matplotlib-users

[Matplotlib-users] only vertical grid lines?

[Jochen V. <vo...@se...>]

Hello,

I used matplotlib to create a graphical representation of the boot
process of my Debian GNU/Linux system.  The result can be found at

    http://seehuhn.de/comp/bootlog.html

and especially in the (2400x1500 sized) picture

    http://seehuhn.de/comp/bootlog2.png

I have a question about matplotlib usage: I would like the picture to
have vertical grid lines at positions 1, 2, 3, 4, ..., not horizonzal
grid lines.  On the horizontal x-Axis I would like to see labels at
positions 0, 5, 10, 15, ... and I would like to get rid of the labels
on the y-Axis.  Is this possible?  How do I do it?

All the best,
Jochen
--=20
http://seehuhn.de/

Re: [Matplotlib-users] only vertical grid lines?

[John H. <jdh...@ac...>]

>>>>> "Jochen" == Jochen Voss <vo...@se...> writes:

    Jochen> Hello, I used matplotlib to create a graphical
    Jochen> representation of the boot process of my Debian GNU/Linux
    Jochen> system.  The result can be found at

    Jochen>     http://seehuhn.de/comp/bootlog.html

    Jochen> and especially in the (2400x1500 sized) picture

    Jochen>     http://seehuhn.de/comp/bootlog2.png

Very nice - you manage to pack a lot of information into the graph.
Edward Tufte would be proud!  It's nice to have folks stress testing
matplotlib

    Jochen> I have a question about matplotlib usage: 

  * I would like the picture to have vertical grid lines at positions
    1, 2, 3, 4, ..., On the horizontal x-Axis I would like to see
    labels at positions 0, 5, 10, 15, ...

The best way to do this would be to use minor ticks on 1,2,3,4,etc and
major ticks on 0,5,10,15 - the grids occur at the locations of the
ticks, and you can customize how you want the ticks labeled.  See
examples/major_minor_demo1.py and the use of the MultipleLocator.  The
Multiple locator places ticks at multiples of some base. For you minor
ticks you would use MultipleLocator(1) and for your major ticks
MultipleLocator(1).  The labels for the minor ticks are off by default
which is what you want, though you can customize this.

See http://matplotlib.sf.net/matplotlib.ticker.html for more info.

  * not horizonzal grid lines.  

The ax.xaxis.grid and ax.yaxis.grid functions can be used to
selectively turn on and off the grids for the respective axes.  The
signature is

    def grid(self, b, which='major'): 

where b is a boolean.  You can use this to control gridding for the
major and minor ticks on a per-axis basis.  So 

  ax.yaxis.grid(False) 



selectively turns off the grids for the y axis major tick grid lines.
The minor grids are off by default.  You can turn them on, eg with

   ax.xaxis.grid(True, which='minor') 

  * and I would like to get rid of the labels on the y-Axis.  Is this
    possible?

ax.set_yticklabels('')

See also the functions xticks and yticks in the matlab interface for
easy customization of tick locations and labels.

Cheers,
JDH

Re: [Matplotlib-users] only vertical grid lines?

[John H. <jdh...@ac...>]

>>>>> "Jochen" == Jochen Voss <vo...@se...> writes:

    Jochen> Hello, I used matplotlib to create a graphical
    Jochen> representation of the boot process of my Debian GNU/Linux
    Jochen> system.  The result can be found at

    Jochen>     http://seehuhn.de/comp/bootlog.html

    Jochen> and especially in the (2400x1500 sized) picture

    Jochen>     http://seehuhn.de/comp/bootlog2.png

After looking again at your plot, particularly the rectangle parts
with text labels, it occurs to me that it would be nice to have a
general purpose class for this (rectangular boxes with text labels).
The class matplotlib.tables.Cell strives for this, but is not feature
complete.  What would be nice would be to be able to specify the
location of the text with respect to the box, with horizontal and
vertical alignment. This wouldn't be too hard, since text already has
these alignment flags with respect to an x,y location ( eg
http://matplotlib.sf.net/screenshots.html#align_text ).  

For full control, one would want to be able to specify the location of
x,y with respect to the rectangle (left, center, bottom, right, etc)
*and* the alignment of the text with respect to the x,y location, so
that one could have

--------------------
|                  |
|    centered      |
|                  |
--------------------

--------------------
|                  |
|                  | to-the-right
|                  |
--------------------


--------------------
|top-left-under    |
|                  | 
|                  |
--------------------


top-left-over
--------------------
|                  |
|                  | 
|                  |
--------------------

and so on.....  All the components are in place to make this
relatively easy from a layout persepctive.  One just has to plug them
together and make a nice, intuitive interface.  

JDH     

Re: [Matplotlib-users] only vertical grid lines?

[Jochen V. <vo...@se...>]

Hello John,

thanks a lot for your help.  My picture is improving.
But one problem is left: the script

    from matplotlib.matlab import *
    from matplotlib.ticker import MultipleLocator

    subplot(211)
    ax=3Dgca()
    ax.xaxis.set_minor_locator(MultipleLocator(1))
    ax.xaxis.set_major_locator(MultipleLocator(5))
    ax.xaxis.grid(True, which=3D"minor")
    ax.yaxis.grid(False)
    semilogy([1,2,3,4,5,6,7,8,9,10,11],[1,2,3,4,5,6,7,8,9,10,11],basey=3D2)

    show()

has vertical grid lines ONLY on the minor ticks, which looks silly.
How do I get regularly spaced grid lines here?

All the best,
Jochen
--=20
http://seehuhn.de/

Re: [Matplotlib-users] only vertical grid lines?

[John H. <jdh...@ac...>]

>>>>> "Jochen" == Jochen Voss <vo...@se...> writes:


    Jochen> has vertical grid lines ONLY on the minor ticks, which
    Jochen> looks silly.  How do I get regularly spaced grid lines
    Jochen> here?

This is a consequence of the fact that the ticker skips minor ticks
which coincide with major ticks.  I am not sure this is the right
behavior, but was implemented to avoid other problems (overlapping
gridlines, for example, with anti-aliasing produces undesirable
results).  All you need to do is turn both major and minor grids on

    ax.xaxis.grid(True, which="minor")
    ax.xaxis.grid(True, which="major")

Should work....

JDH