Adds ability to scale Tilemap heatmap intensity values

[stormpython]

Closes #5946.

At times, data displayed in Tile maps needs to be log transformed in order to properly visualize the results. For a better explanation on this, please read the comments on issue #5946.

The pull requests adds the ability to scale heatmap intensity values using either a linear (default), logarithmic (base 10), or square root scale.

Please Note: With log transformed data, all data points must be greater than or equal to 1 as the log(0) is -Infinity and the log(negative number) is NaN. Also, see here.

This pull accomplishes creating scales using d3 quantitative scales. For more info on the API, please see the documentation.

[trevan]

Make sure the heatIntensity is between 0 and 1 as that is the scale that Leaflet.heat uses. #5236 has more information.

[stormpython]

@trevan thanks. Will make that correction right now.

[stormpython]

Thanks to @trevan's comments regarding heatIntensity value needing to be between 0 and 1, I have removed the previous logic which results in heatIntensity values greater than 1. From now on, all metric values are normalized to a range between 0 (inclusive) and 1 (inclusive). The tests have also been changed to reflect this.

[epixa]

This needs to be rebased on master or have master merged into it.

[w33ble]

The max value is no longer being used, and can be removed

[epixa]

This has conflicts again

[stormpython]

Yeah, I am working on fixing these.

[stormpython]

jenkins, test it.

[stormpython]

@jbudz passing this back to you.

[jbudz]

can we remove this line?

[jbudz]

LGTM

[stormpython]

jenkins, test it

[stormpython]

Failure due to a breaking change in elasticsearch master.

[stormpython]

@w33ble would you mind being the second reviewer here?

[w33ble]

@stormpython this looks fine, the changes are pretty small and the values that come out seem correct. My dataset doesn't look any different with different scales though... do you have an example or use case or something where I can see the varying results?

[stormpython]

@tbragin would you mind checking out this PR? I figure you have an appropriate dataset which needs a log scale for visualizing heat intensities. Its really hard to test the difference without an appropriate dataset.

[tbragin]

I tried it on a dataset that has some variability in concentration. I don't see any effect - screenshots below. I shared my dataset with @stormpython so he can play with it some more.

[stormpython]

Thanks @tbragin

[thomasneirynck]

The current heatmap has issues beyond this, e.g. #8763. Heatmaps already just give only a rough approximation of the underlying quantities of the data points. Rescaling the inputs only compounds possible mispreresentation. To avoid single-color blobs, the intensity of the heatmap should be under the control of the user, but not sure if we should do this by skewing the underlying values.

I will close this for now. We're looking at an overhaul of the mapping visualizations and can pick up #5946 in that context.