Related
I've created following map using MapKit.js, with hundred of custom annotations, clustering (yellow dots) and callout popup on annotation click.
What I want to do, when clicking on the popup link, is simply to zoom in one step and center view on the clicked annotation (in a responsive context).
In Google Maps, that I'm used to, you simply position map by it's center and zoom level.
In MapKit.js, you use a center/region combo, and honestly I can't understand how this works.
Official doc is unclear to me, and I wasn't able to find really enlightling ressource.
If someone could explain to me how we are supposed to manage zoom level using center / region combo, it would be really appreciated.
Thanks :-)
[EDIT]
This center/region thing still doesn't make sense to me, so I've decided to override MapKit.js with a zoom feature.
Thanks to this post, I've manage to implement the zoom calculation, which seems to be ok.
I need now to implement the set zoom action.
No success yet, this math things are so far now ^^
Any help is highly welcomed :-)
Function:
function MapKitJsZoom(map) {
var LN2 = 0.6931471805599453; // ???
var WH = 256; // World Height
var WW = 256; // World Width
var MAX = 21; // Max zoom level
// GET CURRENT ZOOM.
var latToRad = function (lat) {
var sin = Math.sin(lat * Math.PI / 180);
var radX2 = Math.log((1 + sin) / (1 - sin)) / 2;
return Math.max(Math.min(radX2, Math.PI), -Math.PI) / 2;
};
var zoom = function (mapPx, worldPx, fraction) {
return (Math.log(mapPx / worldPx / fraction) / LN2);
};
this.get = function () {
var bounds = map.region.toBoundingRegion();
var latFraction = (latToRad(bounds.northLatitude) - latToRad(bounds.southLatitude)) / Math.PI;
var latZoom = zoom(map.element.clientHeight, WH, latFraction);
var lngDiff = bounds.eastLongitude - bounds.westLongitude;
var lngFraction = ((lngDiff < 0) ? (lngDiff + 360) : lngDiff) / 360;
var lngZoom = zoom(map.element.clientWidth, WW, lngFraction);
return Math.round(Math.min(latZoom, lngZoom, MAX));
};
// SET CURRENT ZOOM
this.set = function (zoom) {
// TODO
// I need to calculate latitude and longitude deltas
// that correspond to required zoom based on viewport size
// (map.element.clientWidth and map.element.clientHeight)
map.region.span = new mapkit.CoordinateSpan(latitudeDelta, longitudeDelta);
};
}
Usage:
var map = new mapkit.Map("map");
map.zoom = new MapKitJsZoom(map);
map.addEventListener('region-change-end', function () {
console.log(map.zoom.get());
});
There are two methods of accomplishing this:
1) set center then change zoom level
var newCenter = new mapkit.Coordinate(37.792446, -122.399360);
map._impl.zoomLevel--;
map.setCenterAnimated(newCenter, true);
2) set region using center and span (delta in degress)
var newCenter = new mapkit.Coordinate(37.792446, -122.399360);
var span = new mapkit.CoordinateSpan(.01);
var region = new mapkit.CoordinateRegion(newCenter, span);
map.setRegionAnimated(region)
I am building a map in d3 and basing it off of this codepen by Andy Barefoot: https://codepen.io/nb123456/pen/zLdqvM?editors=0010. I want to modify the initiateZoom() function so that if I set the lat/lon coordinates for a box surrounding say Ohio, the map will initialize its panning to be over Ohio.
function initiateZoom() {
minZoom = Math.max($("#map-holder").width() / w, $("#map-holder").height() / h);
maxZoom = 20 * minZoom;
zoom
.scaleExtent([minZoom, maxZoom])
.translateExtent([[0, 0], [w, h]])
;
midX = ($("#map-holder").width() - minZoom * w) / 2;
midY = ($("#map-holder").height() - minZoom * h) / 2;//These are the original values
var swlat = 32;
var swlon = -82;
var nelat = 42;
var nelon = -72;
var projectCoordinates = projection([(swlat+nelat)/2, (swlon+nelon)/2]);
/*This did not work
var midX = minZoom*(w-(swlat+nelat)/2) - ($("#map-holder").width()-(swlat+nelat)/2);
var midY = minZoom*(h-(swlon+nelon)/2) - ($("#map-holder").height()-(swlon+nelon)/2);*/
/*Neither did this
var midX = minZoom*(w-projectCoordinates[0])-($("#map-holder").width()-projectCoordinates[0]);
var midY = minZoom*(h-projectCoordinates[1])-($("#map-holder").height()-projectCoordinates[1]);*/
svg.call(zoom.transform, d3.zoomIdentity.translate(midX, midY).scale(minZoom));
}
The idea behind the original approach was to:
1: Get the current pixel display of the map
2: Get the new pixel distance from the map corner to the map point after zoom has been applied
3: The pixel distance of the center of the container to the top of the container
4: subtract the values from 2 and 3
The original post was trying to translate the map so that it would initialize the zoom and pan over the center of the map. I tried to modify this approach first by directly substituting the lat/lon values into the above equations. I also tried first transforming the lat/lon values using the projection and then substituting those values in, with little success. What do I need to do in order to get my desired result?
Setting a translateExtent could be a bad idea because it depends on the zoom scale.
The following replacement works.
function initiateZoom() {
// Define a "minzoom" whereby the "Countries" is as small possible without leaving white space at top/bottom or sides
minZoom = Math.max($("#map-holder").width() / w, $("#map-holder").height() / h);
// set max zoom to a suitable factor of this value
maxZoom = 20 * minZoom;
// set extent of zoom to chosen values
// set translate extent so that panning can't cause map to move out of viewport
zoom
.scaleExtent([minZoom, maxZoom])
.translateExtent([[0, 0], [w, h]])
;
var swlat = 32;
var swlon = -82;
var nelat = 42;
var nelon = -72;
var nwXY = projection([swlon, nelat]);
var seXY = projection([nelon, swlat]);
var zoomScale = Math.min($("#map-holder").width()/(seXY[0]-nwXY[0]), $("#map-holder").height()/(seXY[1]-nwXY[1]))
var projectCoordinates = projection([(swlon+nelon)/2, (swlat+nelat)/2]);
svg.call(zoom.transform, d3.zoomIdentity.translate($("#map-holder").width()*0.5-zoomScale*projectCoordinates[0], $("#map-holder").height()*0.5-zoomScale*projectCoordinates[1]).scale(zoomScale));
}
I'm trying to optimize a Mapbox view for long-distance hiking trails, like the Appalachian Trail or the Pacific Crest Trail. Here's an example, which I've oriented by hand, showing the Senda Pirenáica in Spain:
The area of interest, the viewport, and the pitch are given. I need to find the correct center, bearing, and zoom.
The map.fitBounds method doesn't help me here because it assumes pitch=0 and bearing=0.
I've done some poking around and this seems to be a variation of the smallest surrounding rectangle problem, but I'm stuck on a couple of additional complications:
How do I account for the distorting effect of pitch?
How do I optimize for the aspect ratio of the viewport? Note that taking the viewport narrower or wider would change the bearing of the best solution:
FWIW I'm also using turf-js, which helps me get the convex hull for the line.
This solution results in the path displayed at the correct bearing with a magenta trapezoid outline showing the target "tightest trapezoid" to show the results of the calculations. The extra line coming from the top corner shows where the map.center() value is located.
The approach is as follows:
render the path to the map using the "fitbounds" technique to get an approximate zoom level for the "north up and pitch=0" situation
rotate the pitch to the desired angle
grab the trapezoid from the canvas
This result would look like this:
After this, we want to rotate that trapezoid around the path and find the tightest fit of the trapezoid to the points. In order to test for the tightest fit it is easier to rotate the path rather than the trapezoid so I have taken that approach here. I haven't implemented a "convex hull" on the path to minimize the number of points to rotate but that is something that can be added as an optimization step.
To get the tightest fit, the first step is to move the map.center() so that the path is at the "back" of the view. This is where the most space is in the frustum so it will be easy to manipulate it there:
Next, we measure the distance between the angled trapezoid walls and each point in the path, saving the closest points on both the left and right sides. We then center the path in the view by translating the view horizontally based on these distances, and then scale the view to eliminate that space on both sides as shown by the green trapezoid below:
The scale used to get this "tightest fit" gives us our ranking for whether this is the best view of the path. However, this view may not be the best visually since we pushed the path to the back of the view to determine the ranking. Instead, we now adjust the view to place the path in the vertical center of the view, and scale the view triangle larger accordingly. This gives us the magenta colored "final" view desired:
Finally, this process is done for every degree and the minimum scale value determines the winning bearing, and we take the associated scale and center position from there.
mapboxgl.accessToken = 'pk.eyJ1IjoiZm1hY2RlZSIsImEiOiJjajJlNWMxenowNXU2MzNudmkzMndwaGI3In0.ALOYWlvpYXnlcH6sCR9MJg';
var map;
var myPath = [
[-122.48369693756104, 37.83381888486939],
[-122.48348236083984, 37.83317489144141],
[-122.48339653015138, 37.83270036637107],
[-122.48356819152832, 37.832056363179625],
[-122.48404026031496, 37.83114119107971],
[-122.48404026031496, 37.83049717427869],
[-122.48348236083984, 37.829920943955045],
[-122.48356819152832, 37.82954808664175],
[-122.48507022857666, 37.82944639795659],
[-122.48610019683838, 37.82880236636284],
[-122.48695850372314, 37.82931081282506],
[-122.48700141906738, 37.83080223556934],
[-122.48751640319824, 37.83168351665737],
[-122.48803138732912, 37.832158048267786],
[-122.48888969421387, 37.83297152392784],
[-122.48987674713133, 37.83263257682617],
[-122.49043464660643, 37.832937629287755],
[-122.49125003814696, 37.832429207817725],
[-122.49163627624512, 37.832564787218985],
[-122.49223709106445, 37.83337825839438],
[-122.49378204345702, 37.83368330777276]
];
var myPath2 = [
[-122.48369693756104, 37.83381888486939],
[-122.49378204345702, 37.83368330777276]
];
function addLayerToMap(name, points, color, width) {
map.addLayer({
"id": name,
"type": "line",
"source": {
"type": "geojson",
"data": {
"type": "Feature",
"properties": {},
"geometry": {
"type": "LineString",
"coordinates": points
}
}
},
"layout": {
"line-join": "round",
"line-cap": "round"
},
"paint": {
"line-color": color,
"line-width": width
}
});
}
function Mercator2ll(mercX, mercY) {
var rMajor = 6378137; //Equatorial Radius, WGS84
var shift = Math.PI * rMajor;
var lon = mercX / shift * 180.0;
var lat = mercY / shift * 180.0;
lat = 180 / Math.PI * (2 * Math.atan(Math.exp(lat * Math.PI / 180.0)) - Math.PI / 2.0);
return [ lon, lat ];
}
function ll2Mercator(lon, lat) {
var rMajor = 6378137; //Equatorial Radius, WGS84
var shift = Math.PI * rMajor;
var x = lon * shift / 180;
var y = Math.log(Math.tan((90 + lat) * Math.PI / 360)) / (Math.PI / 180);
y = y * shift / 180;
return [ x, y ];
}
function convertLL2Mercator(points) {
var m_points = [];
for(var i=0;i<points.length;i++) {
m_points[i] = ll2Mercator( points[i][0], points[i][1] );
}
return m_points;
}
function convertMercator2LL(m_points) {
var points = [];
for(var i=0;i<m_points.length;i++) {
points[i] = Mercator2ll( m_points[i][0], m_points[i][1] );;
}
return points;
}
function pointsTranslate(points,xoff,yoff) {
var newpoints = [];
for(var i=0;i<points.length;i++) {
newpoints[i] = [ points[i][0] + xoff, points[i][1] + yoff ];
}
return(newpoints);
}
// note [0] elements are lng [1] are lat
function getBoundingBox(arr) {
var ne = [ arr[0][0] , arr[0][1] ];
var sw = [ arr[0][0] , arr[0][1] ];
for(var i=1;i<arr.length;i++) {
if(ne[0] < arr[i][0]) ne[0] = arr[i][0];
if(ne[1] < arr[i][1]) ne[1] = arr[i][1];
if(sw[0] > arr[i][0]) sw[0] = arr[i][0];
if(sw[1] > arr[i][1]) sw[1] = arr[i][1];
}
return( [ sw, ne ] );
}
function pointsRotate(points, cx, cy, angle){
var radians = angle * Math.PI / 180.0;
var cos = Math.cos(radians);
var sin = Math.sin(radians);
var newpoints = [];
function rotate(x, y) {
var nx = cx + (cos * (x - cx)) + (-sin * (y - cy));
var ny = cy + (cos * (y - cy)) + (sin * (x - cx));
return [nx, ny];
}
for(var i=0;i<points.length;i++) {
newpoints[i] = rotate(points[i][0],points[i][1]);
}
return(newpoints);
}
function convertTrapezoidToPath(trap) {
return([
[trap.Tl.lng, trap.Tl.lat], [trap.Tr.lng, trap.Tr.lat],
[trap.Br.lng, trap.Br.lat], [trap.Bl.lng, trap.Bl.lat],
[trap.Tl.lng, trap.Tl.lat] ]);
}
function getViewTrapezoid() {
var canvas = map.getCanvas();
var trap = {};
trap.Tl = map.unproject([0,0]);
trap.Tr = map.unproject([canvas.offsetWidth,0]);
trap.Br = map.unproject([canvas.offsetWidth,canvas.offsetHeight]);
trap.Bl = map.unproject([0,canvas.offsetHeight]);
return(trap);
}
function pointsScale(points,cx,cy, scale) {
var newpoints = []
for(var i=0;i<points.length;i++) {
newpoints[i] = [ cx + (points[i][0]-cx)*scale, cy + (points[i][1]-cy)*scale ];
}
return(newpoints);
}
var id = 1000;
function convertMercator2LLAndDraw(m_points, color, thickness) {
var newpoints = convertMercator2LL(m_points);
addLayerToMap("id"+id++, newpoints, color, thickness);
}
function pointsInTrapezoid(points,yt,yb,xtl,xtr,xbl,xbr) {
var str = "";
var xleft = xtr;
var xright = xtl;
var yh = yt-yb;
var sloperight = (xtr-xbr)/yh;
var slopeleft = (xbl-xtl)/yh;
var flag = true;
var leftdiff = xtr - xtl;
var rightdiff = xtl - xtr;
var tmp = [ [xtl, yt], [xtr, yt], [xbr,yb], [xbl,yb], [xtl,yt] ];
// convertMercator2LLAndDraw(tmp, '#ff0', 2);
function pointInTrapezoid(x,y) {
var xsloperight = xbr + sloperight * (y-yb);
var xslopeleft = xbl - slopeleft * (y-yb);
if((x - xsloperight) > rightdiff) {
rightdiff = x - xsloperight;
xright = x;
}
if((x - xslopeleft) < leftdiff) {
leftdiff = x - xslopeleft;
xleft = x;
}
if( (y<yb) || (y > yt) ) {
console.log("y issue");
}
else if(xsloperight < x) {
console.log("sloperight");
}
else if(xslopeleft > x) {
console.log("slopeleft");
}
else return(true);
return(false);
}
for(var i=0;i<points.length;i++) {
if(pointInTrapezoid(points[i][0],points[i][1])) {
str += "1";
}
else {
str += "0";
flag = false;
}
}
if(flag == false) console.log(str);
return({ leftdiff: leftdiff, rightdiff: rightdiff });
}
var viewcnt = 0;
function calculateView(trap, points, center) {
var bbox = getBoundingBox(points);
var bbox_height = Math.abs(bbox[0][1] - bbox[1][1]);
var view = {};
// move the view trapezoid so the path is at the far edge of the view
var viewTop = trap[0][1];
var pointsTop = bbox[1][1];
var yoff = -(viewTop - pointsTop);
var extents = pointsInTrapezoid(points,trap[0][1]+yoff,trap[3][1]+yoff,trap[0][0],trap[1][0],trap[3][0],trap[2][0]);
// center the view trapezoid horizontally around the path
var mid = (extents.leftdiff - extents.rightdiff) / 2;
var trap2 = pointsTranslate(trap,extents.leftdiff-mid,yoff);
view.cx = trap2[5][0];
view.cy = trap2[5][1];
var w = trap[1][0] - trap[0][0];
var h = trap[1][1] - trap[3][1];
// calculate the scale to fit the trapezoid to the path
view.scale = (w-mid*2)/w;
if(bbox_height > h*view.scale) {
// if the path is taller than the trapezoid then we need to make it larger
view.scale = bbox_height / h;
}
view.ranking = view.scale;
var trap3 = pointsScale(trap2,(trap2[0][0]+trap2[1][0])/2,trap2[0][1],view.scale);
w = trap3[1][0] - trap3[0][0];
h = trap3[1][1] - trap3[3][1];
view.cx = trap3[5][0];
view.cy = trap3[5][1];
// if the path is not as tall as the view then we should center it vertically for the best looking result
// this involves both a scale and a translate
if(h > bbox_height) {
var space = h - bbox_height;
var scale_mul = (h+space)/h;
view.scale = scale_mul * view.scale;
cy_offset = space/2;
trap3 = pointsScale(trap3,view.cx,view.cy,scale_mul);
trap3 = pointsTranslate(trap3,0,cy_offset);
view.cy = trap3[5][1];
}
return(view);
}
function thenCalculateOptimalView(path) {
var center = map.getCenter();
var trapezoid = getViewTrapezoid();
var trapezoid_path = convertTrapezoidToPath(trapezoid);
trapezoid_path[5] = [center.lng, center.lat];
var view = {};
//addLayerToMap("start", trapezoid_path, '#00F', 2);
// get the mercator versions of the points so that we can use them for rotations
var m_center = ll2Mercator(center.lng,center.lat);
var m_path = convertLL2Mercator(path);
var m_trapezoid_path = convertLL2Mercator(trapezoid_path);
// try all angles to see which fits best
for(var angle=0;angle<360;angle+=1) {
var m_newpoints = pointsRotate(m_path, m_center[0], m_center[1], angle);
var thisview = calculateView(m_trapezoid_path, m_newpoints, m_center);
if(!view.hasOwnProperty('ranking') || (view.ranking > thisview.ranking)) {
view.scale = thisview.scale;
view.cx = thisview.cx;
view.cy = thisview.cy;
view.angle = angle;
view.ranking = thisview.ranking;
}
}
// need the distance for the (cx, cy) from the current north up position
var cx_offset = view.cx - m_center[0];
var cy_offset = view.cy - m_center[1];
var rotated_offset = pointsRotate([[cx_offset,cy_offset]],0,0,-view.angle);
map.flyTo({ bearing: view.angle, speed:0.00001 });
// once bearing is set, adjust to tightest fit
waitForMapMoveCompletion(function () {
var center2 = map.getCenter();
var m_center2 = ll2Mercator(center2.lng,center2.lat);
m_center2[0] += rotated_offset[0][0];
m_center2[1] += rotated_offset[0][1];
var ll_center2 = Mercator2ll(m_center2[0],m_center2[1]);
map.easeTo({
center:[ll_center2[0],ll_center2[1]],
zoom : map.getZoom() });
console.log("bearing:"+view.angle+ " scale:"+view.scale+" center: ("+ll_center2[0]+","+ll_center2[1]+")");
// draw the tight fitting trapezoid for reference purposes
var m_trapR = pointsRotate(m_trapezoid_path,m_center[0],m_center[1],-view.angle);
var m_trapRS = pointsScale(m_trapR,m_center[0],m_center[1],view.scale);
var m_trapRST = pointsTranslate(m_trapRS,m_center2[0]-m_center[0],m_center2[1]-m_center[1]);
convertMercator2LLAndDraw(m_trapRST,'#f0f',4);
});
}
function waitForMapMoveCompletion(func) {
if(map.isMoving())
setTimeout(function() { waitForMapMoveCompletion(func); },250);
else
func();
}
function thenSetPitch(path,pitch) {
map.flyTo({ pitch:pitch } );
waitForMapMoveCompletion(function() { thenCalculateOptimalView(path); })
}
function displayFittedView(path,pitch) {
var bbox = getBoundingBox(path);
var path_cx = (bbox[0][0]+bbox[1][0])/2;
var path_cy = (bbox[0][1]+bbox[1][1])/2;
// start with a 'north up' view
map = new mapboxgl.Map({
container: 'map',
style: 'mapbox://styles/mapbox/streets-v9',
center: [path_cx, path_cy],
zoom: 12
});
// use the bounding box to get into the right zoom range
map.on('load', function () {
addLayerToMap("path",path,'#888',8);
map.fitBounds(bbox);
waitForMapMoveCompletion(function() { thenSetPitch(path,pitch); });
});
}
window.onload = function(e) {
displayFittedView(myPath,60);
}
body { margin:0; padding:0; }
#map { position:absolute; top:0; bottom:0; width:100%; }
<script src='https://api.tiles.mapbox.com/mapbox-gl-js/v0.37.0/mapbox-gl.js'></script>
<link href='https://api.tiles.mapbox.com/mapbox-gl-js/v0.37.0/mapbox-gl.css' rel='stylesheet' />
<div id='map'></div>
Smallest surrounding rectangle would be specific to pitch=0 (looking directly down).
One option is to continue with smallest surrounding rectangle approach and calculate the transformation of the target area - just like a 3d engine does. If this is what you do maybe skim through unity docs to better understand the mechanics of viewing frustum
I feel this wouldn't be appropriate for your problem though as you'd have to re-calculate a 2d rendering of the target area from different angles, a relatively expensive brute force.
Another way to normalize the calculation would be to render a viewport projection into target area plane. See for yourself:
Then all you have to do is "just" figure out the largest size your original convex hull can fit into a trapezoid of that shape (specifically a convex isosceles trapezoid since we don't manipulate camera roll).
This is where I get a little out of depth and don't know where to point you for a calculation. I figure it's at least cheaper to iterate over possible solutions in this 2D space though.
P.S: One more thing to keep in mind is the viewport projection shape will be different depending on FOV (field of view).
This changes when you resize the browser viewport, but the property doesn't seem to be exposed in mapbox-gl-js.
Edit:
After some thought I feel the best mathematical solution can feel a little "dry" in reality. Not being across the use case and, possibly, making some wrong assumptions, I'd ask these questions:
For a route that's roughly a straight line, would it always be panned in so the ends are at bottom left and top right corners? That would be close to "optimal" but could get... boring.
Would you want to keep more of the path closer to the viewport? You can lose route detail if a large portion of it is far away from the viewport.
Would you pick points of interest to focus on? Those could be closer to the viewport.
Perhaps it would be handy to classify different types of routes by shape of hull and create panning presets?
Hopefully this can point you in the right direction with some tweaking.
First I set up the two points we want to show
let pointA = [-70, 43]
let pointB = [-83, 32]
Then I found the middle of those two points. I made my own function for this, but it looks like turf can do this.
function middleCoord(a, b){
let x = (a - b)/2
return _.min([a, b]) + x
}
let center = [middleCoord(pointA[0], pointB[0]), middleCoord(pointA[1], pointB[1])]
I used turfs bearing function to have the view from the 2nd point look at the first point
let p1 = turf.point(pointA)
let p2 = turf.point(pointB)
let points = turf.featureCollection([p1, p2])
let bearing = turf.bearing(p2, p1)
Then I call the map and run the fitBounds function:
var map = new mapboxgl.Map({
container: 'map', // container id
style: 'mapbox://styles/mapbox/outdoors-v10', //hosted style id
center: center, // starting position
zoom: 4, // starting zoom
pitch: 60,
bearing: bearing
})
map.fitBounds([pointA, pointB], {padding: 0, offset: 0})
Here's a codepen: https://codepen.io/thejoshderocher/pen/BRYGXq
To adjust the bearing to best use the screen size is to get the size of the window and adjust the bearing to take the most advantage of the available screen space. If it's a mobile screen in portrait, this bearing work perfect. If you are on a desktop with a wide view you will need to rotate so point A is in one of the top corners.
I'd like to help the user to input an orientation for a segment with OpenLayers.
I have that form where user can input the bearing for a point, but I would like to help him by :
start drawing the first vertice of a segment on the map when the user clicks on a button, (that first vertice being a known point)
then the user just has to click for the second vertice, and bearing is computed automatically.
See the fiddle here or SO snippet below.
I'm almost done : I can compute the bearing when a segment is drawn. But there's an exception at the very end of the script : I can't get OL to draw automatically the first point of my segment.
Thank you to anyone who can help.
<script src="http://openlayers.org/api/OpenLayers.js"></script>
<body>
<div id="map" style="height: 500px"></div>
</body>
<script>
var CONSTANTS = {
MAP_FROM_PROJECTION: new OpenLayers.Projection("EPSG:4326"), // Transform from WGS 1984
MAP_TO_PROJECTION: new OpenLayers.Projection("EPSG:900913") // to Spherical Mercator Projection
};
function radians(n) {
return n * (Math.PI / 180);
}
function degrees(n) {
return n * (180 / Math.PI);
}
function computeBearing(startLat, startLong, endLat, endLong) {
startLat = radians(startLat);
startLong = radians(startLong);
endLat = radians(endLat);
endLong = radians(endLong);
var dLong = endLong - startLong;
var dPhi = Math.log(Math.tan(endLat / 2.0 + Math.PI / 4.0) / Math.tan(startLat / 2.0 + Math.PI / 4.0));
if (Math.abs(dLong) > Math.PI) {
if (dLong > 0.0) dLong = -(2.0 * Math.PI - dLong);
else dLong = (2.0 * Math.PI + dLong);
}
return (degrees(Math.atan2(dLong, dPhi)) + 360.0) % 360.0;
}
map = new OpenLayers.Map("map");
map.addLayer(new OpenLayers.Layer.OSM());
map.setCenter(new OpenLayers.LonLat(3, 47).transform(CONSTANTS.MAP_FROM_PROJECTION, CONSTANTS.MAP_TO_PROJECTION), 6);
var lineLayer = new OpenLayers.Layer.Vector("Line Layer");
map.addLayers([lineLayer]);
var lineControl = new OpenLayers.Control.DrawFeature(lineLayer, OpenLayers.Handler.Path, {
handlerOptions: {
maxVertices: 2,
freehandMode: function(evt) {
return false;
}
},
featureAdded: function(feature) {
var drawnLinePoints = feature.geometry.getVertices();
var lonlat1 = drawnLinePoints[0].transform(CONSTANTS.MAP_TO_PROJECTION, CONSTANTS.MAP_FROM_PROJECTION);
var lonlat2 = drawnLinePoints[1].transform(CONSTANTS.MAP_TO_PROJECTION, CONSTANTS.MAP_FROM_PROJECTION);
var bearingValue = computeBearing(lonlat1.y, lonlat1.x, lonlat2.y, lonlat2.x);
console.log(bearingValue);
}
});
map.addControl(lineControl);
lineControl.activate();
var handler;
for (var i = 0; i < map.controls.length; i++) {
var control = map.controls[i];
if (control.displayClass === "olControlDrawFeature") {
handler = control.handler;
break;
}
}
// Here I have an exception in the console : I would like
// OL to draw hat point automatically.
handler.addPoint(new OpenLayers.Pixel(50, 50));
</script>
OpenLayers.Handler.Path.addPoint works on OpenLayers.Pixel, not OpenLayers.LonLat:
/**
* Method: addPoint
* Add point to geometry. Send the point index to override
* the behavior of LinearRing that disregards adding duplicate points.
*
* Parameters:
* pixel - {<OpenLayers.Pixel>} The pixel location for the new point.
*/
addPoint: function(pixel) {
this.layer.removeFeatures([this.point]);
var lonlat = this.layer.getLonLatFromViewPortPx(pixel);
this.point = new OpenLayers.Feature.Vector(
new OpenLayers.Geometry.Point(lonlat.lon, lonlat.lat)
);
this.line.geometry.addComponent(
this.point.geometry, this.line.geometry.components.length
);
this.layer.addFeatures([this.point]);
this.callback("point", [this.point.geometry, this.getGeometry()]);
this.callback("modify", [this.point.geometry, this.getSketch()]);
this.drawFeature();
delete this.redoStack;
}
I actually see no good way of achieving this other than adding an addPointByLonLat method:
OpenLayers.Handler.Path.prototype.addPointByLonLat = function(lonLat) {
this.layer.removeFeatures([this.point]);
this.point = new OpenLayers.Feature.Vector(
new OpenLayers.Geometry.Point(lonlat.lon, lonlat.lat)
);
this.line.geometry.addComponent(
this.point.geometry, this.line.geometry.components.length
);
this.layer.addFeatures([this.point]);
this.callback("point", [this.point.geometry, this.getGeometry()]);
this.callback("modify", [this.point.geometry, this.getSketch()]);
this.drawFeature();
delete this.redoStack;
};
Or subclass as your own handler class (propbably cleaner).
Notes:
addPoint is not an API method (so addPointByLonLat is also not). This may result in problem on version changes.
Don't use the compressed/minified JS in development and check docs on methods you use.
Next time consider asking on https://gis.stackexchange.com/.
Consider asking for a code review on your JS.
You can also use insertXY(x,y) function in order to insert a point with geographic coordinates
http://dev.openlayers.org/docs/files/OpenLayers/Handler/Path-js.html#OpenLayers.Handler.Path.insertXY
lonlat = new OpenLayers.LonLat(1,45);
lonlat.transform(CONSTANTS.MAP_FROM_PROJECTION,CONSTANTS.MAP_TO_PROJECTION);
handler.createFeature(new OpenLayers.Pixel(100, 100));
handler.insertXY(lonlat.lon,lonlat.lat);
handler.drawFeature();
You can check it here with a fork of your original jsfiddle : http://jsfiddle.net/mefnpbn2/
See this fiddle for the solution.
tldr :
// draw the first point as I needed, as if a user has clicked on the map
handler.modifyFeature(new OpenLayers.Pixel(50, 50), true);
// draw a first point on the map (not clicked).
// This will be the initial point where the "cursor" is on the map, as long as
// the user hasn't hovered onto the map with its mouse. This make the blue
// line showing current segment to appear, without this segment is drawn but
// no feedback is given to the user as long as he hasn't clicked.
handler.addPoint(new OpenLayers.Pixel(50, 50)); //
I have a set of latitude and longitude coordinates in an array. When plotted as points on a Polygon in Google Maps, they form roughly an irregular oval pointing from west to east.
I would like to rotate the polygon to arbitrary degrees with a point near the far left (western) axis as the origin point.
What's important to me is that the overall distances are preserved — the total length and width of the polygon in miles should be preserved even though size of the polygon in pixels will obviously vary because of the map projection.
I've spent many hours Googling and searching on this site for an answer to this question but haven't been able to find one. Strictly speaking this is not a Google Maps issue — it's just a geometry issue related to rotating polygons on a map projection.
Here is more or less (some things dropped/renamed for simplicity) what I've been playing with so far:
function transpose_poly() {
//the polygon I'm transposing; these are actually in lng,lat not lat,lng
var poly = [
[165.2467094000077,11.90511591102683],[165.1960646350092,11.8776472999036],[165.163749733589,11.84385698549516],[165.1260503052001,11.79105209544025],[165.1174590975139,11.73721064669002],[165.1239723221977,11.68443896599666],[165.1455712453083,11.62322111902237],[165.1628352566873,11.60212814234246],[165.2801792409835,11.57112282455308],[165.3469838984795,11.55321856612457],[165.4267372080734,11.49929306693537],[165.565122175408,11.43334434041074],[165.7036021721537,11.37198352732909],[165.7982543390455,11.32757670668951],[165.942446703552,11.2660302522167],[166.0435044916007,11.23207374453692],[166.1628753311194,11.19161490039798],[166.3468035041342,11.13537294275959],[166.432786256031,11.11077339082378],[166.6460225244011,11.09603091173615],[166.9054486129032,11.10353634871533],[167.0953310801652,11.11920326808891],[167.2738338244123,11.14546202299651],[167.5299835821322,11.20017441185735],[167.7587090824888,11.25007287877568],[168.0532186132958,11.32987818697488],[168.3030699093596,11.40339603540862],[168.592055474493,11.49329084618948],[168.8894586866613,11.59767488596071],[169.1097084341002,11.70426500697907],[169.3388671138959,11.8464629880637],[169.47335151263,11.96284699062962],[169.4987805640997,12.00051052731504],[169.5113979458664,12.04997756596092],[169.496674063518,12.07975001861134],[169.4439862794831,12.10473302818016],[169.3792705121883,12.11718325976015],[169.2053586392944,12.12366910168141],[169.0210976722354,12.12171866909852],[168.7390558752391,12.08703266811138],[168.4733370821476,12.04764814638675],[168.1055698159765,12.00021651042535],[167.8745488025422,11.97152786285725],[167.5955303201492,11.94724207538445],[167.1571321566584,11.94152529858467],[166.8673995936747,11.95771709621411],[166.6698153277294,11.98911065050636],[166.4432968316392,12.03361885637251],[166.2604579582592,12.0693923391982],[166.0834945953367,12.09610600014998],[165.9206278637858,12.10688793842689],[165.7421018997999,12.1070118835482],[165.6244752224984,12.09489163127243],[165.4939687494391,12.0642833194958],[165.4055155587918,12.02521842289156],[165.3386147079753,11.97769336095395],[165.278157772496,11.92168821285675],[165.2467094000077,11.90511591102683],
];
var pos = marker.getPosition(); //where it transposes it to
var marker_lat = pos.lat();
var marker_lng = pos.lng();
var angle = document.getElementById("poly_angle").value; //rotational angle
var original_lat = 11.697222; //the original lat/lng of the polygon coordinates above
var original_lng = 165.27194399999996; //used to move the general polygon to a new set of coords as an offset
var new_poly = []; //the transformed polygon
//iterate over polygon array, create new array of Google Maps LatLng objects that are transposed and rotated
for(var i=0; i<poly.length; i++) {
new_poly.push( rotateLatLng( (poly[i][1]-original_lat)+marker_lat,(poly[i][0].lng()-original_lng)+marker_lng,angle));
}
// plot the polygon on the map
poly_obj = new google.maps.Polygon({
paths: poly_new,
strokeColor: "#ff763b",
strokeOpacity: 1,
strokeWeight: 1,
fillColor: "#ff763b",
fillOpacity: 0.25,
map: map,
visible: true
});
}
//this rotating function is cobbled together from code I found... it doesn't really work. It somewhat rotates it but distorts it terribly and gives really bizarre results
function rotateLatLng (pointLat,pointLng,angle) {
var pos = marker.getPosition();
var theX = pointLat;
var theY = pointLng;
var rotationTheta = angle;
var rotationThetaRad = rotationTheta*(Math.PI/180);
var rotationOriginX = pos.lat();
var rotationOriginY = pos.lng();
var newX;
var newY;
if (rotationOriginX == 0 && rotationOriginY == 0) {
newX = theX * Math.cos(rotationThetaRad) - Math.sin(rotationThetaRad) * theY;
newY = theX * Math.sin(rotationThetaRad) + Math.cos(rotationThetaRad) * theY;
} else {
newX = (theX - rotationOriginX) * Math.cos(rotationThetaRad) - (theY - rotationOriginY) * Math.sin(rotationTheta) + rotationOriginX;
newY = (theX - rotationOriginX) * Math.sin(rotationThetaRad) + (theY - rotationOriginY) * Math.cos(rotationTheta) + rotationOriginY;
}
return new google.maps.LatLng(newX,newY);
}
I don't think the above is necessarily the right way to do it at all. Any pointers would be helpful.
A much more simplified version of this problem would be to say that if I had an origin point of lat1,lng1, and another target point of lat2,lng2, how do I calculate lat3,lng3 which is defined as being the same distance between lat1,lng1 and lat2,lng2 but at an arbitrary angle? Because if I knew how to do that, applying that to the entire polygon should be a snap.