Plot overhaul - Githubissues

scverse / scirpy

A scanpy extension to analyse single-cell TCR and BCR data.

https://scirpy.scverse.org/en/latest/

BSD 3-Clause "New" or "Revised" License

206 stars 32 forks source link

Plot overhaul #247

Open grst opened 3 years ago

grst commented 3 years ago

Follow scanpy's plotting conventions more closely.

What should a plotting function return?
What parameters should it have? Do we really need all of show, ax, return_fig?
If you'd design scanpy plotting from scratch, how would you do it?

Supersedes #193.

[ ] Possible to build a common plotting library for the scanpy ecosystem?
[ ] deal with coloring of non-categorical pandas columns

grst commented 8 months ago

Maybe always return the figure instead of the axes? Seems to work well with jupyter (doesn't show anything similar to "<Axes: title={'center': 'Number of cells in airr:receptor_subtype by gex:source'}, xlabel='airr:receptor_subtype', ylabel='Number of cells'>".

grst commented 5 months ago

Accompany with "plotting cookbook",

how to make figures of a certain dimension
how to combine figures into one
how to save figures in high quality
how to set title/suptitle

grst commented 5 months ago

Something like

def plot_mpl():
    with plt.ioff():
        fig, ax = plt.subplots()
        ax.plot(range(10))
        return fig

works well in Jupyter. It shows the figure and one can use the figure object for saving etc. if one wants to.

If someone works in ipython, they'd need to manually call .show() -- but I think this would be ok.