Results 396 comments of Jérôme Dockès

indeed, replacing the simple SVG with plotly figures would jeopardize a key feature of the tablereport: that it tends to always work and display correctly in different environments and screen...

given that there are not only Lists but also Arrays and Structs, and that just "list" without reference to a specific dtype may be ambiguous (e.g. should pandas object series...

it's a function you need, and it already exists (wrap_transformer) -- you just need to come up with a better name & expose it in the public API

> We just need to expose it as an object. do you mean a transformer class, and if so why do you think it is needed rather than just a...

rather than ApplySeparateCols I would suggest ApplyOnEachCol which is shorter and highlights the fact that the transformation is done independently on each column by using the singular and "each" Normally...

yeah maybe that's a good alternative to #1723 and #1714 , letting users easily select such columns and do what encoding they prefer with them