max_y is not inferred correctly when min_y_from_data=true

[openjck]

Steps to reproduce

Download and open this HTML document

Actual result

Screenshot

Description

There's a lot of empty space above the highest data point. The documentation says that by default, "the value [of max_y] is inferred from the dataset." But this doesn't appear to be happening.

If max_y is manually set to 3.2, the chart looks the way I would expect.

Expected result

The chart should look like this (ignoring the weird y-axis labels as described in #821):

[cnwangjie]

I think this may be a bug worth paying attention to.
It is caused by this line

max_val = (max_val < 0) ? max_val + (max_val - max_val * args.inflator) * use_inflator : max_val * (use_inflator ? args.inflator : 1);

in mg_min_max_numerical method.

In the example, the max_val is 3.2 at the beginning. After it multiplied by default inflator (10/9), it becomes 3.555555555555556. So it caused such a result.

If the value of y is bigger, the problem will be more obvious. If you set data as

const exampleData = [
    { x: 1, y: 1003.0 },
    { x: 2, y: 1003.1 },
    { x: 3, y: 1003.2 },
];

will cause the following case. (The max_val becomes 1114.6666666666667)

And the following case may be what we would rather see.

However, for most situations will use_inflator to be enabled. So I think we should make some changes in order to avoid appearing frequently.

I thought it might be possible to multiply inflator by difference between the max_val and the min_val. And let max_val add this value. Insteand of directly multiplying the max_val by inflator.

Also, I think the code of this method is too complicated. Maybe some optimization should make it easier to understand.