matplotlib-users Mailing List for matplotlib

Hello,
Let say we have a figure created by:
plt.plot(range(100))
On WX backend plt.grid(1) or key "G" responds finely for turning on/off the
grid lines. However when I log-scale both axes then plt.grid(1 or 0) or "G"
doesn't respond on minor grid lines.
Is there a way to control this behavior?
Thanks.
-- 
Gökhan

2010年4月15日 KrishnaPribadi <Kri...@ha...>:
> Is there any way to just turn
> off the exponent number in the right corner and force the x tick labels to
> be integers or floating point depending on how close one is zoomed?
Have a look at matplotlib/ticker.py. Most classes in ticker.py are
derived from class ticker.TickHelper. This in turn has a method
.set_axis(). So I think when you write your own formatter class, you
should be able to make use of this, because I guess it's called upon
attachment of the formatter to the axis, by the axis instance it's
attached to. This means, you should be able to retrieve the view
limits from the .axis attribute during __call__() of the formatter, so
you can choose the format used appropriate according to the view span.
Maybe, just as a thought, you should also have a look at the Locator
class you're using, to find out how the decision is made, when to use
integer tick positions and when to use fp ones.
I think you should derive your formatter class from class ticker.Formatter.
hth,
Friedrich

Hello,
I am currently using the sphinxext module to write documentation that includes automatically generated plots. I was wondering if there is a way to override the default savefig behavior for this - for example, I would like all the plots for the documentation to have the bbox_inches='tight' option to avoid any unnecessary whitespace. Is this something that can be done?
Thanks,
Thomas