提高Baidu Map聚合的效率
百度的MAP的例子里提供了一个聚合效果，地址是 http://developer.baidu.com/map/jsdemo.htm#c1_4 ，效果图如下图：
这个效果很赞，但效率很低，当数据量达到5000的时候就难以忍受了，加载和地图缩放都很卡，用户体验很差劲。官方提供的MarkerClusterer.js 文件是这样的：
/** * @fileoverview MarkerClusterer标记聚合器用来解决加载大量点要素到地图上产生覆盖现象的问题，并提高性能。 * 主入口类是<a href="symbols/BMapLib.MarkerClusterer.html">MarkerClusterer</a>， * 基于Baidu Map API 1.2。 * * @author Baidu Map Api Group * @version 1.2 */ /** * @namespace BMap的所有library类均放在BMapLib命名空间下 */ var BMapLib = window.BMapLib = BMapLib || {}; (function(){ /** * 获取一个扩展的视图范围，把上下左右都扩大一样的像素值。 * @param {Map} map BMap.Map的实例化对象 * @param {BMap.Bounds} bounds BMap.Bounds的实例化对象 * @param {Number} gridSize 要扩大的像素值 * * @return {BMap.Bounds} 返回扩大后的视图范围。 */ var getExtendedBounds = function(map, bounds, gridSize){ bounds = cutBoundsInRange(bounds); var pixelNE = map.pointToPixel(bounds.getNorthEast()); var pixelSW = map.pointToPixel(bounds.getSouthWest()); pixelNE.x += gridSize; pixelNE.y -= gridSize; pixelSW.x -= gridSize; pixelSW.y += gridSize; var newNE = map.pixelToPoint(pixelNE); var newSW = map.pixelToPoint(pixelSW); return new BMap.Bounds(newSW, newNE); }; /** * 按照百度地图支持的世界范围对bounds进行边界处理 * @param {BMap.Bounds} bounds BMap.Bounds的实例化对象 * * @return {BMap.Bounds} 返回不越界的视图范围 */ var cutBoundsInRange = function (bounds) { var maxX = getRange(bounds.getNorthEast().lng, -180, 180); var minX = getRange(bounds.getSouthWest().lng, -180, 180); var maxY = getRange(bounds.getNorthEast().lat, -74, 74); var minY = getRange(bounds.getSouthWest().lat, -74, 74); return new BMap.Bounds(new BMap.Point(minX, minY), new BMap.Point(maxX, maxY)); }; /** * 对单个值进行边界处理。 * @param {Number} i 要处理的数值 * @param {Number} min 下边界值 * @param {Number} max 上边界值 * * @return {Number} 返回不越界的数值 */ var getRange = function (i, mix, max) { mix && (i = Math.max(i, mix)); max && (i = Math.min(i, max)); return i; }; /** * 判断给定的对象是否为数组 * @param {Object} source 要测试的对象 * * @return {Boolean} 如果是数组返回true，否则返回false */ var isArray = function (source) { return '[object Array]' === Object.prototype.toString.call(source); }; /** * 返回item在source中的索引位置 * @param {Object} item 要测试的对象 * @param {Array} source 数组 * * @return {Number} 如果在数组内，返回索引，否则返回-1 */ var indexOf = function(item, source){ var index = -1; if(isArray(source)){ if (source.indexOf) { index = source.indexOf(item); } else { for (var i = 0, m; m = source[i]; i++) { if (m === item) { index = i; break; } } } } return index; }; /** *@exports MarkerClusterer as BMapLib.MarkerClusterer */ var MarkerClusterer = /** * MarkerClusterer * @class 用来解决加载大量点要素到地图上产生覆盖现象的问题，并提高性能 * @constructor * @param {Map} map 地图的一个实例。 * @param {Json Object} options 可选参数，可选项包括：<br /> * markers {Array<Marker>} 要聚合的标记数组<br /> * girdSize {Number} 聚合计算时网格的像素大小，默认60<br /> * maxZoom {Number} 最大的聚合级别，大于该级别就不进行相应的聚合<br /> * minClusterSize {Number} 最小的聚合数量，小于该数量的不能成为一个聚合，默认为2<br /> * isAverangeCenter {Boolean} 聚合点的落脚位置是否是所有聚合在内点的平均值，默认为否，落脚在聚合内的第一个点<br /> * styles {Array<IconStyle>} 自定义聚合后的图标风格，请参考TextIconOverlay类<br /> */ BMapLib.MarkerClusterer = function(map, options){ if (!map){ return; } this._map = map; this._markers = []; this._clusters = []; var opts = options || {}; this._gridSize = opts["gridSize"] || 60; this._maxZoom = opts["maxZoom"] || 18; this._minClusterSize = opts["minClusterSize"] || 2; this._isAverageCenter = false; if (opts['isAverageCenter'] != undefined) { this._isAverageCenter = opts['isAverageCenter']; } this._styles = opts["styles"] || []; var that = this; this._map.addEventListener("zoomend",function(){ that._redraw(); }); this._map.addEventListener("moveend",function(){ that._redraw(); }); var mkrs = opts["markers"]; isArray(mkrs) && this.addMarkers(mkrs); }; /** * 添加要聚合的标记数组。 * @param {Array<Marker>} markers 要聚合的标记数组 * * @return 无返回值。 */ MarkerClusterer.prototype.addMarkers = function(markers){ for(var i = 0, len = markers.length; i <len ; i++){ this._pushMarkerTo(markers[i]); } this._createClusters(); }; /** * 把一个标记添加到要聚合的标记数组中 * @param {BMap.Marker} marker 要添加的标记 * * @return 无返回值。 */ MarkerClusterer.prototype._pushMarkerTo = function(marker){ var index = indexOf(marker, this._markers); if(index === -1){ marker.isInCluster = false; this._markers.push(marker);//Marker拖放后enableDragging不做变化，忽略 } }; /** * 添加一个聚合的标记。 * @param {BMap.Marker} marker 要聚合的单个标记。 * @return 无返回值。 */ MarkerClusterer.prototype.addMarker = function(marker) { this._pushMarkerTo(marker); this._createClusters(); }; /** * 根据所给定的标记，创建聚合点 * @return 无返回值 */ MarkerClusterer.prototype._createClusters = function(){ var mapBounds = this._map.getBounds(); var extendedBounds = getExtendedBounds(this._map, mapBounds, this._gridSize); for(var i = 0, marker; marker = this._markers[i]; i++){ if(!marker.isInCluster && extendedBounds.containsPoint(marker.getPosition()) ){ this._addToClosestCluster(marker); } } }; /** * 根据标记的位置，把它添加到最近的聚合中 * @param {BMap.Marker} marker 要进行聚合的单个标记 * * @return 无返回值。 */ MarkerClusterer.prototype._addToClosestCluster = function (marker){ var distance = 4000000; var clusterToAddTo = null; var position = marker.getPosition(); for(var i = 0, cluster; cluster = this._clusters[i]; i++){ var center = cluster.getCenter(); if(center){ var d = this._map.getDistance(center, marker.getPosition()); if(d < distance){ distance = d; clusterToAddTo = cluster; } } } if (clusterToAddTo && clusterToAddTo.isMarkerInClusterBounds(marker)){ clusterToAddTo.addMarker(marker); } else { var cluster = new Cluster(this); cluster.addMarker(marker); this._clusters.push(cluster); } }; /** * 清除上一次的聚合的结果 * @return 无返回值。 */ MarkerClusterer.prototype._clearLastClusters = function(){ for(var i = 0, cluster; cluster = this._clusters[i]; i++){ cluster.remove(); } this._clusters = [];//置空Cluster数组 this._removeMarkersFromCluster();//把Marker的cluster标记设为false }; /** * 清除某个聚合中的所有标记 * @return 无返回值 */ MarkerClusterer.prototype._removeMarkersFromCluster = function(){ for(var i = 0, marker; marker = this._markers[i]; i++){ marker.isInCluster = false; } }; /** * 把所有的标记从地图上清除 * @return 无返回值 */ MarkerClusterer.prototype._removeMarkersFromMap = function(){ for(var i = 0, marker; marker = this._markers[i]; i++){ marker.isInCluster = false; this._map.removeOverlay(marker); } }; /** * 删除单个标记 * @param {BMap.Marker} marker 需要被删除的marker * * @return {Boolean} 删除成功返回true，否则返回false */ MarkerClusterer.prototype._removeMarker = function(marker) { var index = indexOf(marker, this._markers); if (index === -1) { return false; } this._map.removeOverlay(marker); this._markers.splice(index, 1); return true; }; /** * 删除单个标记 * @param {BMap.Marker} marker 需要被删除的marker * * @return {Boolean} 删除成功返回true，否则返回false */ MarkerClusterer.prototype.removeMarker = function(marker) { var success = this._removeMarker(marker); if (success) { this._clearLastClusters(); this._createClusters(); } return success; }; /** * 删除一组标记 * @param {Array<BMap.Marker>} markers 需要被删除的marker数组 * * @return {Boolean} 删除成功返回true，否则返回false */ MarkerClusterer.prototype.removeMarkers = function(markers) { var success = false; for (var i = 0; i < markers.length; i++) { var r = this._removeMarker(markers[i]); success = success || r; } if (success) { this._clearLastClusters(); this._createClusters(); } return success; }; /** * 从地图上彻底清除所有的标记 * @return 无返回值 */ MarkerClusterer.prototype.clearMarkers = function() { this._clearLastClusters(); this._removeMarkersFromMap(); this._markers = []; }; /** * 重新生成，比如改变了属性等 * @return 无返回值 */ MarkerClusterer.prototype._redraw = function () { this._clearLastClusters(); this._createClusters(); }; /** * 获取网格大小 * @return {Number} 网格大小 */ MarkerClusterer.prototype.getGridSize = function() { return this._gridSize; }; /** * 设置网格大小 * @param {Number} size 网格大小 * @return 无返回值 */ MarkerClusterer.prototype.setGridSize = function(size) { this._gridSize = size; this._redraw(); }; /** * 获取聚合的最大缩放级别。 * @return {Number} 聚合的最大缩放级别。 */ MarkerClusterer.prototype.getMaxZoom = function() { return this._maxZoom; }; /** * 设置聚合的最大缩放级别 * @param {Number} maxZoom 聚合的最大缩放级别 * @return 无返回值 */ MarkerClusterer.prototype.setMaxZoom = function(maxZoom) { this._maxZoom = maxZoom; this._redraw(); }; /** * 获取聚合的样式风格集合 * @return {Array<IconStyle>} 聚合的样式风格集合 */ MarkerClusterer.prototype.getStyles = function() { return this._styles; }; /** * 设置聚合的样式风格集合 * @param {Array<IconStyle>} styles 样式风格数组 * @return 无返回值 */ MarkerClusterer.prototype.setStyles = function(styles) { this._styles = styles; this._redraw(); }; /** * 获取单个聚合的最小数量。 * @return {Number} 单个聚合的最小数量。 */ MarkerClusterer.prototype.getMinClusterSize = function() { return this._minClusterSize; }; /** * 设置单个聚合的最小数量。 * @param {Number} size 单个聚合的最小数量。 * @return 无返回值。 */ MarkerClusterer.prototype.setMinClusterSize = function(size) { this._minClusterSize = size; this._redraw(); }; /** * 获取单个聚合的落脚点是否是聚合内所有标记的平均中心。 * @return {Boolean} true或false。 */ MarkerClusterer.prototype.isAverageCenter = function() { return this._isAverageCenter; }; /** * 获取聚合的Map实例。 * @return {Map} Map的示例。 */ MarkerClusterer.prototype.getMap = function() { return this._map; }; /** * 获取所有的标记数组。 * @return {Array<Marker>} 标记数组。 */ MarkerClusterer.prototype.getMarkers = function() { return this._markers; }; /** * 获取聚合的总数量。 * @return {Number} 聚合的总数量。 */ MarkerClusterer.prototype.getClustersCount = function() { var count = 0; for(var i = 0, cluster; cluster = this._clusters[i]; i++){ cluster.isReal() && count++; } return count; }; /** * @ignore * Cluster * @class 表示一个聚合对象，该聚合，包含有N个标记，这N个标记组成的范围，并有予以显示在Map上的TextIconOverlay等。 * @constructor * @param {MarkerClusterer} markerClusterer 一个标记聚合器示例。 */ function Cluster(markerClusterer){ this._markerClusterer = markerClusterer; this._map = markerClusterer.getMap(); this._minClusterSize = markerClusterer.getMinClusterSize(); this._isAverageCenter = markerClusterer.isAverageCenter(); this._center = null;//落脚位置 this._markers = [];//这个Cluster中所包含的markers this._gridBounds = null;//以中心点为准，向四边扩大gridSize个像素的范围，也即网格范围 this._isReal = false; //真的是个聚合 this._clusterMarker = new BMapLib.TextIconOverlay(this._center, this._markers.length, {"styles":this._markerClusterer.getStyles()}); //this._map.addOverlay(this._clusterMarker); } /** * 向该聚合添加一个标记。 * @param {Marker} marker 要添加的标记。 * @return 无返回值。 */ Cluster.prototype.addMarker = function(marker){ if(this.isMarkerInCluster(marker)){ return false; }//也可用marker.isInCluster判断,外面判断OK，这里基本不会命中 if (!this._center){ this._center = marker.getPosition(); this.updateGridBounds();// } else { if(this._isAverageCenter){ var l = this._markers.length + 1; var lat = (this._center.lat * (l - 1) + marker.getPosition().lat) / l; var lng = (this._center.lng * (l - 1) + marker.getPosition().lng) / l; this._center = new BMap.Point(lng, lat); this.updateGridBounds(); }//计算新的Center } marker.isInCluster = true; this._markers.push(marker); var len = this._markers.length; if(len < this._minClusterSize ){ this._map.addOverlay(marker); //this.updateClusterMarker(); return true; } else if (len === this._minClusterSize) { for (var i = 0; i < len; i++) { this._markers[i].getMap() && this._map.removeOverlay(this._markers[i]); } } this._map.addOverlay(this._clusterMarker); this._isReal = true; this.updateClusterMarker(); return true; }; /** * 判断一个标记是否在该聚合中。 * @param {Marker} marker 要判断的标记。 * @return {Boolean} true或false。 */ Cluster.prototype.isMarkerInCluster= function(marker){ if (this._markers.indexOf) { return this._markers.indexOf(marker) != -1; } else { for (var i = 0, m; m = this._markers[i]; i++) { if (m === marker) { return true; } } } return false; }; /** * 判断一个标记是否在该聚合网格范围中。 * @param {Marker} marker 要判断的标记。 * @return {Boolean} true或false。 */ Cluster.prototype.isMarkerInClusterBounds = function(marker) { return this._gridBounds.containsPoint(marker.getPosition()); }; Cluster.prototype.isReal = function(marker) { return this._isReal; }; /** * 更新该聚合的网格范围。 * @return 无返回值。 */ Cluster.prototype.updateGridBounds = function() { var bounds = new BMap.Bounds(this._center, this._center); this._gridBounds = getExtendedBounds(this._map, bounds, this._markerClusterer.getGridSize()); }; /** * 更新该聚合的显示样式，也即TextIconOverlay。 * @return 无返回值。 */ Cluster.prototype.updateClusterMarker = function () { if (this._map.getZoom() > this._markerClusterer.getMaxZoom()) { this._clusterMarker && this._map.removeOverlay(this._clusterMarker); for (var i = 0, marker; marker = this._markers[i]; i++) { this._map.addOverlay(marker); } return; } if (this._markers.length < this._minClusterSize) { this._clusterMarker.hide(); return; } this._clusterMarker.setPosition(this._center); this._clusterMarker.setText(this._markers.length); var thatMap = this._map; var thatBounds = this.getBounds(); this._clusterMarker.addEventListener("click", function(event){ thatMap.setViewport(thatBounds); }); }; /** * 删除该聚合。 * @return 无返回值。 */ Cluster.prototype.remove = function(){ for (var i = 0, m; m = this._markers[i]; i++) { this._markers[i].getMap() && this._map.removeOverlay(this._markers[i]); }//清除散的标记点 this._map.removeOverlay(this._clusterMarker); this._markers.length = 0; delete this._markers; } /** * 获取该聚合所包含的所有标记的最小外接矩形的范围。 * @return {BMap.Bounds} 计算出的范围。 */ Cluster.prototype.getBounds = function() { var bounds = new BMap.Bounds(this._center,this._center); for (var i = 0, marker; marker = this._markers[i]; i++) { bounds.extend(marker.getPosition()); } return bounds; }; /** * 获取该聚合的落脚点。 * @return {BMap.Point} 该聚合的落脚点。 */ Cluster.prototype.getCenter = function() { return this._center; }; })();
这经过测试发现_addToClosestCluster耗时最严重，进一步跟踪到这里，发现是addMarker太占资源了：
跟踪到 Cluster.prototype.addMarker 找到 updateClusterMarker 这个方法
进入 Cluster.prototype.updateClusterMarker 在下图的位置加一个return; 发现效率提供了几百倍，只是地图上再也没有热点了，因为return 屏蔽了下面的绘制。
仔细看了一下这个函数发现这里有个比较严重的问题，如果有5000个点，那么就会调用5000次的addMarker，也就会执行5000次的 setPosition和 setText，这样当然会慢了。应该是在 Cluster数据准备好了一次性绘制到MAP上，例如有10个聚合也只需要绘制10次而已。
于是我加一个聚合的绘制方法：
我同时屏蔽了click事件，因为发现 setViewport 会触发多次的zoomend和moveend，导致聚合被_redraw了很多次！
Cluster.prototype.drawToMap = function(){ this._clusterMarker.setPosition(this._center); this._clusterMarker.setText(this._markers.length); return; var thatMap = this._map; var thatBounds = this.getBounds(); this._clusterMarker.addEventListener("click", function(event){ thatMap.setViewport(thatBounds); //bug! 这里多次出发 moveend _redraw }); }
这个在 MarkerClusterer.prototype._createClusters 最后循环调用一次即可：
MarkerClusterer.prototype._createClusters = function(){ log('_createClusters begin..'); var mapBounds = this._map.getBounds(); var extendedBounds = getExtendedBounds(this._map, mapBounds, this._gridSize); for(var i = 0, marker; marker = this._markers[i]; i++){ //marker 不属于任何聚合 && marker 在map视图范围内 if(!marker.isInCluster && extendedBounds.containsPoint(marker.getPosition()) ){ this._addToClosestCluster(marker); } } log('_createClusters end..'); for(var i = 0, cluster; cluster = this._clusters[i]; i++){ cluster.drawToMap(); //一起绘制到地图上 } };
修改后有个小BUG，需要同时修改下TextIconOverlay脚本，在TextIconOverlay.prototype._updateCss里加一个判断：
/** *更新相应的CSS。 *@return 无返回值。 */ TextIconOverlay.prototype._updateCss = function(){ var style = this.getStyleByText(this._text, this._styles); if(this._domElement){ this._domElement.style.cssText = this._buildCssText(style); } };
经过修改后的MarkerClusterer效率提高了不少，5000个坐标下运行还是比较流畅的，到了3w个坐标还是有点吃力。
1.8w数据
1W数据
修改后的 MarkerClusterer.js
var BMapLib = window.BMapLib = BMapLib || {}; (function() { var getExtendedBounds = function(map, bounds, gridSize) { bounds = cutBoundsInRange(bounds); var pixelNE = map.pointToPixel(bounds.getNorthEast()); var pixelSW = map.pointToPixel(bounds.getSouthWest()); pixelNE.x += gridSize; pixelNE.y -= gridSize; pixelSW.x -= gridSize; pixelSW.y += gridSize; var newNE = map.pixelToPoint(pixelNE); var newSW = map.pixelToPoint(pixelSW); return new BMap.Bounds(newSW, newNE) }; var cutBoundsInRange = function(bounds) { var maxX = getRange(bounds.getNorthEast().lng, -180, 180); var minX = getRange(bounds.getSouthWest().lng, -180, 180); var maxY = getRange(bounds.getNorthEast().lat, -74, 74); var minY = getRange(bounds.getSouthWest().lat, -74, 74); return new BMap.Bounds(new BMap.Point(minX, minY), new BMap.Point(maxX, maxY)) }; var getRange = function(i, mix, max) { mix && (i = Math.max(i, mix)); max && (i = Math.min(i, max)); return i }; var isArray = function(source) { return '[object Array]' === Object.prototype.toString.call(source) }; var indexOf = function(item, source) { var index = -1; if (isArray(source)) { if (source.indexOf) { index = source.indexOf(item) } else { for (var i = 0, m; m = source[i]; i++) { if (m === item) { index = i; break } } } } return index }; var MarkerClusterer = BMapLib.MarkerClusterer = function(map, options) { if (!map) { return } this._map = map; this._markers = []; this._clusters = []; var opts = options || {}; this._gridSize = opts["gridSize"] || 60; this._maxZoom = opts["maxZoom"] || 18; this._minClusterSize = opts["minClusterSize"] || 2; this._isAverageCenter = false; if (opts['isAverageCenter'] != undefined) { this._isAverageCenter = opts['isAverageCenter'] } this._styles = opts["styles"] || []; var that = this; this._map.addEventListener("zoomend", function() { that._redraw() }); this._map.addEventListener("moveend", function() { that._redraw() }); var mkrs = opts["markers"]; isArray(mkrs) && this.addMarkers(mkrs) }; MarkerClusterer.prototype.addMarkers = function(markers) { for (var i = 0, len = markers.length; i < len; i++) { this._pushMarkerTo(markers[i]) } this._createClusters() }; MarkerClusterer.prototype._pushMarkerTo = function(marker) { var index = indexOf(marker, this._markers); if (index === -1) { marker.isInCluster = false; this._markers.push(marker) } }; MarkerClusterer.prototype.addMarker = function(marker) { this._pushMarkerTo(marker); this._createClusters() }; MarkerClusterer.prototype._createClusters = function() { var mapBounds = this._map.getBounds(); var extendedBounds = getExtendedBounds(this._map, mapBounds, this._gridSize); for (var i = 0, marker; marker = this._markers[i]; i++) { if (!marker.isInCluster && extendedBounds.containsPoint(marker.getPosition())) { this._addToClosestCluster(marker) } } for (var i = 0, cluster; cluster = this._clusters[i]; i++) { cluster.drawToMap() } }; MarkerClusterer.prototype._addToClosestCluster = function(marker) { var distance = 4000000; var clusterToAddTo = null; var position = marker.getPosition(); for (var i = 0, cluster; cluster = this._clusters[i]; i++) { var center = cluster.getCenter(); if (center) { var d = this._map.getDistance(center, marker.getPosition()); if (d < distance) { distance = d; clusterToAddTo = cluster } } } if (clusterToAddTo && clusterToAddTo.isMarkerInClusterBounds(marker)) { clusterToAddTo.addMarker(marker) } else { var cluster = new Cluster(this); cluster.addMarker(marker); this._clusters.push(cluster) } }; MarkerClusterer.prototype._clearLastClusters = function() { for (var i = 0, cluster; cluster = this._clusters[i]; i++) { cluster.remove() } this._clusters = []; this._removeMarkersFromCluster() }; MarkerClusterer.prototype._removeMarkersFromCluster = function() { for (var i = 0, marker; marker = this._markers[i]; i++) { marker.isInCluster = false } }; MarkerClusterer.prototype._removeMarkersFromMap = function() { for (var i = 0, marker; marker = this._markers[i]; i++) { marker.isInCluster = false; this._map.removeOverlay(marker) } }; MarkerClusterer.prototype._removeMarker = function(marker) { var index = indexOf(marker, this._markers); if (index === -1) { return false } this._map.removeOverlay(marker); this._markers.splice(index, 1); return true }; MarkerClusterer.prototype.removeMarker = function(marker) { var success = this._removeMarker(marker); if (success) { this._clearLastClusters(); this._createClusters() } return success }; MarkerClusterer.prototype.removeMarkers = function(markers) { var success = false; for (var i = 0; i < markers.length; i++) { var r = this._removeMarker(markers[i]); success = success || r } if (success) { this._clearLastClusters(); this._createClusters() } return success }; MarkerClusterer.prototype.clearMarkers = function() { this._clearLastClusters(); this._removeMarkersFromMap(); this._markers = [] }; MarkerClusterer.prototype._redraw = function() { this._clearLastClusters(); this._createClusters() }; MarkerClusterer.prototype.getGridSize = function() { return this._gridSize }; MarkerClusterer.prototype.setGridSize = function(size) { this._gridSize = size; this._redraw() }; MarkerClusterer.prototype.getMaxZoom = function() { return this._maxZoom }; MarkerClusterer.prototype.setMaxZoom = function(maxZoom) { this._maxZoom = maxZoom; this._redraw() }; MarkerClusterer.prototype.getStyles = function() { return this._styles }; MarkerClusterer.prototype.setStyles = function(styles) { this._styles = styles; this._redraw() }; MarkerClusterer.prototype.getMinClusterSize = function() { return this._minClusterSize }; MarkerClusterer.prototype.setMinClusterSize = function(size) { this._minClusterSize = size; this._redraw() }; MarkerClusterer.prototype.isAverageCenter = function() { return this._isAverageCenter }; MarkerClusterer.prototype.getMap = function() { return this._map }; MarkerClusterer.prototype.getMarkers = function() { return this._markers }; MarkerClusterer.prototype.getClustersCount = function() { var count = 0; for (var i = 0, cluster; cluster = this._clusters[i]; i++) { cluster.isReal() && count++ } return count }; function Cluster(markerClusterer) { this._markerClusterer = markerClusterer; this._map = markerClusterer.getMap(); this._minClusterSize = markerClusterer.getMinClusterSize(); this._isAverageCenter = markerClusterer.isAverageCenter(); this._center = null; this._markers = []; this._gridBounds = null; this._isReal = false; this._clusterMarker = new BMapLib.TextIconOverlay(this._center, this._markers.length, { "styles": this._markerClusterer.getStyles() }) } Cluster.prototype.addMarker = function(marker) { if (this.isMarkerInCluster(marker)) { return false } if (!this._center) { this._center = marker.getPosition(); this.updateGridBounds() } else { if (this._isAverageCenter) { var l = this._markers.length + 1; var lat = (this._center.lat * (l - 1) + marker.getPosition().lat) / l; var lng = (this._center.lng * (l - 1) + marker.getPosition().lng) / l; this._center = new BMap.Point(lng, lat); this.updateGridBounds() } } marker.isInCluster = true; this._markers.push(marker); var len = this._markers.length; if (len < this._minClusterSize) { this._map.addOverlay(marker); return true } else if (len === this._minClusterSize) { for (var i = 0; i < len; i++) { this._markers[i].getMap() && this._map.removeOverlay(this._markers[i]) } } this._map.addOverlay(this._clusterMarker); this._isReal = true; this.updateClusterMarker(); return true }; Cluster.prototype.isMarkerInCluster = function(marker) { if (this._markers.indexOf) { return this._markers.indexOf(marker) != -1 } else { for (var i = 0, m; m = this._markers[i]; i++) { if (m === marker) { return true } } } return false }; Cluster.prototype.isMarkerInClusterBounds = function(marker) { return this._gridBounds.containsPoint(marker.getPosition()) }; Cluster.prototype.isReal = function(marker) { return this._isReal }; Cluster.prototype.updateGridBounds = function() { var bounds = new BMap.Bounds(this._center, this._center); this._gridBounds = getExtendedBounds(this._map, bounds, this._markerClusterer.getGridSize()) }; Cluster.prototype.drawToMap = function() { this._clusterMarker.setPosition(this._center); this._clusterMarker.setText(this._markers.length); return; var thatMap = this._map; var thatBounds = this.getBounds(); this._clusterMarker.addEventListener("click", function(event) { thatMap.setViewport(thatBounds) }) } Cluster.prototype.updateClusterMarker = function() { if (this._map.getZoom() > this._markerClusterer.getMaxZoom()) { this._clusterMarker && this._map.removeOverlay(this._clusterMarker); for (var i = 0, marker; marker = this._markers[i]; i++) { this._map.addOverlay(marker) } return } if (this._markers.length < this._minClusterSize) { this._clusterMarker.hide(); return } return; this._clusterMarker.setPosition(this._center); this._clusterMarker.setText(this._markers.length); var thatMap = this._map; var thatBounds = this.getBounds(); this._clusterMarker.addEventListener("click", function(event) { thatMap.setViewport(thatBounds) }) }; Cluster.prototype.remove = function() { for (var i = 0, m; m = this._markers[i]; i++) { this._markers[i].getMap() && this._map.removeOverlay(this._markers[i]) } this._map.removeOverlay(this._clusterMarker); this._markers.length = 0; delete this._markers } Cluster.prototype.getBounds = function() { var bounds = new BMap.Bounds(this._center, this._center); for (var i = 0, marker; marker = this._markers[i]; i++) { bounds.extend(marker.getPosition()) } return bounds }; Cluster.prototype.getCenter = function() { return this._center } })();
