone know what must be done to get this working again?
Thanks,
-Ben

Yip! I updated my code to "pylab.plot()" and so on.
Currently I use the following versions:
* the actual matplotlib 0.98 - the old 0.87.7 (installed via synaptics
had a memory leak)
* the stable openbabel release 2.1.1 (beta available...)
*numpy 1.1.0 (actual version; compiled from source)
*python 2.5.1
I am working on ubuntu linux 7.04 (feisty fawn) - and I do not use IDLE
or anything like that
Greets,
Flo
Massimo Sandal wrote:
>> Da: mat...@li... per conto di Florian Koelling
>> Inviato: gio 26/06/2008 14.04
>> A: Darren Dale; mat...@li...
>> Oggetto: Re: [Matplotlib-users] Bug: mad interference between matplotlib and openbabel
>> 
> 
> 
>> Thanks for your fast replies! As suggested I removed all "from import*"
>> statements. I used "import pylab" and "import pylab as p" statements but
>> it do s not work either. :-@
>> 
>
> What errors are you getting?
>
> Have you updated accordingly your calls to pylab functions? (e.g. did you update "plot()" with "pylab.plot()" when you used "import pylab")?
>
> m.
> 

On Thu, Jun 26, 2008 at 7:04 AM, Florian Koelling
<flo...@tu...> wrote:
> Thanks for your fast replies! As suggested I removed all "from import*"
> statements. I used "import pylab" and "import pylab as p" statements but
> it do s not work either. :-@
It was a good guess. You have a software version conflict apparently,
and yet you haven't provided us any version numbers of the packages
you are using -- that makes it pretty hard to progress. One thing I
notice is that you import Numeric, which is the predecessor to numpy.
All recent versions of matplotlib/pylab import numpy by default,
though some older versions can be configured to use Numeric, numarray
or numpy using a numerix setting. It is possible that you are seeing
a Numeric/numpy conflict somewhere. I am not sure what openbabel and
pybel are doing internally vis-a-vis numpy/numeric, though I do see
that there are some problems passing numpy scalars to openbabel at
http://openbabel.org/wiki/Python.
> Could you recommend other "easy to handle " python plotting libs?
chaco is quite nice - http://code.enthought.com/chaco.
But we would like to sort out this problem for future reference, so if
you'd be willing to stick with it and help us figure out where the
conflict is coming from, it will help the future users. My first
suggestion is to see if you need Numeric, or if numpy will suffice,
and make sure you are using it consistently throughout. My second
suggestion is to figure out of there are any gui toolkits in the
openbabel chain, and if so make sure your backend setting in
matplotlib is consistent with them (though the gui mainloop will not
be triggered by simply importing pylab, you must also call pylab.show,
so I don't know why that would cause a problem per se).
In any follow-up emails, please post the versions of all software you
are importing. One could way to get the matplotlib version info is to
create a simple script::
 import pylab
 pylab.plot([1,2,3])
 pylab.show()
and run it with::
 > python myscript.py --verbose-helpful
and post the output here.
JDH

Thanks for pointing this out. This is now fixed in SVN.
Cheers,
Mike
izak marais wrote:
> The vertical and horizontal hatch styles are swapped. According to 
> the docs:
> | - vertical
> - - horizontal
> but '|' gives horizontal and '-' vertical patterns. The reaming styles 
> appear to work correctly.
>
>
> ----- Original Message ----
> From: John Hunter <jd...@gm...>
> To: izak marais <iza...@ya...>
> Cc: mat...@li...
> Sent: Wednesday, June 25, 2008 3:49:56 PM
> Subject: Re: [Matplotlib-users] histogram hatching
>
> On Wed, Jun 25, 2008 at 6:05 AM, izak marais <iza...@ya... 
> <mailto:iza...@ya...>> wrote:
> > Hi,
> >
> > I see there is a "hatch: unknown" kwarg mentioned in the hist()
> > documentation. Can anyone shed some light on how to use this please? I
> > assume it might be used to generate monochrome rectangles with
> > differentiating hatched fills?
>
> The rectangles generates by "hist" are matplotlib.patch.Rectangle
> instances, and the "hatch" property is controlled by the "set_hatch"
> method. Here is the docstring which should tell you most everything
> you need to know. Note the caveat at the end that it is only
> currently supported on postscript -- patches for other backends much
> obliged
>
> def set_hatch(self, h):
> """
> Set the hatching pattern
>
> hatch can be one of::
>
> / - diagonal hatching
> \ - back diagonal
> | - vertical
> - - horizontal
> # - crossed
> x - crossed diagonal
>
> Letters can be combined, in which case all the specified
> hatchings are done. If same letter repeats, it increases the
> density of hatching in that direction.
>
> CURRENT LIMITATIONS:
>
> 1. Hatching is supported in the PostScript backend only.
>
> 2. Hatching is done with solid black lines of width 0.
>
>
>
> ------------------------------------------------------------------------
>
> -------------------------------------------------------------------------
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://sourceforge.net/services/buy/index.php
> ------------------------------------------------------------------------
>
> _______________________________________________
> 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

Thanks for your fast replies! As suggested I removed all "from import*"
statements. I used "import pylab" and "import pylab as p" statements but
it do s not work either. :-@
Could you recommend other "easy to handle " python plotting libs?
Thanx!
Darren Dale wrote:
>
>
> On Thu, Jun 26, 2008 at 5:51 AM, Florian Koelling
> <flo...@tu... <mailto:flo...@tu...>> wrote:
>
> Hi folks!
>
> I' m working on a code to identify ligand's information from pdb
> crystal
> structures. For this purpose I use the openbabel package. I' m
> encoding
> the ligand's information in smarts.
> For checking if the smarts are assigned correctly I 'm writing a sdf
> file (to control whether the features are placed correctly).
>
> This part works fine - but I get into deep trouble when I try to plot
> simple count statistics of my smarts, when I use the "from pylab
> import*" , or "from pylab import matplotlib" statements (trouble
> appears
> just after IMPORTING - I do not use a pylab function in this state of
> the program yet):
>
>
> Avoid all use of "from package import *". When you import *, strange
> things can happen, since one import can overwrite things from a
> previous import. Try "import pylab" or "import pylab as plt", that way
> all the pylab functionality remains in its own namespace and cant be
> overwritten, and can't overwrite anything else. Finally, "from pylab
> import matplotlib" won't work, matplotlib is a top-level package.

Hi Florian,
 you should try "import pylab" or "import pylab as P". "from pylab 
import *" might overwrite some functions or methods ... if a pylab 
function/method has the same name as one in a module imported with * before.
See the following example:
#file a.py
def foo():
 print "I'm a"
# file b.py
def foo():
 print "I'm b"
# file import_ab.py
from a import *
foo()
from b import *
foo()
The output is:
I'm a
I'm b
as the function foo() is overwritten by the second import...
Manuel
Florian Koelling wrote:
> Hi folks!
> 
> I' m working on a code to identify ligand's information from pdb crystal
> structures. For this purpose I use the openbabel package. I' m encoding
> the ligand's information in smarts.
> For checking if the smarts are assigned correctly I 'm writing a sdf
> file (to control whether the features are placed correctly).
> 
> This part works fine - but I get into deep trouble when I try to plot
> simple count statistics of my smarts, when I use the "from pylab
> import*" , or "from pylab import matplotlib" statements (trouble appears
> just after IMPORTING - I do not use a pylab function in this state of
> the program yet):
> 
> Smart features are not recognized anymore, the feature coordinates are 
> shifted and so on - But I don' t know why.
> 
> my code snip:
> 
> 
> 
> 
> import os
> from numpy import*
> from time import*
> 
> #### SMARTING
> import numpy
> 
> import openbabel
> from openbabel import*
> import pybel
> from pybel import* 
> 
> #### LIGAND DETECTION
> import Bio.PDB
> from Bio.PDB import*
> import Numeric
> from Bio.PDB.StructureBuilder import *
> 
> 
> #from pylab import* --> RAISES THE PROBLEMS !!!!
> 
> 
> def projecting_smarts(isolated_ligand, method ,color, output):
> 
> 
> #### Bio PDB -> 2 get ligand coords
> parser =PDBParser()
> 
> structure = parser.get_structure('s', isolated_ligand)
> atom = structure.get_atoms() 
> ###
> 
> 
> print "smarting successfully started!"
> 
> mol = readfile("pdb", isolated_ligand).next()
> 
> 
> smart_list = [] 
> 
> result_acceptor = method.findall(mol) #finds acceptors
> 
> 
> for i in result_acceptor: #Result
> Donor= list from tuples
> list_element1 = (list(i)[0]) # -> 
> resolve hierarchy 
> smart_list.append(list_element1)
> 
> print smart_list,"ACCEPTOR_LIST" #list of smart features
> 
> 
> 
> 
> Maybe one of you might know what to do!
> 
> Thanx alot!
> 
> flo
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> -------------------------------------------------------------------------
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://sourceforge.net/services/buy/index.php
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Hi folks!
I' m working on a code to identify ligand's information from pdb crystal
structures. For this purpose I use the openbabel package. I' m encoding
the ligand's information in smarts.
For checking if the smarts are assigned correctly I 'm writing a sdf
file (to control whether the features are placed correctly).
This part works fine - but I get into deep trouble when I try to plot
simple count statistics of my smarts, when I use the "from pylab
import*" , or "from pylab import matplotlib" statements (trouble appears
just after IMPORTING - I do not use a pylab function in this state of
the program yet):
 
Smart features are not recognized anymore, the feature coordinates are 
shifted and so on - But I don' t know why.
my code snip:
import os
from numpy import*
from time import*
#### SMARTING
import numpy
import openbabel
from openbabel import*
import pybel
from pybel import* 
#### LIGAND DETECTION
import Bio.PDB
from Bio.PDB import*
import Numeric
from Bio.PDB.StructureBuilder import *
#from pylab import* --> RAISES THE PROBLEMS !!!!
 
def projecting_smarts(isolated_ligand, method ,color, output):
 
 
#### Bio PDB -> 2 get ligand coords
 parser =PDBParser()
 
 structure = parser.get_structure('s', isolated_ligand)
 atom = structure.get_atoms() 
###
 print "smarting successfully started!"
 
 mol = readfile("pdb", isolated_ligand).next()
 
 smart_list = [] 
 
 result_acceptor = method.findall(mol) #finds acceptors
 
 
 for i in result_acceptor: #Result
Donor= list from tuples
 list_element1 = (list(i)[0]) # -> 
resolve hierarchy 
 smart_list.append(list_element1)
 
 print smart_list,"ACCEPTOR_LIST" #list of smart features
Maybe one of you might know what to do!
Thanx alot!
flo

The vertical and horizontal hatch styles are swapped. According to the docs:
 | - vertical
 - - horizontal
but '|' gives horizontal and '-' vertical patterns. The reaming styles appear to work correctly.
----- Original Message ----
From: John Hunter <jd...@gm...>
To: izak marais <iza...@ya...>
Cc: mat...@li...
Sent: Wednesday, June 25, 2008 3:49:56 PM
Subject: Re: [Matplotlib-users] histogram hatching
On Wed, Jun 25, 2008 at 6:05 AM, izak marais <iza...@ya...> wrote:
> Hi,
>
> I see there is a "hatch: unknown" kwarg mentioned in the hist()
> documentation. Can anyone shed some light on how to use this please? I
> assume it might be used to generate monochrome rectangles with
> differentiating hatched fills?
The rectangles generates by "hist" are matplotlib.patch.Rectangle
instances, and the "hatch" property is controlled by the "set_hatch"
method. Here is the docstring which should tell you most everything
you need to know. Note the caveat at the end that it is only
currently supported on postscript -- patches for other backends much
obliged
 def set_hatch(self, h):
 """
 Set the hatching pattern
 hatch can be one of::
 / - diagonal hatching
 \ - back diagonal
 | - vertical
 - - horizontal
 # - crossed
 x - crossed diagonal
 Letters can be combined, in which case all the specified
 hatchings are done. If same letter repeats, it increases the
 density of hatching in that direction.
 CURRENT LIMITATIONS:
 1. Hatching is supported in the PostScript backend only.
 2. Hatching is done with solid black lines of width 0.
 

On 26-Jun-08, at 4:13 AM, David Warde-Farley wrote:
> Hi all,
>
> I'm making multiple calls to the plot() command, actually plotting
> parallel horizontal lines, and using a consistent 'linewidth'
> parameter, but for some reason I get different line widths each time
> (or the same for a couple). I assumed it was some sort of wonky
> behaviour caused by antialiasing, but aa=False doesn't help. Can
> anyone tell me why this might be happening?
It was my own fault, I had set ylim to be the location of the first 
and last lines respectively and thus the thicker lines were getting 
clipped. Please disregard the last message. :)
David

Hi all,
I'm making multiple calls to the plot() command, actually plotting 
parallel horizontal lines, and using a consistent 'linewidth' 
parameter, but for some reason I get different line widths each time 
(or the same for a couple). I assumed it was some sort of wonky 
behaviour caused by antialiasing, but aa=False doesn't help. Can 
anyone tell me why this might be happening?
David

When I draw a Line2D instance (using plot()), I have set up an interactive
way of clicking on one of the points to return the x-coordinate of the
point. There might be a simpler way to do it, but the way I figured out is:
def on_pick(self, event): # called when clicking on a Line2D point
 return [event.artist.get_xdata()[pt].data[0] for pt in
event.artist.contains(event.mouseevent)[1].values()]
my basic interpretation is that I get the point index in the line and then
get the xcoord for that point.
It would be nice to get the xcoord for a scatter point as well, but I
haven't had luck figuring that one out. I can do
event.artist.contains(event.mouseevent) to get the index for the
scatterpoint, but apparently there's no get_xdata() tied to the
regularpolygon instance.
paul

Thank you very much for the info. 
As an aside, after saving it one can use epstopdf to convert the crosshatched eps figure to pdf. If you want to edit it, inkscape can import pdf and save to svg.
----- Original Message ----
From: John Hunter <jd...@gm...>
To: izak marais <iza...@ya...>
Cc: mat...@li...
Sent: Wednesday, June 25, 2008 3:49:56 PM
Subject: Re: [Matplotlib-users] histogram hatching
On Wed, Jun 25, 2008 at 6:05 AM, izak marais <iza...@ya...> wrote:
> Hi,
>
> I see there is a "hatch: unknown" kwarg mentioned in the hist()
> documentation. Can anyone shed some light on how to use this please? I
> assume it might be used to generate monochrome rectangles with
> differentiating hatched fills?
The rectangles generates by "hist" are matplotlib.patch.Rectangle
instances, and the "hatch" property is controlled by the "set_hatch"
method. Here is the docstring which should tell you most everything
you need to know. Note the caveat at the end that it is only
currently supported on postscript -- patches for other backends much
obliged
 def set_hatch(self, h):
 """
 Set the hatching pattern
 hatch can be one of::
 / - diagonal hatching
 \ - back diagonal
 | - vertical
 - - horizontal
 # - crossed
 x - crossed diagonal
 Letters can be combined, in which case all the specified
 hatchings are done. If same letter repeats, it increases the
 density of hatching in that direction.
 CURRENT LIMITATIONS:
 1. Hatching is supported in the PostScript backend only.
 2. Hatching is done with solid black lines of width 0.
 

Hi All,
I am using the function vlines() to create a number of lines at 
sampling points below my sinewave (top right graph)
ie.
 tt = axUpper2.vlines(t[xi],result[0], result[xi] )
were t is the time along x axis, and xi is the sample points and 
'result' is the sine
This returns a "matplotlib.collections.LineCollection"
print tt
<matplotlib.collections.LineCollection instance at 0x01FA9E68>
I also have a slider for changing some input parameters, which will 
affect both amplitude and offset of 'result', .. however once things 
change I cannot work out how to change the data in my line collection. I 
need to get rid of the old lines, and create a new set or update the 
values in the current lines based on the new "result".
It's not obvious to me how I should go about it.
my first instinct was to try and iterate on the linecollection
 for l in tt:
 l.set_ydata(result[xi])
"TypeError: iteration over non-sequence"
I found out a collection is not iterable
I can call vlines() on each change, ..but I still don't know how to get 
rid of the old lines.
Can someone give me a pointer to functions I should look into, or make 
suggestion how to work this another way?.
Thanks
Steve

> I think I have this fixed in svn -- also, I rewrote your example to
> use csv2rec (which returns record arrays). Thought you might be
> interested:
>
> import matplotlib.mlab as mlab
> import matplotlib.pyplot as plt
> import matplotlib.dates as mdates
> import matplotlib.cbook as cbook
> r1 = mlab.csv2rec('test1.dat', names='date,value', delimiter=' ',
> converterd={'date':cbook.todate('%Y%m%d')})
> r2 = mlab.csv2rec('test2.dat', names='date,value', delimiter=' ',
> converterd={'date':cbook.todate('%Y%m%d')})
>
> print 'dtype', r1.dtype
> fig = plt.figure()
> ax1 = fig.add_subplot(111)
> ax2 = ax1.twinx()
>
> # produce plot
> line1, = ax1.plot(r1.date, r1.value, 'bo-')
> line2, = ax2.plot(r2.date, r2.value, 'bo-')
>
> # set up axes
> ax1.xaxis.set_major_locator(mdates.DayLocator(range(0, 31, 2)))
> ax1.xaxis.set_minor_locator(mdates.DayLocator(range(0, 31, 1)))
> ax1.xaxis.set_major_formatter(mdates.DateFormatter('%Y/%m/%d'))
>
> plt.show()
Thanks John, although the scale on the x-axis isn't shown using the
above example.
Cheers
Adam