Add ticks.integerSteps option to linear scale.

[etimberg]

When true, and stepSize, is not specified, the generated step size will be increased
to the nearest integer value.

Resolves #4103

[benmccann]

My main question with this PR is how do we want tick generation and the autoskipper to interact? E.g. the time scale has a lot of tick generation logic, but I think @simonbrunel would prefer that we remove most of it and rely on the autoskipper instead. I'm wondering if we want this logic in the autoSkipper in that case? E.g. an option like skipNonIntegerTicks or something along those lines

[etimberg]

That's a good question. I haven't really thought about the auto skipper. In an imagined redesigned auto skipper, having extra options would be nice. But reducing the overall number of ticks generated does have other implications like the fact that the auto skipper would have less things to consider which may make it faster

[etimberg]

@simonbrunel @benmccann I rebased this so it should pass the CI now

[simonbrunel]

@etimberg what do you think about a more generic option that allows to choose the rounding step, for example precision, a bit like jasmine toBeCloseTo:

• if undefined or false, the input is unchanged (eq. integerSteps: undefined/false)
• if 0, the input is rounded to the upper value (eq. integerSteps: true)
• if > 1, The number of decimal points to keep

[etimberg]

Sure, that works for me

[jcopperfield]

@etimberg I think it would be better rename the previous precision variable to factor and reuse it, instead of introducing the variable spacePrecision. This to improve readability and as it serves the same purpose later in the function.

function generateTicks(generationOptions, dataRange) {
	var ticks = [];
	// To get a "nice" value for the tick spacing, we will use the appropriately named
	// "nice number" algorithm. See http://stackoverflow.com/questions/8506881/nice-label-algorithm-for-charts-with-minimum-ticks
	// for details.

	var factor = 1;
	var spacing;

	if (generationOptions.stepSize && generationOptions.stepSize > 0) {
		spacing = generationOptions.stepSize;
	} else {
		var niceRange = helpers.niceNum(dataRange.max - dataRange.min, false);
		spacing = helpers.niceNum(niceRange / (generationOptions.maxTicks - 1), true);
		if (generationOptions.precision !== undefined) {
			factor = Math.pow(10, generationOptions.precision);
			spacing = Math.ceil(spacing * factor) / factor;
		}
	}
	var niceMin = Math.floor(dataRange.min / spacing) * spacing;
	var niceMax = Math.ceil(dataRange.max / spacing) * spacing;

	// If min, max and stepSize is set and they make an evenly spaced scale use it.
	if (generationOptions.min && generationOptions.max && generationOptions.stepSize) {
		// If very close to our whole number, use it.
		if (helpers.almostWhole((generationOptions.max - generationOptions.min) / generationOptions.stepSize, spacing / 1000)) {
			niceMin = generationOptions.min;
			niceMax = generationOptions.max;
		}
	}

	var numSpaces = (niceMax - niceMin) / spacing;
	// If very close to our rounded value, use it.
	if (helpers.almostEquals(numSpaces, Math.round(numSpaces), spacing / 1000)) {
		numSpaces = Math.round(numSpaces);
	} else {
		numSpaces = Math.ceil(numSpaces);
	}

	if (spacing < 1) {
		factor = Math.pow(10, spacing.toString().length - 2);
		niceMin = Math.round(niceMin * factor) / factor;
		niceMax = Math.round(niceMax * factor) / factor;
	}
	ticks.push(generationOptions.min !== undefined ? generationOptions.min : niceMin);
	for (var j = 1; j < numSpaces; ++j) {
		ticks.push(Math.round((niceMin + j * spacing) * factor) / factor);
	}
	ticks.push(generationOptions.max !== undefined ? generationOptions.max : niceMax);

	return ticks;
}

[benmccann]

I think precision is more understandable to end users. E.g. if you want to round to the nearest .01 it's pretty easy to see that this is two decimal places, but I'm not sure we would want to make the user calculate that they should pass in 100

[jcopperfield]

@benmccann I am not sure I understand your comment. My proposal wasn't to change the precision-option given by the user, but to use the variable precision inside the function, instead of introducing an extra variable spacingPrecision.

[jcopperfield]

Or more precisely renaming the previous use of precision to factor and use this for both tick spacing as tick rounding, as they both serve the same purpose.

[simonbrunel]

I agree with @jcopperfield, it's more readable renaming internal variables precision -> factor and spacingPrecision -> precision since these have different units / meanings.

[etimberg]

I agree, will make this change

[saadismail]

In which version will this PR will be live in the stable version? I can see that this was merged into master back in Apr 2018 but https://www.chartjs.org/docs/latest/axes/cartesian/linear.html is still outdated and current version (2.8.0) does not have this.

[benmccann]

@saadismail I'm not sure why you say https://www.chartjs.org/docs/latest/axes/cartesian/linear.html is outdated. I see "precision" on that page as was added in this PR

[saadismail]

@benmccann I am sorry, I did not read the whole conversation in the older thread and tried to look for integerSteps instead of precision. Thank you for the help and sorry for the inconvenience.