[Feature] 关于 sampling 采样细节保留的一些建议

[Yurunsoft]

What problem does this feature solve?

我们的需求场景有2个：

1. 16个温度曲线渲染在单独的dom中，每个曲线最多可能有30000个点，每1s刷新一次数据
2. 3条曲线渲染在单独的dom中，其中2条曲线最多可能有30000个点，第3条曲线最多可能有60000个点，每500ms刷新一次数据

场景 1 要求尽量不丢失细节，对高低温位置比较敏感。目前用 5.6.0-rc.1 的新 minmax 算法，放大后可能和不采样时有所出入。
场景 2 用 5.6.0-rc.1 的新 minmax 算法基本满足需求了。

建议开启采样后，如果放大了，缩放后展示区域的点数<=最大能渲染的点数，就使用原始数据去渲染。

或者可以根据放大缩小倍数，动态采样，这样能更好保留细节。

What does the proposed API look like?

不改变 API，或者增加可配置项。

[echarts-bot]

@Yurunsoft It seems you are not using English, I've helped translate the content automatically. To make your issue understood by more people and get helped, we'd like to suggest using English next time. 🤗

TRANSLATED

TITLE

[Feature] Some suggestions for preserving sampling details

BODY

What problem does this feature solve?

We have 2 demand scenarios:

1. 16 temperature curves are rendered in a separate dom. Each curve may have up to 30,000 points, and the data is refreshed every 1s.
2. 3 curves are rendered in separate dom, 2 of which may have up to 30,000 points, and the third curve may have up to 60,000 points, and the data is refreshed every 500ms.

Scenario 1 requires that details are not lost as much as possible and is sensitive to high and low temperature locations. Currently, the new minmax algorithm of 5.6.0-rc.1 is used, and the amplification may be different from that without sampling.
Scenario 2 uses the new minmax algorithm of 5.6.0-rc.1 to basically meet the needs.

It is recommended that after you turn on sampling, if you zoom in and the number of points in the zoomed display area <= the maximum number of points that can be rendered, use the original data for rendering.

Or you can dynamically sample according to the zoom factor, which can better preserve details.

What does the proposed API look like?

No changes to the API or adding configurable items.