matplotlib basic question

[orangeDinosaur]

Hi,

I am exploring the possibility of using python as a replacement of
MATLAB when I leave school. So, I've been playing with matplotlib and
have run into some weird behavior after recently installing python
2.5.1 and matplotlib 0.90 on my Windows XP machine. Here's an example
of what I see:

from matplotlib.pylab import *
x=arange(-2*pi,2*pi,pi/24)
y=sin(x)
plot(x,y)

show()

So, first off, what's up with the [<matplotlib.lines.Line2D instance
at 0x017C38C8>] line that shows up after my plot command? And second,
when I call show(), a new figure pops up with my sin wave -- seems all
right, yes? But I'm not given another >>> prompt in IDLE until or
unless I close the figure that popped up with the show() call.

So, after closing the figure I type this:

and this time another figure pops up with my sine wave again and I get
a prompt as well. But now, the figure window is completely
unresponsive -- I can't even close it without getting the "your
program is not repsonding" business. What am I missing? This
behavior so far seems pretty unintuitive.

Any clarification is appreciated!

trevis

 

[cfriedalek]

So, first off, what's up with the [<matplotlib.lines.Line2D instance
at 0x017C38C8>] line that shows up after my plot command? And second,
when I call show(), a new figure pops up with my sin wave -- seems all
right, yes? But I'm not given another >>> prompt in IDLE until or
unless I close the figure that popped up with the show() call.

This may not be strictly correct but thats a reference to the plot
instance which is subsequently passed to show().
If you don't want to see it put a semicolon at the end of your command
eg. plot(range(5));

The issue with IDLE is to due with lack of connection between the
python interpreter event loop and the event loop of matplotlib (I
think). For me the solution was to install the ipython shell (http://
ipython.scipy.org/moin/). If you run ipython with the -pylab flag then
you can matplotlib interactively. For example

plot(range(5),'bo')
show()
clf()
plot([1,2,3,4,5],range(0,10,2),'r-')
(no need for another show() command since the graphics display is
already visible ... unless of course I deleted it in between plot
commands)

hth

 

[Rob Clewley]

So, first off, what's up with the [<matplotlib.lines.Line2D instance
at 0x017C38C8>] line that shows up after my plot command? And second,
when I call show(), a new figure pops up with my sin wave -- seems all
right, yes? But I'm not given another >>> prompt in IDLE until or
unless I close the figure that popped up with the show() call.

This may not be strictly correct but thats a reference to the plot
instance which is subsequently passed to show().
If you don't want to see it put a semicolon at the end of your command
eg. plot(range(5));

My understanding is that the semicolon trick is specific to IPython,
and does not work in IDLE. I don't know about in other environments.
You can just assign the list of returned plot objects to a variable to
make it invisible at the command line although these return values
(like all return values) are not "shown in the command window" (i.e.
passed to stdout) if they are returned from calls made in a script.

The issue with IDLE is to due with lack of connection between the
python interpreter event loop and the event loop of matplotlib (I
think). For me the solution was to install the ipython shell (http://
ipython.scipy.org/moin/). If you run ipython with the -pylab flag then
you can matplotlib interactively. For example

plot(range(5),'bo')
show()
clf()
plot([1,2,3,4,5],range(0,10,2),'r-')
(no need for another show() command since the graphics display is
already visible ... unless of course I deleted it in between plot
commands)

The OP appears to be using IDLE, for which the things to try are:

In site-packages/matplotlib/backends/backend_tkagg.py uncomment the
line #os.environ['PYTHONINSPECT'] = '1'

Set interactive=True in share/matplotlib/.matplotlibrc

Start IDLE with the -n flag

In site-packages/matplotlib/backends/backend_tkagg.py comment out the line

Tk.mainloop()

in the function "show"

-Rob

 

[Colin J. Williams]

Hi,

I am exploring the possibility of using python as a replacement of
MATLAB when I leave school. So, I've been playing with matplotlib and
have run into some weird behavior after recently installing python
2.5.1 and matplotlib 0.90 on my Windows XP machine. Here's an example
of what I see:

from matplotlib.pylab import *
x=arange(-2*pi,2*pi,pi/24)
y=sin(x)
plot(x,y)

show()

So, first off, what's up with the [<matplotlib.lines.Line2D instance
at 0x017C38C8>] line that shows up after my plot command? And second,
when I call show(), a new figure pops up with my sin wave -- seems all
right, yes? But I'm not given another >>> prompt in IDLE until or
unless I close the figure that popped up with the show() call.

So, after closing the figure I type this:

and this time another figure pops up with my sine wave again and I get
a prompt as well. But now, the figure window is completely
unresponsive -- I can't even close it without getting the "your
program is not repsonding" business. What am I missing? This
behavior so far seems pretty unintuitive.

Any clarification is appreciated!

trevis

Trevis,

You might look at Numpy, which deals with multi-dimensional arrays.

It has a small matrix component and some progress has been made with
MATLAB amenable problems.

Colin W.

 

[Pete Forman]

> [...] But now, the figure window is completely unresponsive -- I
> can't even close it without getting the "your program is not
> repsonding" business. What am I missing? This behavior so far
> seems pretty unintuitive.

The best way out of this is to use IPython. It also needs a backend
which remains responsive, WxAgg works but Tk did not last time I
tried. IPython 0.8.1 is a release candidate which fixes some Windows
issues in 0.8.0. If you want a stable package that has all the parts
present out of the box then look at Enthought.

 

[orangeDinosaur]

OK, I'll go with the enthought installation. This seems to be the
path of least resistance. For those of you who have been in my
position, is there a reason NOT to go with the enthought installation
and do things piecemeal instead?

thanks,
trevis

[...] But now, the figure window is completely unresponsive -- I
can't even close it without getting the "your program is not
repsonding" business. What am I missing? This behavior so far
seems pretty unintuitive.

The best way out of this is to use IPython. It also needs a backend
which remains responsive, WxAgg works but Tk did not last time I
tried. IPython 0.8.1 is a release candidate which fixes some Windows
issues in 0.8.0. If you want a stable package that has all the parts
present out of the box then look at Enthought.

 

[Colin J. Williams]

OK, I'll go with the enthought installation. This seems to be the
path of least resistance. For those of you who have been in my
position, is there a reason NOT to go with the enthought installation
and do things piecemeal instead?

thanks,
trevis

[...] But now, the figure window is completely unresponsive -- I
can't even close it without getting the "your program is not
repsonding" business. What am I missing? This behavior so far
seems pretty unintuitive.

The best way out of this is to use IPython. It also needs a backend
which remains responsive, WxAgg works but Tk did not last time I
tried. IPython 0.8.1 is a release candidate which fixes some Windows
issues in 0.8.0. If you want a stable package that has all the parts
present out of the box then look at Enthought.

It's rather heavy and may include stuff you don't need.

I'm not sure that scipy has been updated to Python 2.5

Colin W.

 

[Robert Kern]

I'm not sure that scipy has been updated to Python 2.5

? scipy certainly works with 2.5. Are you referring to something else perhaps?

--
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
that is made terrible by our own mad attempt to interpret it as though it had
an underlying truth."
-- Umberto Eco

 

[Tommy Grav]

? scipy certainly works with 2.5. Are you referring to something
else perhaps?

A side question: Is there any plans of updating the scipy.org
Superpack bundle
for Mac OS X to work with 2.5?

Cheers
Tommy

 

[Robert Kern]

A side question: Is there any plans of updating the scipy.org
Superpack bundle
for Mac OS X to work with 2.5?

You would have to ask Chris Fonnesbeck. It's not an "official" binary inasmuch
as scipy has official binaries.

--
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
that is made terrible by our own mad attempt to interpret it as though it had
an underlying truth."
-- Umberto Eco

 

[Tommy Grav]

You would have to ask Chris Fonnesbeck. It's not an "official"
binary inasmuch
as scipy has official binaries.

Okidoki. This is the last reason for me not upgrading to 2.5 as I
have a hard
time compiling things from scratch

Cheers
Tommy

 

[Pete Forman]

> ? scipy certainly works with 2.5. Are you referring to something
> else perhaps?

Yes, the Python Enthought Edition was being discussed and it is
currently based on Python 2.4.3.

http://code.enthought.com/enthon/

 

[Robert Kern]

Yes, the Python Enthought Edition was being discussed and it is
currently based on Python 2.4.3.

http://code.enthought.com/enthon/

I'm quite familiar with Python Enthought Edition as I work at Enthought. I was
confused because Colin referenced scipy specifically rather than simply saying
that the current version of Python Enthought Edition was not based on 2.5.

--
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
that is made terrible by our own mad attempt to interpret it as though it had
an underlying truth."
-- Umberto Eco