Add defaults to arguments #11
Conversation
|
With this functionality, we get something where: geodataframe.plot('variate')
legendgram(geodataframe.variate)yields something that should look OK 😄 |
| f = plt.gcf() | ||
| if ax is None: | ||
| ax = plt.gca() | ||
| if pal is None: |
There was a problem hiding this comment.
Maybe I am missing something here, but why do you use None and an if clause for a default? You could just directly assign 'viridis' to the parameter in the api. This way users also see how the input out to look like?
There was a problem hiding this comment.
Just curious maybe I can learn something :)
There was a problem hiding this comment.
Good question! I had intended to somehow detect the current matplotlib palette and use that if pal=None. But, I couldn't quite figure out how to do that. So, I deleted the exploration in this if branch & simply used pal = 'viridis'.
Ideally, you'd use pal=None and if pal is None, somehow check what colormap the user is using currently. I don't quite know how to do that.
There was a problem hiding this comment.
Ok, yeah that sounds great!
do you think something like
'''
fig, ax = plt.subplots()
im = ax.imshow(a, cmap="viridis")
cm = im.get_cmap()
print(cm.name)
"""
through
'Matplotlib.cm.get_cmap()' could work?
Happy to leave this as is and deal with it later. Maybe add a note, so others can keep finding solutions? :)
There was a problem hiding this comment.
Ah, well that's simole... how I missed it idk.
No reason to merge an incomplete if that works! I'll take a look tomorrow.
There was a problem hiding this comment.
@ljwolf Is this still under consideration? I put together a small gist demonstrating a helper function built on top of @slumnitz's idea. Basically, it uses the most recent colormap of an image, if the axis has any associated images. Otherwise, it defaults to viridis.
There was a problem hiding this comment.
Yes, totally! push on top of this so that it updates!
Co-authored-by: James Gaboardi <jgaboardi@gmail.com>
|
EDIT @ljwolf :
|
|
@ljwolf gentle ping to see if there is still interest in getting this merged. |
This adds a few enhancements relating to default arguments & alternative position/size options for
loc. It also moves the examples folder tonotebooks, to keep in line with other PySAL projects.defaults
breaksthat is 10 evenly-spaced percentilespalthat is'viridis'position/size options
interpret a
loc = (x,y,w,h)tuple as the first two coordinates being the anchor (left bottom corner of the new axis) and the second two being the width & height of the new axis, relative to the existing axis. So,(0,0,.5,.25)would nominally start in the lefthand corner, stretch to the middle of the plot, and grow to the bottom half.We could make that better for UX by providing a separate width ratio thing, but I think since position & width/height affect each other depending on the later matplotlib spacing calls (e.g.
f.tight_layout()can really squeeze/realign the subaxis), I liked keeping this tuple as a two or four.