I want to use d3 for the next task:
display rotating globe with donut chart in center of every country. It should be possible to interact with globe (select country, zoom, rotate).
Seems d3 provide an easy way to implement every part of it but I can not get donuts part working as I need.
There is an easy way draw donut chart with the help of d3.arc:
var arc = d3.arc();
var data = [3, 23, 17, 35, 4];
var radius = 15/scale;
var _arc = arc.innerRadius(radius - 7/scale)
.outerRadius(radius).context(donutsContext);
var pieData = pie(data);
for (var i = 0; i < pieData.length; i++) {
donutsContext.beginPath();
donutsContext.fillStyle = color(i);
_arc(pieData[i]);
}
by with code as it is donuts are displayed on a plane on top of the globe, like:
globe with donut
while I want them to be 'wrapped' around the globe
There is d3.geoCircle method that can be projected to globe correctly. I got 'ring' projected correctly to the globe with the help of two circles:
var circle = d3.geoCircle()
.center(centroid)
.radius(2);
var outerCircle = circle();
var circle = d3.geoCircle()
.center(centroid)
.radius(1);
var innerCircle = circle();
var interCircleCoordinates = [];
for (var i = innerCircle.coordinates[0].length - 1; i >= 0; i--) {
interCircleCoordinates.push(innerCircle.coordinates[0][i]);
}
outerCircle.coordinates.push(interCircleCoordinates);
globe with rings
but I really need to get a donut.
The other way I tried is getting image from donuts and wrapping this image around globe with the help of pixels manipulation:
var image = new Image;
image.onload = onload;
image.src = img;
function onload() {
window.dx = image.width;
window.dy = image.height;
context.drawImage(image, 0, 0, dx, dy);
sourceData = context.getImageData(0, 0, dx, dy).data;
target = context.createImageData(width, height);
targetData = target.data;
for (var y = 0, i = -1; y < height; ++y) {
for (var x = 0; x < width; ++x) {
var p = projection.invert([x, y]), λ = p[0], φ = p[1];
if (λ > 180 || λ < -180 || φ > 90 || φ < -90) { i += 4; continue; }
var q = ((90 - φ) / 180 * dy | 0) * dx + ((180 + λ) / 360 * dx | 0) << 2;
var r = sourceData[q];
var g = sourceData[++q];
var b = sourceData[++q];
targetData[++i] = r;
targetData[++i] = g;
targetData[++i] = b;
targetData[++i] = 125;//
}
}
context.clearRect(0,0, width, height);
context.putImageData(target, 0, 0);
};
by this way I get extremely slow rotating and interaction with a globe for a globe size I need (1000px)
So my questions are:
Is there is some way to project donuts that are generated with the help of d3.arc to a sphere (globe, orthographic projection)?
Is there is some way to get a donut from geoCircle?
Maybe there is some other way to achieve my goal I do not see
There is one way that comes to mind to display donuts on a globe. The key challenge is that d3 doesn't project three dimensional objects very well - with one exception, geographic features. Consequently, an "easy" solution is to convert your pie charts into geographic features and project them with the rest of your features.
To do this you need to:
Use a pie/donut generator as you normally would
Go along the paths generated to get points approximating the pie shape.
Convert the points to long/lat points
Assemble those points into geojson
Project them onto the map.
The first point is easy enough, just make a pie chart with an inner radius.
Now you have to select each path and find points along its perimeter using path.getPointAtLength(), this will be dependent on path length, so path.getTotalLength() will be handy (and corners are important, so you might want to incorporate a little bit of complexity for these corner cases to ensure you get them)).
Once you have the points, you need the use of a second projection, azimuthal equidistant would be best. If the pie chart is centered on [0,0] in svg coordinate space, rotate the azimuthal (don't center), so that the centroid coordinate is located at [0,0] in svg space (you can use translates on the pies to position them, but it will just add extra steps). Take each point and run it through projection.invert() using the second projection. You will need to update the projection for each donut chart as each one will have a different geographic centroid.
Once you have lat long points, it's easy - you've already done it with the geo circle function - convert to geojson and project with the orthographic projection.
This approach gave me something like:
Notes: Depending on your data, it might be easiest to preprocess your data into geojson and store that as opposed to calculating the geojson each page load.
You are using canvas, while you don't need to actually use an svg, you need to still be able to access svg functions like getPointAtLength, you do not need to have an svg or display svg elements by using a custom element replicating a path :
document.createElementNS(d3.namespaces.svg, 'path');
Oh, and make sure the second projection's translate is set - the default is [480,250] for all (most?) d3 projections, that will throw things off if unaccounted for.
Related
Following my post last week three.js How to programatically produce a plane from dataset I come back to the community to solve a problem of definition of surface occupied on the ground by a 3D building.
The solution proposed in comments in this post works for this building but is not universal.
To make it universal I chose the following method: when the walls are built I create their clone in another group (see this previous post for walls creation)
// prepare the clones
var clones = new THREE.Group();
scene.add(clones);
var num=0;
// drawing the real walls
var wallGeometry = new THREE.PlaneGeometry(size,(hstair*batims[i][1]));
val = 0xFFFFFF;
opa = 0.5;
if(deltaX > deltaY){val = 0x000000; opa = 0.05;} // shaded wall
var wallMaterial = new THREE.MeshBasicMaterial({color:val,transparent:true, opacity:opa, side:THREE.DoubleSide});
var walls = new THREE.Mesh(wallGeometry, wallMaterial);
walls.position.set((startleft+endleft)/2,(hstair*batims[i][1])/2,(startop+endtop)/2);
walls.rotation.y = -rads;
scene.add(walls);
// add the pseudo-walls to scene
var cloneGeometry=new THREE.PlaneGeometry(long,3);
var cloneMaterial=new THREE.MeshBasicMaterial({color:0xff0000,transparent:true,opacity:0.5,side:THREE.DoubleSide});
var clone=new THREE.Mesh(pseudomursGeometry,pseudomursMaterial);
clone.position.set((startleft+endleft)/2,3,(startop+endtop)/2);
clone.rotation.y=-rads;
clones.add(clone);
num++;
The idea is now to rotate this pseudo-building so that the longest wall is vertical, which allows me to determine the exact floor area occupied with its boundingBox:
var angle=turn=0;
for(i=0; i<dists.length; i++) { // dists is the array of wall lengths
if(dists[i]==longs[0]){ // longs is the reordered lengths array
angle=angles[i][1]; // angle of the longest wall
}
}
// we can now rotate the whole group to put the longest wall vertical
if(angle>0){
turn = angle*-1+(Math.PI/2);
}
else {
turn = angle+(Math.PI/2);
}
clones.rotation.y=turn;
It works perfectly as long as the building has a right angle, whatever its shape: triangle, rectangle, bevel, right angle polygons,
var boundingBox = new THREE.Box3().setFromObject(clones);
var thisarea = boundingBox.getSize();
// area size gives the expected result
console.log('AREA SIZE = '+thisarea.x+' '+thisarea.y+' '+thisarea.z);
...but not when there are no more right angles, for example a trapezoid
The reason is that we rotate the group, and not the cloned walls. I can access and rotate each wall by
for(n=0;n<num;n++){
thisangle = clones.children[n].rotation.y;
clones.children[n].rotation.y = turn-thisangle;
}
But the result is wrong for the others pseudo-walls:
So the question is: how to turn each red pseudo-wall so that the longest one is vertical and the others remain correctly positioned in relation to it? In this way, any building with any shape can be reproduced in 3D with its internal equipment. Any idea on how to achieve this result?
A weird & ugly but well-working solution:
// 1. determines which is the longest side
for(i=0; i<dists.length; i++) {
if(dists[i]==longs[0]){
longest=i;
break; // avoid 2 values if rectangle
}
}
// 2. the group is rotated until the longest side has an angle in degrees
// close to 0 or 180
var letsturn = setInterval(function() {
clones.rotation.y += 0.01;
var group_rotation = THREE.Math.radToDeg(clones.rotation.y); // degrees
var stop = Math.round(angles[longest][0] - group_rotation);
// 3. stop when longest wall is vertical
if( (stop>=179 && stop<=181) || (stop>=-1 && stop<=1) ) {
clearInterval(letsturn);
createPlane() // we can now use boundingBox in reliability
}
}, 1);
et voilà.
Matrix operations performed on the GPU can be pretty hard to debug because GPU operations don't really allow for console logs.
I've written one designed for a real time 2D rendering engine based on a very simple form of I guess what could be called ray casting and am having trouble figuring out what's wrong with it (it's outputting [0,0,0,255,0,0,0,255,...] instead of populating colors).
this.thread.x is the index of the current unit (color channel) in the matrix being operated on.
scene is a buffer made up of 6-unit clumps, each value containing, in order:
The type of entity, always 1 for "sprite" in this case.
The sprite ID, corresponding the the index in this.constants.textures containing the buffer for the entity's sprite.
X offset, the left edge of the sprite
Y offset, the top edge of the sprite
width of the sprite
height of the sprite
bufferWidth is the width of the render area multiplied by 4 channels.
this.constants.textures is an array containing buffers of each sprite which the sprite IDs from the scene refer to.
Note: For those curious, this is being done with GPU.js, a JavaScript lib that converts a JS func into GLSL code to be run via WebGL.
function(scene, sceneLength, bufferWidth) {
var channel = this.thread.x % 4;
if (channel === 3) {
return 255;
}
var x = this.thread.x % bufferWidth;
var y = Math.floor(this.thread.x / bufferWidth);
for (let i1 = 0; i1 < sceneLength; i1 += 6) {
var id = scene[i1 + 1];
var x1 = scene[i1 + 2];
var y1 = scene[i1 + 3];
var w1 = scene[i1 + 4];
var h1 = scene[i1 + 5];
var r1 = scene[i1 + 6];
var offsetX1 = x1 - x;
if (offsetX1 > 0 && offsetX1 < w1) {
var offsetY1 = y1 - y;
if (offsetY1 > 0 && offsetY1 < h1) {
var c1 = offsetY1 * w1 * 4 + offsetX1 * 4;
var c1R = c1 - (c1 % 4);
var c1A = c1R + 3;
if (this.constants.textures[id][c1A] != 0) {
return this.constants.textures[id][c1];
}
}
}
}
return 0;
}
Explanation for the concept I'm trying to implement:
With a matrix operation, when you want to draw a sprite if you were to perform a pass on the entire render area, you'd be doing far more work than necessary. If you break the rendering area down into chunks and only update the sections involved in the sprite being drawn, that would be a fairly decent way to do it. It would certainly be good enough for real time game rendering. This would be a multi-pass approach, where sprites are rendered one at a time.
Alternatively, for what seems to me to be the most optimal approach possible, instead of that, we can use a single-pass approach that performs a single matrix operation for the entire rendering area, evaluating for each color channel what should be there based on doing a very basic form of collision detection with each sprite in the scene and the relevant pixel in that sprite.
You're calculating your sprite offsets backwards, the calculations should be:
var offsetX1 = x - x1;
and
var offsetY1 = y - y1;
The offsets should increase as x and y increase (assuming the sprite co-ordinates have the same co-ordinate system as the screen co-ordinates), so you shouldn't be subtracting x and y.
I need to convert the position and rotation on a 3d object to screen position and rotation. I can convert the position easily but not the rotation. I've attempted to convert the rotation of the camera but it does not match up.
Attached is an example plunkr & conversion code.
The white facebook button should line up with the red plane.
https://plnkr.co/edit/0MOKrc1lc2Bqw1MMZnZV?p=preview
function toScreenPosition(position, camera, width, height) {
var p = new THREE.Vector3(position.x, position.y, position.z);
var vector = p.project(camera);
vector.x = (vector.x + 1) / 2 * width;
vector.y = -(vector.y - 1) / 2 * height;
return vector;
}
function updateScreenElements() {
var btn = document.querySelector('#btn-share')
var pos = plane.getWorldPosition();
var vec = toScreenPosition(pos, camera, canvas.width, canvas.height);
var translate = "translate3d("+vec.x+"px,"+vec.y+"px,"+vec.z+"px)";
var euler = camera.getWorldRotation();
var rotate = "rotateX("+euler.x+"rad)"+
" rotateY("+(euler.y)+"rad)"+
" rotateY("+(euler.z)+"rad)";
btn.style.transform= translate+ " "+rotate;
}
... And a screenshot of the issue.
I would highly recommend not trying to match this to the camera space, but instead to apply the image as a texture map to the red plane, and then use a raycast to see whether a click goes over the plane. You'll save yourself headache in translating and rotating and then hiding the symbol when it's behind the cube, etc
check out the THREEjs examples to see how to use the Raycaster. It's a lot more flexible and easier than trying to do rotations and matching. Then whatever the 'btn' onclick function is, you just call when you detect a raycast collision with the plane
I would like to position cubes in a rectangular/square like grid. I'm having trouble trying to create some methodology in depending on what I pick through an HTML form input (checkboxes) to have it arrange left to right and up to down, a series of cubes, in a prearranged grid all on the same plane.
What measurement units is three.js in? Right now, I'm setting up my shapes using the built-in geometries, for instance.
var planeGeometry = new THREE.PlaneGeometry(4, 1, 1, 1);
The 4 and 1; I'm unsure what that measures up to in pixels, although I do see it rendered. I'm resorting to eyeballing it (guess and checking) every time so that it looks acceptable.
Without a fair bit of extra math THREE is not measured in pixels.
To make a simple grid (I leave optimizations, colors, etc for future refinements) try something like:
var hCount = from_my_web_form('horiz'),
vCount = from_my_web_form('vert'),
size = 1,
spacing = 1.3;
var grid = new THREE.Object3d(); // just to hold them all together
for (var h=0; h<hCount; h+=1) {
for (var v=0; v<vCount; v+=1) {
var box = new THREE.Mesh(new THREE.BoxGeometry(size,size,size),
new THREE.MeshBasicMaterial());
box.position.x = (h-hCount/2) * spacing;
box.position.y = (v-vCount/2) * spacing;
grid.add(box);
}
}
scene.add(grid);
Is there a way to determine the size and position of a model and then auto-center and scale the model so that it is positioned at the origin and within the view of the camera? I find that when I import a Collada model from Sketchup, if the model was not centered at the origin in Sketchup, then it is not centered in three.js. While that makes sense, it would be nice to auto-center to origin after importing.
I've seen some discussion in the different file loaders about getting the bounds of the imported model, but I have been unable to find any references to how to do that.
The scaling issue is less important, but I feel like it relates to a bounds function, which is why I asked it too.
EDIT:
More info after playing around a bit and a few more google searches...
The code for my callback function on loading the collada file now looks like this:
loader.load(mURL, function colladaReady( collada ) {
dae = collada.scene;
skin = collada.skins[ 0 ];
dae.scale.x = dae.scale.y = dae.scale.z = 1;
dae.updateMatrix();
//set arbitrary min and max for comparison
var minX = 100000;
var minY = 100000;
var minZ = 100000;
var maxX = 0;
var maxY = 0;
var maxZ = 0;
var geometries = collada.dae.geometries;
for(var propName in geometries){
if(geometries.hasOwnProperty(propName) && geometries[propName].mesh){
dae.geometry = geometries[propName].mesh.geometry3js;
dae.geometry.computeBoundingBox();
bBox = dae.geometry.boundingBox;
if(bBox.min.x < minX) minX = bBox.min.x;
if(bBox.min.y < minY) minY = bBox.min.x;
if(bBox.min.z < minZ) minZ = bBox.min.z;
if(bBox.max.x > maxX) maxX = bBox.max.x;
if(bBox.max.y > maxY) maxY = bBox.max.x;
if(bBox.max.z > maxZ) maxZ = bBox.max.z;
}
}
//rest of function....
This is generating some interesting data about the model. I can get an overall extreme coordinate for the model, which I'm assuming (probably incorrectly) would be close to an overall bounding box for the model. But trying to do anything with those coordinates (like averaging and moving the model to the averages) generates inconsistent results.
Also, it seems inefficient to have to loop through every geometry for a model, is there a better way? If not, can this logic be applied to other loaders?
You can use THREE.Box3#setFromObject to get the bounding box of any Object3D, including an imported model, without having to loop through the geometries yourself. So you could do something like
var bBox = new THREE.Box3().setFromObject(collada.scene);
to get the extreme bounding box of the model; then you could use any of the techniques in the answers that gaitat linked in order to set the camera position correctly. For instance, you could follow this technique (How to Fit Camera to Object) and do something like:
var height = bBox.size().y;
var dist = height / (2 * Math.tan(camera.fov * Math.PI / 360));
var pos = collada.scene.position;
camera.position.set(pos.x, pos.y, dist * 1.1); // fudge factor so you can see the boundaries
camera.lookAt(pos);
Quick fiddle: http://jsfiddle.net/p19r9re2/ .
try geometry.center()
center: function () {
var offset = new Vector3();
return function center() {
this.computeBoundingBox();
this.boundingBox.getCenter( offset ).negate();
this.translate( offset.x, offset.y, offset.z );
return this;
};
}(),