Auto min value on log scale

[nickofthyme]

Is your feature request related to a problem? Please describe.
When using a log scale the min value is hard coded to 1.

elastic-charts/src/scales/constants.ts

Line 40 in f99ba1f

export const LOG_MIN_ABS_DOMAIN = 1;

This causes issues in charts with a dataset with values less than 1 causing the values to not be shown.

https://codesandbox.io/s/happy-fog-pkd1j?file=/src/App.tsx

Describe the solution you'd like
EC should detect the min value from the data. Also, allow users to set the min log value via settings.

Describe alternatives you've considered
Setting explicit min domain value.

Kibana Cross Issues
elastic/kibana#83917

Checklist

• this request is checked against already exist requests
• every related Kibana issue is listed under Kibana Cross Issues list
• kibana cross issue tag is associated to the issue if any kibana cross issue is present

[th0ger]

Yes this seems to be related with (if not introduced by) PR #896

I would also add to the solution in PR #896 , that setting "dummy 1 baseline for log y scale" is not ideal, since there can be values 0<x<1. It would be better to set that baseline value to the minimum of the y-axis (whether that value is found via autoscaling, fixing it by "axis extents" or any other means).

My comments only apply for the log-scale case.

[nickofthyme]

Yup totally, this baseline value was added to match the kibana functionality for log scales. I think the intent was to limit the compression of data when most of the data is in a higher order of magnitude while only a few are in a much lower order of magnitude. But really this is hiding potentially valuable data. I could see this as an optional value but not set by default.

[th0ger]

Yes, for example in some datasets 0 may be stored like 1e-17 or so due to limited precision of calculations.

But it should be up to the user to set an (optional) lowest value that the ymin autoscaling can obtain. I don't mean a fixed value, but a lowest threshold for the autoscaling.

But in fact, one could simply filter the data for y<y_threshold. So is it really needed?

[markov00]

🎉 This issue has been resolved in version 25.0.0 🎉

The release is available on:

• GitHub release
• npm package (@latest dist-tag)

Your semantic-release bot 📦🚀

[markov00]

🎉 This issue has been resolved in version 25.1.0 🎉

The release is available on:

• GitHub release
• npm package (@latest dist-tag)

Your semantic-release bot 📦🚀