Just wondering what kind of smoothing factor matplotlib use by default when we create a line graph with around 200,200 x, y data points. I didnt use any soothing as it was smoothened already and wanna know the exact way how matplotlib does that.
For the AGG backend we “simplify” the lines by dropping any points that “do not change the plot” before rasterizing. That is if you have N co-linear points Matplotlib will draw a line from the first to last and drop the (redundent) points in the middle (and as of mpl2.1 this includes back-tracking). See Performance — Matplotlib 3.5.1 documentation for details of how to control what the tolerances for “small” changes.
I do not recall off the top of my head how much we simplify the paths on the way out to esp/pdf/svg.