ANN: matplotlib-0.87.4 (bugfix release for enthon)

···

On 7/7/06, John Hunter <jdhunter@...4...> wrote:

We'd like to do a bugfix release for the next release of enthought
python, which will include the latest mpl. Apparently, there is a
problem with 0.87.3 and numpy which has been fixed in svn.

If there is anything we should wait on, let us know, otherwise we'll
probably try to roll out 0.87.4 early next week.

Thanks,
JDH

===============================================================
2006-07-11 Released 0.87.4 at revision 2558

2006-07-07 Fixed a usetex bug with older versions of latex - DSD

2006-07-07 Add compatibility for NumPy 1.0 - TEO

2006-06-29 Added a Qt4Agg backend. Thank you James Amundson - DSD

2006-06-26 Fixed a usetex bug. On windows, usetex will prcess
           postscript output in the current directory rather than
           in a temp directory. This is due to the use of spaces
           and tildes in windows paths, which cause problems with
           latex. The subprocess module is no longer used. - DSD

2006-06-22 Various changes to bar(), barh(), and hist().
           Added 'edgecolor' keyword arg to bar() and barh().
           The x and y args in barh() have been renamed to width
           and bottom respectively, and their order has been swapped
           to maintain a (position, value) order ala matlab. left,
           height, width and bottom args can now all be scalars or
           sequences. barh() now defaults to edge alignment instead
           of center alignment. Added a keyword arg 'align' to bar(),
           barh() and hist() that controls between edge or center bar
           alignment. Fixed ignoring the rcParams['patch.facecolor']
           for bar color in bar() and barh(). Fixed ignoring the
           rcParams['lines.color'] for error bar color in bar()
           and barh(). Fixed a bug where patches would be cleared
           when error bars were plotted if rcParams['axes.hold']
           was False. - MAS

2006-06-22 Added support for numerix 2-D arrays as alternatives to
           a sequence of (x,y) tuples for specifying paths in
           collections, quiver, contour, pcolor, transforms.
           Fixed contour bug involving setting limits for
           color mapping. Added numpy-style all() to numerix. - EF

2006-06-20 Added custom FigureClass hook to pylab interface - see
           examples/custom_figure_class.py

2006-06-16 Added colormaps from gist (gist_earth, gist_stern,
           gist_rainbow, gist_gray, gist_yarg, gist_heat, gist_ncar) - JW

2006-06-16 Added a pointer to parent in figure canvas so you can
           access the container with fig.canvas.manager. Useful if
           you want to set the window title, eg in gtk
           fig.canvas.manager.window.set_title, though a GUI neutral
           method would be preferable JDH

2006-06-16 Fixed colorbar.py to handle indexed colors (i.e.,
           norm = no_norm()) by centering each colored region
           on its index. - EF

2006-06-15 Added scalex and scaley to Axes.autoscale_view to support
           selective autoscaling just the x or y axis, and supported
           these command in plot so you can say plot(something,
           scaley=False) and just the x axis will be autoscaled.
           Modified axvline and axhline to support this, so for
           example axvline will no longer autoscale the y axis. JDH

2006-06-13 Fix so numpy updates are backward compatible - TEO

2006-06-12 Updated numerix to handle numpy restructuring of
           oldnumeric - TEO

2006-06-12 Updated numerix.fft to handle numpy restructuring
           Added ImportError to numerix.linear_algebra for numpy -TEO

2006-06-11 Added quiverkey command to pylab and Axes, using
           QuiverKey class in quiver.py. Changed pylab and Axes
           to use quiver2 if possible, but drop back to the
           newly-renamed quiver_classic if necessary. Modified
           examples/quiver_demo.py to illustrate the new quiver
           and quiverkey. Changed LineCollection implementation
           slightly to improve compatibility with PolyCollection. - EF

2006-06-11 Fixed a usetex bug for windows, running latex on files
           with spaces in their names or paths was failing - DSD

2006-06-09 Made additions to numerix, changes to quiver to make it
           work with all numeric flavors. - EF

2006-06-09 Added quiver2 function to pylab and method to axes,
           with implementation via a Quiver class in quiver.py.
           quiver2 will replace quiver before the next release;
           it is placed alongside it initially to facilitate
           testing and transition. See also
           examples/quiver2_demo.py. - EF

2006-06-08 Minor bug fix to make ticker.py draw proper minus signs
           with usetex - DSD

what about the mplot3d. (http://www.scipy.org/Cookbook/Matplotlib/mplot3D)
it works nice under matplotlib0.86.2
but it can't work under matplotlib 0.87.3(with numpy or numeric).

···

On 7/12/06, Charlie Moad <cwmoad@...287...> wrote:

On 7/7/06, John Hunter <jdhunter@...4...> wrote:
>
> We'd like to do a bugfix release for the next release of enthought
> python, which will include the latest mpl. Apparently, there is a
> problem with 0.87.3 and numpy which has been fixed in svn.
>
> If there is anything we should wait on, let us know, otherwise we'll
> probably try to roll out 0.87.4 early next week.
>
> Thanks,
> JDH

matplotlib · PyPI
matplotlib - Browse Files at SourceForge.net

===============================================================
2006-07-11 Released 0.87.4 at revision 2558

2006-07-07 Fixed a usetex bug with older versions of latex - DSD

2006-07-07 Add compatibility for NumPy 1.0 - TEO

2006-06-29 Added a Qt4Agg backend. Thank you James Amundson - DSD

2006-06-26 Fixed a usetex bug. On windows, usetex will prcess
          postscript output in the current directory rather than
          in a temp directory. This is due to the use of spaces
          and tildes in windows paths, which cause problems with
          latex. The subprocess module is no longer used. - DSD

2006-06-22 Various changes to bar(), barh(), and hist().
          Added 'edgecolor' keyword arg to bar() and barh().
          The x and y args in barh() have been renamed to width
          and bottom respectively, and their order has been swapped
          to maintain a (position, value) order ala matlab. left,
          height, width and bottom args can now all be scalars or
          sequences. barh() now defaults to edge alignment instead
          of center alignment. Added a keyword arg 'align' to bar(),
          barh() and hist() that controls between edge or center bar
          alignment. Fixed ignoring the rcParams['patch.facecolor']
          for bar color in bar() and barh(). Fixed ignoring the
          rcParams['lines.color'] for error bar color in bar()
          and barh(). Fixed a bug where patches would be cleared
          when error bars were plotted if rcParams['axes.hold']
          was False. - MAS

2006-06-22 Added support for numerix 2-D arrays as alternatives to
          a sequence of (x,y) tuples for specifying paths in
          collections, quiver, contour, pcolor, transforms.
          Fixed contour bug involving setting limits for
          color mapping. Added numpy-style all() to numerix. - EF

2006-06-20 Added custom FigureClass hook to pylab interface - see
          examples/custom_figure_class.py

2006-06-16 Added colormaps from gist (gist_earth, gist_stern,
          gist_rainbow, gist_gray, gist_yarg, gist_heat, gist_ncar) - JW

2006-06-16 Added a pointer to parent in figure canvas so you can
          access the container with fig.canvas.manager. Useful if
          you want to set the window title, eg in gtk
          fig.canvas.manager.window.set_title, though a GUI neutral
          method would be preferable JDH

2006-06-16 Fixed colorbar.py to handle indexed colors (i.e.,
          norm = no_norm()) by centering each colored region
          on its index. - EF

2006-06-15 Added scalex and scaley to Axes.autoscale_view to support
          selective autoscaling just the x or y axis, and supported
          these command in plot so you can say plot(something,
          scaley=False) and just the x axis will be autoscaled.
          Modified axvline and axhline to support this, so for
          example axvline will no longer autoscale the y axis. JDH

2006-06-13 Fix so numpy updates are backward compatible - TEO

2006-06-12 Updated numerix to handle numpy restructuring of
          oldnumeric - TEO

2006-06-12 Updated numerix.fft to handle numpy restructuring
          Added ImportError to numerix.linear_algebra for numpy -TEO

2006-06-11 Added quiverkey command to pylab and Axes, using
          QuiverKey class in quiver.py. Changed pylab and Axes
          to use quiver2 if possible, but drop back to the
          newly-renamed quiver_classic if necessary. Modified
          examples/quiver_demo.py to illustrate the new quiver
          and quiverkey. Changed LineCollection implementation
          slightly to improve compatibility with PolyCollection. - EF

2006-06-11 Fixed a usetex bug for windows, running latex on files
          with spaces in their names or paths was failing - DSD

2006-06-09 Made additions to numerix, changes to quiver to make it
          work with all numeric flavors. - EF

2006-06-09 Added quiver2 function to pylab and method to axes,
          with implementation via a Quiver class in quiver.py.
          quiver2 will replace quiver before the next release;
          it is placed alongside it initially to facilitate
          testing and transition. See also
          examples/quiver2_demo.py. - EF

2006-06-08 Minor bug fix to make ticker.py draw proper minus signs
          with usetex - DSD

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
matplotlib-users List Signup and Options

It would be useful to have plot accept a 2-D array as well. Would
patches for this be considered, or is there some reason why this can't
work? At the moment, doing

P.plot(z) where z is Nx2,

raises

RuntimeError: xdata and ydata must be the same length

So it doesn't look as though this functionality will override any
current feature.

Regards
Stéfan

···

On Tue, Jul 11, 2006 at 10:53:59PM -0400, Charlie Moad wrote:

2006-06-22 Added support for numerix 2-D arrays as alternatives to
           a sequence of (x,y) tuples for specifying paths in
           collections, quiver, contour, pcolor, transforms.
           Fixed contour bug involving setting limits for
           color mapping. Added numpy-style all() to numerix. - EF

The modules have been imported into the matplotlib repository, and
they are semi-functional. They still need a lot of work though.
There is a new example, simple3d_oo.py.

- Charlie

···

On 7/11/06, imcs ee <imcsee@...287...> wrote:

what about the mplot3d. (http://www.scipy.org/Cookbook/Matplotlib/mplot3D)
it works nice under matplotlib0.86.2
but it can't work under matplotlib 0.87.3(with numpy or numeric).

On 7/12/06, Charlie Moad <cwmoad@...287...> wrote:
> On 7/7/06, John Hunter <jdhunter@...4...> wrote:
> >
> > We'd like to do a bugfix release for the next release of enthought
> > python, which will include the latest mpl. Apparently, there is a
> > problem with 0.87.3 and numpy which has been fixed in svn.
> >
> > If there is anything we should wait on, let us know, otherwise we'll
> > probably try to roll out 0.87.4 early next week.
> >
> > Thanks,
> > JDH
>
> matplotlib · PyPI
> matplotlib - Browse Files at SourceForge.net
>
> ===============================================================
> 2006-07-11 Released 0.87.4 at revision 2558
>
> 2006-07-07 Fixed a usetex bug with older versions of latex - DSD
>
> 2006-07-07 Add compatibility for NumPy 1.0 - TEO
>
> 2006-06-29 Added a Qt4Agg backend. Thank you James Amundson - DSD
>
> 2006-06-26 Fixed a usetex bug. On windows, usetex will prcess
> postscript output in the current directory rather than
> in a temp directory. This is due to the use of spaces
> and tildes in windows paths, which cause problems with
> latex. The subprocess module is no longer used. - DSD
>
> 2006-06-22 Various changes to bar(), barh(), and hist().
> Added 'edgecolor' keyword arg to bar() and barh().
> The x and y args in barh() have been renamed to width
> and bottom respectively, and their order has been swapped
> to maintain a (position, value) order ala matlab. left,
> height, width and bottom args can now all be scalars or
> sequences. barh() now defaults to edge alignment instead
> of center alignment. Added a keyword arg 'align' to bar(),
> barh() and hist() that controls between edge or center bar
> alignment. Fixed ignoring the rcParams['patch.facecolor']
> for bar color in bar() and barh(). Fixed ignoring the
> rcParams['lines.color'] for error bar color in bar()
> and barh(). Fixed a bug where patches would be cleared
> when error bars were plotted if rcParams['axes.hold']
> was False. - MAS
>
> 2006-06-22 Added support for numerix 2-D arrays as alternatives to
> a sequence of (x,y) tuples for specifying paths in
> collections, quiver, contour, pcolor, transforms.
> Fixed contour bug involving setting limits for
> color mapping. Added numpy-style all() to numerix. - EF
>
> 2006-06-20 Added custom FigureClass hook to pylab interface - see
> examples/custom_figure_class.py
>
> 2006-06-16 Added colormaps from gist (gist_earth, gist_stern,
> gist_rainbow, gist_gray, gist_yarg, gist_heat, gist_ncar) - JW
>
> 2006-06-16 Added a pointer to parent in figure canvas so you can
> access the container with fig.canvas.manager. Useful if
> you want to set the window title, eg in gtk
> fig.canvas.manager.window.set_title, though a GUI neutral
> method would be preferable JDH
>
> 2006-06-16 Fixed colorbar.py to handle indexed colors (i.e.,
> norm = no_norm()) by centering each colored region
> on its index. - EF
>
> 2006-06-15 Added scalex and scaley to Axes.autoscale_view to support
> selective autoscaling just the x or y axis, and supported
> these command in plot so you can say plot(something,
> scaley=False) and just the x axis will be autoscaled.
> Modified axvline and axhline to support this, so for
> example axvline will no longer autoscale the y axis. JDH
>
> 2006-06-13 Fix so numpy updates are backward compatible - TEO
>
> 2006-06-12 Updated numerix to handle numpy restructuring of
> oldnumeric - TEO
>
> 2006-06-12 Updated numerix.fft to handle numpy restructuring
> Added ImportError to numerix.linear_algebra for numpy -TEO
>
> 2006-06-11 Added quiverkey command to pylab and Axes, using
> QuiverKey class in quiver.py. Changed pylab and Axes
> to use quiver2 if possible, but drop back to the
> newly-renamed quiver_classic if necessary. Modified
> examples/quiver_demo.py to illustrate the new quiver
> and quiverkey. Changed LineCollection implementation
> slightly to improve compatibility with PolyCollection. - EF
>
> 2006-06-11 Fixed a usetex bug for windows, running latex on files
> with spaces in their names or paths was failing - DSD
>
> 2006-06-09 Made additions to numerix, changes to quiver to make it
> work with all numeric flavors. - EF
>
> 2006-06-09 Added quiver2 function to pylab and method to axes,
> with implementation via a Quiver class in quiver.py.
> quiver2 will replace quiver before the next release;
> it is placed alongside it initially to facilitate
> testing and transition. See also
> examples/quiver2_demo.py. - EF
>
> 2006-06-08 Minor bug fix to make ticker.py draw proper minus signs
> with usetex - DSD
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> Matplotlib-users mailing list
> Matplotlib-users@lists.sourceforge.net
> matplotlib-users List Signup and Options
>

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
matplotlib-users List Signup and Options

If 2-D is allowed,
I hope N×K is allowed,
not just N×2.

Cheers,
Alan Isaac

···

On Wed, 12 Jul 2006, Stefan van der Walt apparently wrote:

It would be useful to have plot accept a 2-D array as well. Would
patches for this be considered, or is there some reason why this can't
work? At the moment, doing
P.plot(z) where z is Nx2

Alan G Isaac wrote:

It would be useful to have plot accept a 2-D array as well. Would patches for this be considered, or is there some reason why this can't work? At the moment, doing P.plot(z) where z is Nx2

If 2-D is allowed,
I hope N×K is allowed,
not just N×2.

I have thought about this, but haven't gotten around to doing it. To do the whole job, we could handle a single NxK argument, and we could also handle pairs:
Nx1, NxK
NxK, Nx1
NxK, NxK

I think it is clear what the behavior should be with any of the pairs; essentially they should be broadcast to the NxK, NxK case, resulting in K lines.

In the single-argument NxK case, I think matlab plots subsequent columns against the first column. Is this what you would like? The alternative would be to plot each column against the row index, which is what happens in the Nx1 case, but I suspect the matlab-compatibility argument wins out here. Also, this maintains consistency with the original request for Nx2, which I assume was for plotting the second column against the first. I certainly would not want to have NxK behavior change as K goes from 2 to values larger than 2.

Eric

···

On Wed, 12 Jul 2006, Stefan van der Walt apparently wrote:

Yes.

Cheers,
Alan Isaac

···

On Wed, 12 Jul 2006, Eric Firing apparently wrote:

In the single-argument NxK case, I think matlab plots
subsequent columns against the first column. Is this what
you would like?

Stefan van der Walt wrote:

2006-06-22 Added support for numerix 2-D arrays as alternatives to
          a sequence of (x,y) tuples for specifying paths in
          collections, quiver, contour, pcolor, transforms.
          Fixed contour bug involving setting limits for
          color mapping. Added numpy-style all() to numerix. - EF

It would be useful to have plot accept a 2-D array as well. Would
patches for this be considered, or is there some reason why this can't
work? At the moment, doing

P.plot(z) where z is Nx2,

raises

RuntimeError: xdata and ydata must be the same length

So it doesn't look as though this functionality will override any
current feature.

Stefan,

To reply more directly to your proposal now that I have thought about it more: although I see the logic in it, I don't see much gain from your Nx2 idea; it not very hard to simply write P.plot(z[:,0], z[:,1]). Furthermore, implementing your idea would not conflict with present behavior, but it would interfere with an alternative version of fancy argument handling that I think would be much more generally useful.

Eric

···

On Tue, Jul 11, 2006 at 10:53:59PM -0400, Charlie Moad wrote:

Hi Eric

···

On Wed, Jul 12, 2006 at 05:59:16PM -1000, Eric Firing wrote:

To reply more directly to your proposal now that I have thought about it
more: although I see the logic in it, I don't see much gain from your
Nx2 idea; it not very hard to simply write P.plot(z[:,0], z[:,1]).
Furthermore, implementing your idea would not conflict with present
behavior, but it would interfere with an alternative version of fancy
argument handling that I think would be much more generally useful.

Could you expand on what you mean with alternative fancy argument handling?

Thanks
Stéfan

Thanks, I found the description in the other thread.

Regards
Stéfan

···

On Wed, Jul 12, 2006 at 05:59:16PM -1000, Eric Firing wrote:

To reply more directly to your proposal now that I have thought about it
more: although I see the logic in it, I don't see much gain from your
Nx2 idea; it not very hard to simply write P.plot(z[:,0], z[:,1]).
Furthermore, implementing your idea would not conflict with present
behavior, but it would interfere with an alternative version of fancy
argument handling that I think would be much more generally useful.