位置不断变化的openlayers集群非常慢 [英] openlayers clustering is very slow with ever changing positions

问题描述

在openlayers(5/6)中使用clusterSource时，我可以很好地对功能进行聚类.

when using a clusterSource in openlayers (5/6) i can cluster features nicely.

除了作为我的clustersource来源的vectorsource以外，它会不断更改数据. 我有一个websocket可以为我提供所需的位置信息(例如，海上运输的船只或Flightradar24上的飞机) 底层源代码的每一次更改都让我有种感觉，我的聚类工作就完成了. 有没有办法邮寄该更新?还是手动调用?

Except the vectorsource that is the source of my clustersource changes data constantly. I have a websockets that feeds me positions of stuff (like ships in marinetrafic or airplanes on flightradar24) I got the feeling with every change in the underlying source, my clustering gets done all over. Is there a way to postphone that update? or call it manually?

我可以缓冲输入流并一次处理更多的消息(例如每3秒更新1000个位置) 我希望在处理完这1000条消息之后进行聚类. 现在它变得如此缓慢，以至于我感觉到它在每条消息中都会发生(所以每3秒1000次)

I can buffer the input stream and process a higher number of messages at once (like 1000 position updates every 3 seconds) And i want the clustering to happen AFTER i processed those 1000 messages. It gets so slow now that i got the feeling it happens every message (so 1000 times every 3 seconds)

另一种优化方法是仅对当前可见的项目进行聚类(在聚类距离一侧的边缘周围有一个缓冲区) 我认为这将节省大量计算.

An other optimalisation would be to only cluster the items that are currently visible (with a buffer around the edges the side of the cluster distance) That would save a lot of calculation in my opinion.

这里有人可以帮我吗?

现在我得到了这样的信息(以元代码):

right now i got something like this (in meta-code):

const features = {};  //my own reference collection
const buffer = [];  //it's get filled somewhere else (constantly)

function processBuffer() {

  //here we like to pause the clustering logic

  buffer.forEach(function(message){
    let feature = features[message.id];

    if (feature) {
      //existing feature
      feature.getGeometry().setCoordinates(message.coord);
    } else {
      //new feature
      feature = ... create feature ...
      vectorSource.addFeature(feature);
    }
  });

  //here we like to continue the clustering logic

  //reset buffer
  buffer.length = 0;
}

setInterval(processBuffer, 3000);

推荐答案

使用.clear()和.addFeatures()通过数组替换源中的所有功能应该更加有效.

Replacing all the features in the source via an array using .clear() and .addFeatures() should be much more efficient.

这篇关于位置不断变化的openlayers集群非常慢的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！