Bug when using categorical axes and colors in plotly.express

According to the documentation:

By default, Plotly Express lays out categorical data in the order in which it appears in the underlying data.

However, when using categorical axes in conjunction with colors, things do not show up in the order of the underlying data anymore.

import plotly.express as px
pdata_x = ["A","B",None,"B","C",None,"C","D", None, "D", "E", None]
pdata_y = [1,1,None,2,2,None,3,3,None,4,4,None]
pdata_color = [1,1,1,2,2,2,2,2,2,1,1,1]
px.line(x=pdata_x, y=pdata_y, color=pdata_color).show()

Output is

When not(!) using colors, we would get the correct output:

px.line(x=pdata_x, y=pdata_y ).show()

For anyone who comes across this problem, a temporary fix is explicitely specifying the category order:

px.line(x=pdata_x, y=pdata_y, color=pdata_color, category_orders={"x":["A","B","C","D"]}).show()

plotly==4.14.3
Python 3.8.0 (default, Feb 25 2021, 22:10:10)

Start by running the supplied Python and plotly.express px.line example with categorical x values and colors, then compare it with the no-color and explicit category_orders cases. Trace how categorical ordering is determined for colored lines. Done means the default order follows the underlying data even when color is used, while the documented explicit-order workaround continues to work.