Better support for plotting libraries

Siuba works with plotnine directly after >>, as

from siuba import _, mutate
from siuba.data import mtcars
import plotnine as p9

(mtcars
    >> mutate(hp_per_cyl = _.hp / _.cyl)
    >> p9.ggplot(p9.aes(x='cyl', y='hp_per_cyl')) + p9.geom_point()
)

… but needs a pipe to work with other plotting libraries like seaborn, altair, lets-plot, plotly, etc., as

from siuba import pipe
import plotly.express as px

(mtcars
    >> mutate(hp_per_cyl = _.hp / _.cyl)
    >> pipe(lambda d: px.scatter(data_frame=d, x='cyl', y='hp_per_cyl'))
)

Could other plotting libraries be supported directly after the >>?

[machow]

I think R side-steps the need for a pipe function by using non-standard evaluation 😬

[essicolo]

Agreed. Or is naming pipe otherwise an option? Is there a grammar for a similar function in dplyr?