Cutoff of cells in scatter plots when specifying custom figsize

[WeilerP]

When plotting the latent space, for example the velocity stream embedding, with specifying figsize can lead to cells being cut off at the border (see picture below).

import scvelo as scv

scv.settings.verbosity = 3
scv.settings.presenter_view = True
scv.settings.set_figure_params('scvelo')

adata = scv.datasets.pancreas()

scv.pp.filter_and_normalize(adata, min_shared_counts=20, n_top_genes=2000)
scv.pp.moments(adata, n_pcs=30, n_neighbors=30)

scv.tl.velocity(adata, mode='deterministic')
scv.tl.velocity_graph(adata)

scv.pl.velocity_embedding_stream(
    adata, legend_loc="right", basis='umap', fontsize=20, dpi=300, figsize=[4,4]
)

Versions

scvelo==0.2.4  scanpy==1.8.2  anndata==0.7.8  loompy==3.0.6  numpy==1.20.3  scipy==1.7.3  matplotlib==3.5.0  sklearn==1.0.1  pandas==1.3.4

[WeilerP]

@seigfried, I managed to reproduce the issue reported in #765 using the Pancreas dataset with figsize=[4, 4]. I suspect you can fix the issue by using a different figure size.

[seigfried]

@WeilerP It works with smaller figsizes so I will stick to those for now. Thanks for your help

[hyjforesight]

Hello @WeilerP,
We have the same issue. Is it possible to remove the limit of figsize? I mean, no matter what the value of figsize, it can generate the plot without cutting off at the border?
Thanks!
Best,
YJ

[WeilerP]

@hyjforesight, there is always a figure size set. This is a standard matplotlib argument/workflow.

[esfandyari]

@hyjforesight @WeilerP it's been a while, but in case someone is facing the same problem: for me, using a different figsize didn't solve the problem, but adding the parameter add_margin=0.2 did the trick (scv v0.2.5, matplotlib v3.5.3).