matplotlib-users Mailing List for matplotlib

>>>>> "Christian" == Christian Meesters <mee...@un...> writes:
 Christian> Hi, two bar plot questions from my side: - Is it
 Christian> possible to draw bar plots without surronding lines?
 Christian> How? - I'd like to draw the bars a bit transparent,
 Christian> but 'alpha' does not work. Is there an other way?
The trick is to make the edge and face colors the same. Setting the
linewidth to 0 can also help but doesn't work on every backend. You
can set alpha as on any Artist
 rects = bar([1,2,3], [4,5,6])
 setp(rects, facecolor='b', edgecolor='b', linewidth=0, alpha=0.5)
Call getp(anyartist) on any Artist instance to see what
properties are settable and what their current values are.
Should help :-)
JDH

>>>>> "Sascha" == Sascha <sas...@gm...> writes:
 Sascha> In my application, I am creating a second axes using
 Sascha> pylab's twinx and plot various lines on both axes. I'd
 Sascha> like to take advantage of the automatic change of the line
 Sascha> color that happens when using the plot()
 Sascha> command. Unfortunately, this is done only within one axes
 Sascha> i.e. when plotting on the other axes, the first line color
 Sascha> is used again. This results in multiple lines that have
 Sascha> the some color.
 Sascha> Is there a smart way to make Matplotlib use the next color
 Sascha> in sequence for the first line on a new axes?
There is no way to do this automatically w/o changing the internals of
the line cycle code. At present each Axes gets its own color cycle
counter, and twinx creates a new axes instance with a shared cycle.
You could hack the twinx code to share the counter
ax1._get_lines.count = ax2._get_lines.count
but this would break after a cla in the current implementation since
the _get_lines_count attr is reset there. The relevant code is in
Axes.cla, axes._process_plot_var_args and pylab.twinx if you want to
try and fix this in a sensible way and submit a patch.
JDH

>>>>> "Sascha" == Sascha <sas...@gm...> writes:
 Sascha> I am having some issues creating axis labels. When I set
 Sascha> horizontalalignment='right' (which is the default), the y
 Sascha> axis label is positioned correctly for the left
 Sascha> axis. Using 'center', the text is positioned too close to
 Sascha> the axis so that multiline text runs into the tick
 Sascha> labels. For the right axis (created with pylab.twinx(),
 Sascha> it's the other way around - 'right' is too close to the
 Sascha> tick labels and 'center' works better.
 Sascha> Any hints what I can do to position the labels correctly?
Try adjusting the tick "pad". See the matplotlib rc file and the
rcParams dictionary. The relevant beast is tick.major.pad
JDH

 phil> I created an instance of Figure then add
 phil> some subplots to it. 
 ...snip
 phil> Now i create another instance of Figure.
 phil> self.new_fig = Figure(figsize=(320,200),
 phil> dpi=80)
 phil> My aim is to add a Subplot instance to the
 phil> new Figure instance and changind the
 phil> number of rows and columns and the
 phil> position of the Subplot. I could have
 phil> done: self.new_fig.add_subplot(a)
The important thing to understand is that there is no concept of
"number of rows" of a figure, or "number of columns". The Subplot is
an instance of an Axes, and an Axes has a position property. The
position is left, bottom, width, height in fractional (0,1)
coordinates. When you say Subplot(111), matplotlib computes the
position but the figure does not have the property that it has one row
and one column. For example, you could do
ax1 = subplot(111)
ax2 = axes([0.8, 0.8, 0.15, 0.15])
and have another axes over the subplot. How many rows and columns
does the figure have? It's not particularly well defined, since it
has two axes, but not on a regular grid. Each figure maintains a list
of axes, and you can move them from one to another, and you can manage
their sizes in figure they live in with ax.get_position() and
ax.set_position()
 fig1.delaxes(ax1)
 ax1.set_position(something)
 ax2.set_position(something_else)
 fig2.add_axes(ax1)
If you want to set the position for a given subplot geometry (eg 211)
you can add the following method to axes.Subplot
 def change_geometry(self, numrows, numcols, num):
 'change subplot geometry, eg from 1,1,1 to 2,2,3'
 self._rows = rows
 self._cols = cols
 self._num = num
 self.update_params()
and call it like
 ax2.change_geometry(2,2,3)
You may want to make sure each axes instance you create has an unique
"label". The "current axes" machinery works by comaring the args and
kwargs you pass to add_axes. This is documented with an example in
the Figure.add_axes method.
JDH