Currently, our implementation of assert_xydata() uses the same tolerances for comparing the x-data and y-data. Currently, it takes the argument tolerance and applies it to both the x-data and the y-data.
This could cause issues when the x-axis and y-axis have values with different magnitudes.
For example, if the y-axis of a plot represents elavation in feet, the values might be in the thousands. If the x-axis represents time in minutes, the values might be tiny fractions of days (because Matplotlib converts all time data to days sine epoch). In this case, a reasonable tolerance for the y-axis would be orders of magnitude greater than a reasonable tolerance for the x-axis.
In cases like this, specifying a tolerance is only really useful if we can specify seperate tolerances for the x-data and y-data.
We should consider having different tolerance arguments for each axis.
Currently, our implementation of
assert_xydata()uses the same tolerances for comparing the x-data and y-data. Currently, it takes the argumenttoleranceand applies it to both the x-data and the y-data.This could cause issues when the x-axis and y-axis have values with different magnitudes.
For example, if the y-axis of a plot represents elavation in feet, the values might be in the thousands. If the x-axis represents time in minutes, the values might be tiny fractions of days (because Matplotlib converts all time data to days sine epoch). In this case, a reasonable tolerance for the y-axis would be orders of magnitude greater than a reasonable tolerance for the x-axis.
In cases like this, specifying a tolerance is only really useful if we can specify seperate tolerances for the x-data and y-data.
We should consider having different tolerance arguments for each axis.