[Matplotlib-devel] GSoC 2017: Layout Manager

Hannah story645 at gmail.com
Sat Apr 1 21:03:42 EDT 2017


Using the unit interface, CategoricalConvertor maps strings to numbers. If
unit support can be extended to more plotting functions, then the norm can
work directly in data space (so with strings). This means users can just
use a colormap without having to explicitly set a norm. A norm then would
just be used to specify which of the data to register and which to leave
out, or the data/color mapping-I started prototyping this in  one of the
PRs.

On Apr 1, 2017 4:54 AM, "Kanchana Ranasinghe" <kahnchana at gmail.com> wrote:

Hi,

Really sorry for constantly troubling you over the weekend. Another small
question about something I'm confused about. Regarding categorical data
support for the norm and cmap functions.

Currently, norm maps numerical data to [0,1] interval and cmap assigns
different colors to each point of [0,1] interval. This integrates with the
various scalarMappables to plot data. In making categorical units support,
norm will map strings to [0,1] interval and cmaps will remain the same. Is
this correct?

Thank you

With Regards
Kanchana Ranasinghe


On Fri, Mar 31, 2017 at 11:38 PM, Kanchana Ranasinghe <kahnchana at gmail.com>
wrote:

> Thanks a lot. I got it.
>
> With Regards
> Kanchana Ranasinghe
>
>
> On Fri, Mar 31, 2017 at 11:34 PM, Hannah <story645 at gmail.com> wrote:
>
>> Kind of? I think we may be talking about the same thing from different
>> angles. I'm thinking about any function that takes in (and therefore
>> returns) a cmap & norm that can then be passed into scalerMappables to make
>> a colorbar. This includes stuff like imshow, matshow, pcolor, pcolormesh,
>> scatter, and probably others.
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-devel/attachments/20170401/19d871c0/attachment-0001.html>


More information about the Matplotlib-devel mailing list