Re: [Matplotlib-users] MultipleLocator and MaxNLocator

Ryan, I stopped using gnuplot, because it requires the data to be
formatted in very specific ways :) I remember having functions just to
format the input data correctly for heat plots and the wrapper scripts
I wrote were quite convoluted. Matplotlib has its advantages for sure.
Otherwise I would not have switched. I'm just frustrated with being
back at the start line. Thanks for your help and bearing with my
impatience.

On Sat, Feb 14, 2015 at 7:42 PM, Ryan Nelson <rne...@gm...> wrote:
> Yep. I see your problem. My function and Eric's object should help here.
>
> A sore-spot with many folks coming over to Matplotlib from "X" is the fact
> that MPL does not calculate the size of text until the plot is generated.
> That means it doesn't always get text positioning, etc. exactly correct.
> That takes a little getting used to, and for me, it is minor.
>
> Admit it, Gnuplot as it's quirks as well :)  I always hated that it wouldn't
> cut off some markers at the edge of the screen. For example, with Gnuplot
> 4.6rev5 the following
> plot x with points ps 7
> Leads to a bunch of markers running over the axes limits. (Maybe there is a
> way to fix this now. Many years ago that was not the case.)
>
> Ryan
>
> On Sat, Feb 14, 2015 at 2:18 PM, Tommy Carstensen
> <tom...@gm...> wrote:
>>
>> Ryan, my use case is indeed that I want to avoid overlapping ticks and
>> I want to avoid them by not displaying them. Here is a guy with the
>> same problem:
>>
>> http://stackoverflow.com/questions/9422587/overlapping-y-axis-tick-label-and-x-axis-tick-label-in-matplotlib
>>
>> Here is the problem at the top left of my plot:
>> www.tommycarstensen.com/matplotlib.png
>>
>> I'll just set the ticks manually. Sadly seems like the easiest thing to
>> do.
>>
>> On Sat, Feb 14, 2015 at 7:01 PM, Ryan Nelson <rne...@gm...>
>> wrote:
>> > Tommy,
>> >
>> > I'm sorry. I forgot to hit send all *again*. Below is my original
>> > message,
>> > but the function I wrote is updated because it wasn't exactly
>> > correct....
>> >
>> > Ah. I was working on something to help out, so I'm just seeing Eric's
>> > very
>> > elegant solution, which I have yet to try. However, I feel like you
>> > might
>> > run into some problems if you always drop the first tick. For example,
>> > try
>> > this plot:
>> > ______________
>> > import matplotlib.pyplot as plt
>> > from matplotlib.ticker import MultipleLocator
>> > import numpy as np
>> > xs = np.linspace(2,12,1000)
>> > ys = np.sin(xs)
>> > n = 5
>> > fig = plt.figure()
>> > ax1 = fig.add_subplot(111)
>> > ax1.plot(xs, ys)
>> > ax1.xaxis.set_major_locator(MultipleLocator(5))
>> > plt.show()
>> > _____________
>> > In this case, dropping the first tick will result in only one tick on
>> > the
>> > screen.
>> >
>> > What is your use-case? Are you annoyed that the axis labels are
>> > overlapping
>> > at the far left? If that's the case, here's a little function
>> > (trimticks)
>> > that I whipped up that might help. It drops the far left or far right
>> > label
>> > if it is exactly at the edge of the axes. Should work for y axes as
>> > well.
>> > _____________
>> > def trimticks(ax, n=5):
>> >     xmin, xmax = ax.get_xlim()
>> >     if xmin%n == 0:
>> >         xmin = xmin+n
>> >     else:
>> >         xmin = xmin + n - xmin%n
>> >
>> >     if not xmax%n == 0:
>> >         xmax = xmax + n - xmax%n
>> >
>> >     ticks = np.arange(xmin, xmax, n)
>> >     ax.set_xticks(ticks)
>> >
>> > import matplotlib.pyplot as plt
>> > from matplotlib.ticker import MultipleLocator
>> > import numpy as np
>> > xs = np.linspace(0,20,10000)
>> > ys = np.sin(xs)
>> > fig = plt.figure()
>> > ax1 = fig.add_subplot(111)
>> > ax1.plot(xs, ys)
>> > trimticks(ax1)
>> > plt.show()
>> >
>> > ___________________
>> >
>> >
>> > On Sat, Feb 14, 2015 at 1:45 PM, Tommy Carstensen
>> > <tom...@gm...> wrote:
>> >>
>> >> Erik, that doesn't seem to work either. I tried this:
>> >>
>> >> import matplotlib.pyplot as plt
>> >> from matplotlib.ticker import MultipleLocator
>> >> class TrimmedMultipleLocator(MultipleLocator):
>> >>      def tick_values(self, vmin, vmax):
>> >>          return MultipleLocator.tick_values(self, vmin, vmax)[2:]
>> >> fig = plt.figure()
>> >> ax1 = fig.add_subplot(111)
>> >> ax1.xaxis.set_major_locator(TrimmedMultipleLocator(5))
>> >> #xticks[0].label1.set_visible(False)
>> >> #xticks[-1].label1.set_visible(False)
>> >> #ax1.set_xticks(ax1.xaxis.get_major_ticks()[1:-1])
>> >> ax1.plot(list(range(21)))
>> >> plt.show()
>> >>
>> >> Here is an example of the use of prune='lower', but it does not allow
>> >> one to set the tick step size:
>> >>
>> >>
>> >> http://stackoverflow.com/questions/9422587/overlapping-y-axis-tick-label-and-x-axis-tick-label-in-matplotlib
>> >>
>> >> I think my best bet is to just set those ticks manually.
>> >>
>> >> On Sat, Feb 14, 2015 at 6:19 PM, Eric Firing <ef...@ha...>
>> >> wrote:
>> >> > On 2015/02/14 7:33 AM, Tommy Carstensen wrote:
>> >> >> Thanks again Ryan. That's exactly what I want to achieve; i.e.
>> >> >> remove
>> >> >> the tick at 0 and only keep 5 and 10. Your solution works, but it's
>> >> >> a
>> >> >> bit of hack to use magic constants. I could however get those values
>> >> >> from the xlim.
>> >> >>
>> >> >> Eric, I would describe the desired tick placement algorithm as
>> >> >> removing the first tick on the axis. It can be achieved like this:
>> >> >> ax1.xaxis.set_major_locator(MaxNLocator(prune='lower'))
>> >> >
>> >> > Aha! The problem is that the MaxNLocator is the only one with the
>> >> > prune
>> >> > kwarg.  It could be added to the MultipleLocator.  For now, though,
>> >> > you
>> >> > can make your own specialized Locator, hardwired to omit the first
>> >> > tick,
>> >> > like this:
>> >> >
>> >> > from matplotlib.ticker import MultipleLocator
>> >> >
>> >> > class TrimmedMultipleLocator(MultipleLocator):
>> >> >      def tick_values(self, vmin, vmax):
>> >> >          return MultipleLocator.tick_values(self, vmin, vmax)[1:]
>> >> >
>> >> > then just use
>> >> >
>> >> > ax1.xaxis.set_major_locator(TrimmedMultipleLocator(5))
>> >> >
>> >> > I haven't tested it--but give it a try.  What it is doing is making a
>> >> > subclass of MultipleLocator, and altering only the one little bit of
>> >> > its
>> >> > behavior that you want to modify.  Everything else is automatically
>> >> > inherited from the base class, MultipleLocator.
>> >> >
>> >> > Eric
>> >> >
>> >> >
>> >> >>
>> >> >> But that then overrides this:
>> >> >> ax1.xaxis.set_major_locator(MultipleLocator(5))
>> >> >>
>> >> >> On Sat, Feb 14, 2015 at 5:27 PM, Ryan Nelson <rne...@gm...>
>> >> >> wrote:
>> >> >>> Tommy, (Sorry for the doubleup. I just realized I forgot to hit
>> >> >>> reply-all.)
>> >> >>>
>> >> >>> Do you want to remove the tick at 0 and only have 5,10, etc.? Could
>> >> >>> you just
>> >> >>> do something like this instead:
>> >> >>>
>> >> >>> import matplotlib.pyplot as plt
>> >> >>> from matplotlib.ticker import MultipleLocator
>> >> >>> fig = plt.figure()
>> >> >>> ax1 = fig.add_subplot(111)
>> >> >>> ax1.set_xticks(range(5,11,5))
>> >> >>> ax1.plot(range(11))
>> >> >>> plt.show()
>> >> >>>
>> >> >>> Ryan
>> >> >>>
>> >> >>>
>> >> >>> On Sat, Feb 14, 2015 at 10:47 AM, Tommy Carstensen
>> >> >>> <tom...@gm...> wrote:
>> >> >>>>
>> >> >>>> Thanks for you answer Eric. I had to get some sleep before trying
>> >> >>>> out
>> >> >>>> things. I currently have the code below, but it does not remove
>> >> >>>> the
>> >> >>>> zero value tick. It removes the tick at 5 and 10 however.
>> >> >>>>
>> >> >>>> import matplotlib.pyplot as plt
>> >> >>>> from matplotlib.ticker import MultipleLocator
>> >> >>>> fig = plt.figure()
>> >> >>>> ax1 = fig.add_subplot(111)
>> >> >>>> ax1.xaxis.set_major_locator(MultipleLocator(5))
>> >> >>>> xticks = ax1.xaxis.get_major_ticks()
>> >> >>>> #xticks[0].label1.set_visible(False)
>> >> >>>> #xticks[-1].label1.set_visible(False)
>> >> >>>> ax1.set_xticks(ax1.get_xticks()[1:-1])
>> >> >>>> ax1.plot(list(range(11)))
>> >> >>>> plt.show()
>> >> >>>>
>> >> >>>> On Sat, Feb 14, 2015 at 2:01 AM, Eric Firing <ef...@ha...>
>> >> >>>> wrote:
>> >> >>>>> On 2015/02/13 3:29 PM, Tommy Carstensen wrote:
>> >> >>>>>> Is it possible to combine MultipleLocator and MaxNLocator? One
>> >> >>>>>> seems
>> >> >>>>>> to erase the effect of the other.
>> >> >>>>>
>> >> >>>>> They are for different situations.  MultipleLocator is for when
>> >> >>>>> you
>> >> >>>>> know
>> >> >>>>> what you want your tick interval to be; MaxNLocator is for when
>> >> >>>>> you
>> >> >>>>> don't know that, but you do know roughly how many ticks you want,
>> >> >>>>> and
>> >> >>>>> what sort of numerical intervals are acceptable.
>> >> >>>>>
>> >> >>>>> Eric
>> >> >>>>>
>> >> >>>>>
>> >> >>>>>
>> >> >>>>>
>> >> >>>>>
>> >> >>>>> ------------------------------------------------------------------------------
>> >> >>>>> Dive into the World of Parallel Programming. The Go Parallel
>> >> >>>>> Website,
>> >> >>>>> sponsored by Intel and developed in partnership with Slashdot
>> >> >>>>> Media,
>> >> >>>>> is
>> >> >>>>> your
>> >> >>>>> hub for all things parallel software development, from weekly
>> >> >>>>> thought
>> >> >>>>> leadership blogs to news, videos, case studies, tutorials and
>> >> >>>>> more.
>> >> >>>>> Take
>> >> >>>>> a
>> >> >>>>> look and join the conversation now.
>> >> >>>>> http://goparallel.sourceforge.net/
>> >> >>>>> _______________________________________________
>> >> >>>>> Matplotlib-users mailing list
>> >> >>>>> Mat...@li...
>> >> >>>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>> >> >>>>
>> >> >>>>
>> >> >>>>
>> >> >>>>
>> >> >>>> ------------------------------------------------------------------------------
>> >> >>>> Dive into the World of Parallel Programming. The Go Parallel
>> >> >>>> Website,
>> >> >>>> sponsored by Intel and developed in partnership with Slashdot
>> >> >>>> Media,
>> >> >>>> is
>> >> >>>> your
>> >> >>>> hub for all things parallel software development, from weekly
>> >> >>>> thought
>> >> >>>> leadership blogs to news, videos, case studies, tutorials and
>> >> >>>> more.
>> >> >>>> Take a
>> >> >>>> look and join the conversation now.
>> >> >>>> http://goparallel.sourceforge.net/
>> >> >>>> _______________________________________________
>> >> >>>> Matplotlib-users mailing list
>> >> >>>> Mat...@li...
>> >> >>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>> >> >>>
>> >> >>>
>> >> >>
>> >> >>
>> >> >>
>> >> >> ------------------------------------------------------------------------------
>> >> >> Dive into the World of Parallel Programming. The Go Parallel
>> >> >> Website,
>> >> >> sponsored by Intel and developed in partnership with Slashdot Media,
>> >> >> is
>> >> >> your
>> >> >> hub for all things parallel software development, from weekly
>> >> >> thought
>> >> >> leadership blogs to news, videos, case studies, tutorials and more.
>> >> >> Take a
>> >> >> look and join the conversation now.
>> >> >> http://goparallel.sourceforge.net/
>> >> >> _______________________________________________
>> >> >> Matplotlib-users mailing list
>> >> >> Mat...@li...
>> >> >> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>> >> >>
>> >> >
>> >> >
>> >> >
>> >> >
>> >> > ------------------------------------------------------------------------------
>> >> > Dive into the World of Parallel Programming. The Go Parallel Website,
>> >> > sponsored by Intel and developed in partnership with Slashdot Media,
>> >> > is
>> >> > your
>> >> > hub for all things parallel software development, from weekly thought
>> >> > leadership blogs to news, videos, case studies, tutorials and more.
>> >> > Take
>> >> > a
>> >> > look and join the conversation now.
>> >> > http://goparallel.sourceforge.net/
>> >> > _______________________________________________
>> >> > Matplotlib-users mailing list
>> >> > Mat...@li...
>> >> > https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>> >>
>> >>
>> >>
>> >> ------------------------------------------------------------------------------
>> >> Dive into the World of Parallel Programming. The Go Parallel Website,
>> >> sponsored by Intel and developed in partnership with Slashdot Media, is
>> >> your
>> >> hub for all things parallel software development, from weekly thought
>> >> leadership blogs to news, videos, case studies, tutorials and more.
>> >> Take a
>> >> look and join the conversation now. http://goparallel.sourceforge.net/
>> >> _______________________________________________
>> >> Matplotlib-users mailing list
>> >> Mat...@li...
>> >> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>> >
>> >
>> >
>> >
>> > ------------------------------------------------------------------------------
>> > Dive into the World of Parallel Programming. The Go Parallel Website,
>> > sponsored by Intel and developed in partnership with Slashdot Media, is
>> > your
>> > hub for all things parallel software development, from weekly thought
>> > leadership blogs to news, videos, case studies, tutorials and more. Take
>> > a
>> > look and join the conversation now. http://goparallel.sourceforge.net/
>> > _______________________________________________
>> > Matplotlib-users mailing list
>> > Mat...@li...
>> > https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>> >
>
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming. The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>