Unexpected inverted y-axis from `FigureResampler`

[predict-idlab]

Unexpected inverted y-axis from `FigureResampler`

[darynwhite]

Well, that did fix it, and you were correct. Despite it being NaNs and float32 data, Pandas had it be of dtype: object.

Your suggestion of calling it with .astype("float32") worked before pulling the PyPi update, but the changes from #90 also fixed it.

Thanks! I look forward to integrating this into our workflows!

[jvdd]

Hi @darynwhite,

Sounds like an awesome project!
Thanks for pointing out this issue :)

This is indeed unexpected / weird behavior. Plotly-Resampler should behave exactly the same as plotly.

I believe in your exanple plotly-resampler considers the datatype of hf_y as string (categorical) - as your y-axis is inverted and takes weird non-uniform steps.
Does it work as expected when you call .astype("float32") on the series that you pass as hf_y?.

Can you give me more details of the datatype of hf_y (and perhaps also hf_x)?
Giving a minimal reproducible example would be even more awesome!

Cheers, Jeroen

[jonasvdd]

Hi @darynwhite (@jvdd),

Thank you for submitting this issue!

I digged into it, and was able to reconstruct a minimal example of the undesired behavior.
The problem indeed is that your input hf_y data is of type object, which is cast to string by this code snippet below ⬇️

plotly-resampler/plotly_resampler/figure_resampler/figure_resampler_interface.py

Lines 677 to 682 in cc951c1

	# If the categorical or string-like hf_y data is of type object (happens
	# when y argument is used for the trace constructor instead of hf_y), we
	# transform it to type string as such it will be sent as categorical data
	# to the downsampling algorithm
	if hf_y.dtype == "object":
	hf_y = hf_y.astype("str")

But as this is diverging behavior from plotly w.r.t. numeric series of dtype object, I will fix this in a future PR!
Will create a release fix later this day!

[jonasvdd]

Is fixed in plotly-resampler 0.7.2.2