Develop Reference

given an array representing a hierachy, output data into a tree form in JS

Given a data file which has an array representing a hierarchy. Create a tree data structure by writing a script in Javascript. Output the data in tree form:
Data file:
["transportation.cars.Mazda",
"transportation.cars.Honda",
"transportation.cars.Toyota",
"transportation.train.lightRail",
"transportation.train.rapidTransit",
"transportation.waterVehicle.ferry",
"transportation.waterVehicle.boats"
...]
Output in tree form:
root
transportation
cars
Mazda
Honda
Toyota
train
lightRail
rapidTransit
waterVehicle
ferry
boats
My attempt:
var root = new Node('root');
var arr = ["transportation.cars.Mazda",
"transportation.cars.Honda",
"transportation.cars.Toyota",
"transportation.train.lightRail",
"transportation.train.rapidTransit",
"transportation.waterVehicle.ferry",
"transportation.waterVehicle.boats"
]
for(var i of arr){
var res=i.split(".");
root.addChild(new Node(res[0]));
res[0].addChild(new Node(res[1]));
res[1].addChild(new Node(res[2]));
}
this.addChild = function(node) {
node.setParentNode(this);
this.children[this.children.length] = node;
}
console.log(root);
I am trying to create a tree structure using JavaScript, but it does not has the same function as in Java (i.e. it does not have class method unless using Typescript. )

You can use something similar to a trie tree. The way you add a node would have to be much more specific. But it's possible with something like this.
function Node(word)
{
this.value = word;
this.children = {};
}
function AddDotChain(chain)
{
let arr = chain.split('.');
let currentNode = this;
function recurse(currentIndex)
{
if(currentIndex === arr.length)
{
return;
}
let currentWord = arr[currentIndex];
if(currentNode.children[currentWord])
{
currentNode = currentNode[currentWord];
return recurse(currentIndex + 1);
}
let child = new Node(currentWord);
currentNode.children[currentWord] = child;
currentNode = child;
return recurse(currentIndex + 1);
}
}
Where you just slap the entire chain in there without splitting it. There's probably a flaw in my logic somewhere but the overall idea should work. This can also be done iteritavely if you wanna reduce the overhead of recursion. Forgive the messiness, Tried to type this as fast as possible.
Here's a sloppy sloppy implementation on repl.it.

You can do it, with a data structure as Tree, you only need loop over the array of string that contains the data and split them by dot and then add each item to the tree instance that will be created when you execute the function that take your array and output as a Tree data structure.
this code can help you
var arr = ["transportation.cars.Mazda",
"transportation.cars.Honda",
"transportation.cars.Toyota",
"transportation.train.lightRail",
"transportation.train.rapidTransit",
"transportation.waterVehicle.ferry",
"transportation.waterVehicle.boats"
];
function Node(data) {
this.data = data;
this.children = [];
}
function Tree(data) {
this.root = null;
}
Tree.prototype.contains = function(data) {
return this.find(data) ? true : false;
}
Tree.prototype.add = function(data, node) {
const newNode = new Node(data);
if (!this.root) {
this.root = newNode;
return;
}
const parent = node ? this.find(node) : null;
if (parent) {
if (!this.contains(data)) {
parent.children.push(newNode);
}
}
}
Tree.prototype.find = function(data) {
if (this.root) {
const queue = [this.root];
while(queue.length) {
const node = queue.shift();
if (node && node.data === data) {
return node;
}
for(var i = 0; i < node.children.length; i++) {
const child = node.children[i];
queue.push(child);
}
}
}
return null;
}
function createTreeOfTransportation(arr) {
const tree = new Tree();
for(var i = 0; i < arr.length; i++) {
const element = arr[i];
const nodes = element.split('.');
for (var j = 0; j < nodes.length; j++) {
const currentNode = nodes[j];
const parent = nodes[j-1];
console.log(j, parent);
tree.add(currentNode, parent);
}
}
return tree;
}
console.log(createTreeOfTransportation(arr));

copy an object doesn't effect on my code

I suppose to copy a object source, while copy changes, sourceshould not change.The source code goes as follow:
layoutTreemap: function(source) {
var copy = jQuery.extend(true,{},source);
var select_polygon = this.get_selected_polygon();
var vt = d3.layout.voronoitreemap()
var layoutNodes = vt(copy);
return layoutNodes;
}
d3.layout.voronoitreemap = function() {
var hierarchy = d3.layout.hierarchy().sort(null),
root_polygon = [[0,0],[500,0],[500,500],[0,500]],
iterations = 100,
somenewvariable = 0;
function voronoitreemap(d, depth) {
var nodes = hierarchy(d),
root = nodes[0];
root.polygon = root_polygon;
root.site = null;
if (depth != null){
max_depth = depth;
}
else{
max_depth = "Infinity";
}
computeDiagramRecursively(root, 0);
return nodes;
}
function computeDiagramRecursively(node, level) {
var children = node.children;
if(node.parent) node.parent = null;
if (children && children.length && level < max_depth) {
node.sites = VoronoiTreemap.init(node.polygon, node); // can't say dataset, how about node?
VoronoiTreemap.normalizeSites(node.sites);
VoronoiTreemap.sites = node.sites;
VoronoiTreemap.setClipPolygon(node.polygon);
VoronoiTreemap.useNegativeWeights = false;
VoronoiTreemap.cancelOnAreaErrorThreshold = true;
var polygons = VoronoiTreemap.doIterate(iterations);
// set children polygons and sites
for (var i = 0; i < children.length; i++) {
children[i].polygon = polygons[i];
children[i].site = VoronoiTreemap.sites[i];
computeDiagramRecursively(children[i], (level + 1));
}
}
}
....
return d3_layout_hierarchyRebind(voronoitreemap, hierarchy);
}
But after execute vt(copy), the source has been changed.

This gives issues
root.polygon = root_polygon;
children[i].polygon = polygons[i];
You are copying arrays you think. but actually you are copying the reference to the array. now you have two pointers to the same object(imagine two people using the same fork at dinner)
You need to change this into
root.polygon = root_polygon.slice();
children[i].polygon = polygons[i].slice();
this way the array gets copied instead of referenced. Now each dinner guest has its own fork.

after picking selected range and getting the parent node of the selection is there a way to find the absolute path? [duplicate]

Say I've selected a span tag in a large html document. If I treat the entire html document as a big nested array, I can find the position of the span tag through array indexes. How can I output the index path to that span tag? eg: 1,2,0,12,7 using JavaScript.
Also, how can I select the span tag by going through the index path?

This will work. It returns the path as an array instead of a string.
Updated per your request.
You can check it out here: http://jsbin.com/isata5/edit (hit preview)
// get a node's index.
function getIndex (node) {
var parent=node.parentElement||node.parentNode, i=-1, child;
while (parent && (child=parent.childNodes[++i])) if (child==node) return i;
return -1;
}
// get a node's path.
function getPath (node) {
var parent, path=[], index=getIndex(node);
(parent=node.parentElement||node.parentNode) && (path=getPath(parent));
index > -1 && path.push(index);
return path;
}
// get a node from a path.
function getNode (path) {
var node=document.documentElement, i=0, index;
while ((index=path[++i]) > -1) node=node.childNodes[index];
return node;
}
This example should work on this page in your console.
var testNode=document.getElementById('comment-4007919');
console.log("testNode: " + testNode.innerHTML);
var testPath=getPath(testNode);
console.log("testPath: " + testPath);
var testFind=getNode(testPath);
console.log("testFind: " + testFind.innerHTML);

Using jquery:
var tag = $('#myspan_id');
var index_path = [];
while(tag) {
index_path.push(tag.index());
tag = tag.parent();
}
index_path = index_path.reverse();

Using the DOM
node = /*Your span element*/;
var currentNode = node;
var branch = [];
var cn; /*Stores a Nodelist of the current parent node's children*/
var i;
while (currentNode.parentNode !== null)
{
cn = currentNode.parentNode.childNodes;
for (i=0;i<cn.length;i++)
{
if (cn[i] === currentNode)
{
branch.push(i);
break;
}
}
currentNode = currentNode.parentNode;
}
cn = document.childNodes;
for (i=0;i<cn.length;i++)
{
if (cn[i] === currentNode)
{
branch.push(i);
break;
}
}
node.innerHTML = branch.reverse().join(",");

composedPath for native event.
(function (E, d, w) {
if (!E.composedPath) {
E.composedPath = function () {
if (this.path) {
return this.path;
}
var target = this.target;
this.path = [];
while (target.parentNode !== null) {
this.path.push(target);
target = target.parentNode;
}
this.path.push(d, w);
return this.path;
};
}
})(Event.prototype, document, window);
use:
var path = event.composedPath()

record and retrieve html element / node path using javascript

Say I've selected a span tag in a large html document. If I treat the entire html document as a big nested array, I can find the position of the span tag through array indexes. How can I output the index path to that span tag? eg: 1,2,0,12,7 using JavaScript.
Also, how can I select the span tag by going through the index path?

This will work. It returns the path as an array instead of a string.
Updated per your request.
You can check it out here: http://jsbin.com/isata5/edit (hit preview)
// get a node's index.
function getIndex (node) {
var parent=node.parentElement||node.parentNode, i=-1, child;
while (parent && (child=parent.childNodes[++i])) if (child==node) return i;
return -1;
}
// get a node's path.
function getPath (node) {
var parent, path=[], index=getIndex(node);
(parent=node.parentElement||node.parentNode) && (path=getPath(parent));
index > -1 && path.push(index);
return path;
}
// get a node from a path.
function getNode (path) {
var node=document.documentElement, i=0, index;
while ((index=path[++i]) > -1) node=node.childNodes[index];
return node;
}
This example should work on this page in your console.
var testNode=document.getElementById('comment-4007919');
console.log("testNode: " + testNode.innerHTML);
var testPath=getPath(testNode);
console.log("testPath: " + testPath);
var testFind=getNode(testPath);
console.log("testFind: " + testFind.innerHTML);

Using jquery:
var tag = $('#myspan_id');
var index_path = [];
while(tag) {
index_path.push(tag.index());
tag = tag.parent();
}
index_path = index_path.reverse();

Using the DOM
node = /*Your span element*/;
var currentNode = node;
var branch = [];
var cn; /*Stores a Nodelist of the current parent node's children*/
var i;
while (currentNode.parentNode !== null)
{
cn = currentNode.parentNode.childNodes;
for (i=0;i<cn.length;i++)
{
if (cn[i] === currentNode)
{
branch.push(i);
break;
}
}
currentNode = currentNode.parentNode;
}
cn = document.childNodes;
for (i=0;i<cn.length;i++)
{
if (cn[i] === currentNode)
{
branch.push(i);
break;
}
}
node.innerHTML = branch.reverse().join(",");

composedPath for native event.
(function (E, d, w) {
if (!E.composedPath) {
E.composedPath = function () {
if (this.path) {
return this.path;
}
var target = this.target;
this.path = [];
while (target.parentNode !== null) {
this.path.push(target);
target = target.parentNode;
}
this.path.push(d, w);
return this.path;
};
}
})(Event.prototype, document, window);
use:
var path = event.composedPath()

How to get nodes lying inside a range with javascript?

I'm trying to get all the DOM nodes that are within a range object, what's the best way to do this?
var selection = window.getSelection(); //what the user has selected
var range = selection.getRangeAt(0); //the first range of the selection
var startNode = range.startContainer;
var endNode = range.endContainer;
var allNodes = /*insert magic*/;
I've been been thinking of a way for the last few hours and came up with this:
var getNextNode = function(node, skipChildren){
//if there are child nodes and we didn't come from a child node
if (node.firstChild && !skipChildren) {
return node.firstChild;
}
if (!node.parentNode){
return null;
}
return node.nextSibling
|| getNextNode(node.parentNode, true);
};
var getNodesInRange = function(range){
var startNode = range.startContainer.childNodes[range.startOffset]
|| range.startContainer;//it's a text node
var endNode = range.endContainer.childNodes[range.endOffset]
|| range.endContainer;
if (startNode == endNode && startNode.childNodes.length === 0) {
return [startNode];
};
var nodes = [];
do {
nodes.push(startNode);
}
while ((startNode = getNextNode(startNode))
&& (startNode != endNode));
return nodes;
};
However when the end node is the parent of the start node it returns everything on the page. I'm sure I'm overlooking something obvious? Or maybe going about it in totally the wrong way.
MDC/DOM/range

Here's an implementation I came up with to solve this:
function getNextNode(node)
{
if (node.firstChild)
return node.firstChild;
while (node)
{
if (node.nextSibling)
return node.nextSibling;
node = node.parentNode;
}
}
function getNodesInRange(range)
{
var start = range.startContainer;
var end = range.endContainer;
var commonAncestor = range.commonAncestorContainer;
var nodes = [];
var node;
// walk parent nodes from start to common ancestor
for (node = start.parentNode; node; node = node.parentNode)
{
nodes.push(node);
if (node == commonAncestor)
break;
}
nodes.reverse();
// walk children and siblings from start until end is found
for (node = start; node; node = getNextNode(node))
{
nodes.push(node);
if (node == end)
break;
}
return nodes;
}

The getNextNode will skip your desired endNode recursively if its a parent node.
Perform the conditional break check inside of the getNextNode instead:
var getNextNode = function(node, skipChildren, endNode){
//if there are child nodes and we didn't come from a child node
if (endNode == node) {
return null;
}
if (node.firstChild && !skipChildren) {
return node.firstChild;
}
if (!node.parentNode){
return null;
}
return node.nextSibling
|| getNextNode(node.parentNode, true, endNode);
};
and in while statement:
while (startNode = getNextNode(startNode, false , endNode));

The Rangy library has a Range.getNodes([Array nodeTypes[, Function filter]]) function.

I made 2 additional fixes based on MikeB's answer to improve the accuracy of the selected nodes.
I'm particularly testing this on select all operations, other than range selection made by dragging the cursor along text spanning across multiple elements.
In Firefox, hitting select all (CMD+A) returns a range where it's startContainer & endContainer is the contenteditable div, the difference is in the startOffset & endOffset where it's respectively the index of the first and the last child node.
In Chrome, hitting select all (CMD+A) returns a range where it's startContainer is the first child node of the contenteditable div, and the endContainer is the last child node of the contenteditable div.
The modifications I've added work around the discrepancies between the two. You can see the comments in the code for additional explanation.
function getNextNode(node) {
if (node.firstChild)
return node.firstChild;
while (node) {
if (node.nextSibling) return node.nextSibling;
node = node.parentNode;
}
}
function getNodesInRange(range) {
// MOD #1
// When the startContainer/endContainer is an element, its
// startOffset/endOffset basically points to the nth child node
// where the range starts/ends.
var start = range.startContainer.childNodes[range.startOffset] || range.startContainer;
var end = range.endContainer.childNodes[range.endOffset] || range.endContainer;
var commonAncestor = range.commonAncestorContainer;
var nodes = [];
var node;
// walk parent nodes from start to common ancestor
for (node = start.parentNode; node; node = node.parentNode)
{
nodes.push(node);
if (node == commonAncestor)
break;
}
nodes.reverse();
// walk children and siblings from start until end is found
for (node = start; node; node = getNextNode(node))
{
// MOD #2
// getNextNode might go outside of the range
// For a quick fix, I'm using jQuery's closest to determine
// when it goes out of range and exit the loop.
if (!$(node.parentNode).closest(commonAncestor)[0]) break;
nodes.push(node);
if (node == end)
break;
}
return nodes;
};

below code solve your problem
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<title>payam jabbari</title>
<script src="http://code.jquery.com/jquery-2.0.2.min.js" type="text/javascript"></script>
<script type="text/javascript">
$(document).ready(function(){
var startNode = $('p.first').contents().get(0);
var endNode = $('span.second').contents().get(0);
var range = document.createRange();
range.setStart(startNode, 0);
range.setEnd(endNode, 5);
var selection = document.getSelection();
selection.addRange(range);
// below code return all nodes in selection range. this code work in all browser
var nodes = range.cloneContents().querySelectorAll("*");
for(var i=0;i<nodes.length;i++)
{
alert(nodes[i].innerHTML);
}
});
</script>
</head>
<body>
<div>
<p class="first">Even a week ago, the idea of a Russian military intervention in Ukraine seemed far-fetched if not totally alarmist. But the arrival of Russian troops in Crimea over the weekend has shown that he is not averse to reckless adventures, even ones that offer little gain. In the coming days and weeks</p>
<ol>
<li>China says military will respond to provocations.</li>
<li >This Man Has Served 20 <span class="second"> Years—and May Die—in </span> Prison for Marijuana.</li>
<li>At White House, Israel's Netanyahu pushes back against Obama diplomacy.</li>
</ol>
</div>
</body>
</html>

Annon, great work. I've modified the original plus included Stefan's modifications in the following.
In addition, I removed the reliance on Range, which converts the function into an generic algorithm to walk between two nodes. Plus, I've wrapped everything into a single function.
Thoughts on other solutions:
Not interested in relying on jquery
Using cloneNode lifts the results to a fragment, which prevents many operations one might want to conduct during filtering.
Using querySelectAll on a cloned fragment is wonky because the start or end nodes may be within a wrapping node, hence the parser may not have the closing tag?
Example:
<div>
<p>A</p>
<div>
<p>B</p>
<div>
<p>C</p>
</div>
</div>
</div>
Assume start node is the "A" paragraph, and the end node is the "C" paragraph
. The resulting cloned fragment would be:
<p>A</p>
<div>
<p>B</p>
<div>
<p>C</p>
and we're missing closing tags? resulting in funky DOM structure?
Anyway, here's the function, which includes a filter option, which should return TRUE or FALSE to include/exclude from results.
var getNodesBetween = function(startNode, endNode, includeStartAndEnd, filter){
if (startNode == endNode && startNode.childNodes.length === 0) {
return [startNode];
};
var getNextNode = function(node, finalNode, skipChildren){
//if there are child nodes and we didn't come from a child node
if (finalNode == node) {
return null;
}
if (node.firstChild && !skipChildren) {
return node.firstChild;
}
if (!node.parentNode){
return null;
}
return node.nextSibling || getNextNode(node.parentNode, endNode, true);
};
var nodes = [];
if(includeStartAndEnd){
nodes.push(startNode);
}
while ((startNode = getNextNode(startNode, endNode)) && (startNode != endNode)){
if(filter){
if(filter(startNode)){
nodes.push(startNode);
}
} else {
nodes.push(startNode);
}
}
if(includeStartAndEnd){
nodes.push(endNode);
}
return nodes;
};

I wrote the perfect code for this and it works 100% for every node :
function getNodesInSelection() {
var range = window.getSelection().getRangeAt(0);
var node = range.startContainer;
var ranges = []
var nodes = []
while (node != null) {
var r = document.createRange();
r.selectNode(node)
if(node == range.startContainer){
r.setStart(node, range.startOffset)
}
if(node == range.endContainer){
r.setEnd(node, range.endOffset)
}
ranges.push(r)
nodes.push(node)
node = getNextElementInRange(node, range)
}
// do what you want with ranges and nodes
}
here are some helper functions
function getClosestUncle(node) {
var parent = node.parentElement;
while (parent != null) {
var uncle = parent.nextSibling;
if (uncle != null) {
return uncle;
}
uncle = parent.nextElementSibling;
if (uncle != null) {
return uncle;
}
parent = parent.parentElement
}
return null
}
function getFirstChild(_node) {
var deep = _node
while (deep.firstChild != null) {
deep = deep.firstChild
}
return deep
}
function getNextElementInRange(currentNode, range) {
var sib = currentNode.nextSibling;
if (sib != null && range.intersectsNode(sib)) {
return getFirstChild(sib)
}
var sibEl = currentNode.nextSiblingElemnent;
if (sibEl != null && range.intersectsNode(sibEl)) {
return getFirstChild(sibEl)
}
var uncle = getClosestUncle(currentNode);
var nephew = getFirstChild(uncle)
if (nephew != null && range.intersectsNode(nephew)) {
return nephew
}
return null
}

bob. the function only returns the startNode and endNode. the nodes in between do not get pushed to the array.
seems the while loop returns null on getNextNode() hence that block never gets executed.

here is function return you array of sub-ranges
function getSafeRanges(range) {
var doc = document;
var commonAncestorContainer = range.commonAncestorContainer;
var startContainer = range.startContainer;
var endContainer = range.endContainer;
var startArray = new Array(0),
startRange = new Array(0);
var endArray = new Array(0),
endRange = new Array(0);
// ##### If start container and end container is same
if (startContainer == endContainer) {
return [range];
} else {
for (var i = startContainer; i != commonAncestorContainer; i = i.parentNode) {
startArray.push(i);
}
for (var i = endContainer; i != commonAncestorContainer; i = i.parentNode) {
endArray.push(i);
}
}
if (0 < startArray.length) {
for (var i = 0; i < startArray.length; i++) {
if (i) {
var node = startArray[i - 1];
while ((node = node.nextSibling) != null) {
startRange = startRange.concat(getRangeOfChildNodes(node));
}
} else {
var xs = doc.createRange();
var s = startArray[i];
var offset = range.startOffset;
var ea = (startArray[i].nodeType == Node.TEXT_NODE) ? startArray[i] : startArray[i].lastChild;
xs.setStart(s, offset);
xs.setEndAfter(ea);
startRange.push(xs);
}
}
}
if (0 < endArray.length) {
for (var i = 0; i < endArray.length; i++) {
if (i) {
var node = endArray[i - 1];
while ((node = node.previousSibling) != null) {
endRange = endRange.concat(getRangeOfChildNodes(node));
}
} else {
var xe = doc.createRange();
var sb = (endArray[i].nodeType == Node.TEXT_NODE) ? endArray[i] : endArray[i].firstChild;
var end = endArray[i];
var offset = range.endOffset;
xe.setStartBefore(sb);
xe.setEnd(end, offset);
endRange.unshift(xe);
}
}
}
var topStartNode = startArray[startArray.length - 1];
var topEndNode = endArray[endArray.length - 1];
var middleRange = getRangeOfMiddleElements(topStartNode, topEndNode);
startRange = startRange.concat(middleRange);
response = startRange.concat(endRange);
return response;
}

With generator and document.createTreeWalker:
function *getNodeInRange(range) {
let [start, end] = [range.startContainer, range.endContainer]
if (start.nodeType < Node.TEXT_NODE || Node.COMMENT_NODE < start.nodeType) {
start = start.childNodes[range.startOffset]
}
if (end.nodeType < Node.TEXT_NODE || Node.COMMENT_NODE < end.nodeType) {
end = end.childNodes[range.endOffset-1]
}
const relation = start.compareDocumentPosition(end)
if (relation & Node.DOCUMENT_POSITION_PRECEDING) {
[start, end] = [end, start]
}
const walker = document.createTreeWalker(
document, NodeFilter.SHOW_ALL
)
walker.currentNode = start
yield start
while (walker.parentNode()) yield walker.currentNode
if (!start.isSameNode(end)) {
walker.currentNode = start
while (walker.nextNode()) {
yield walker.currentNode
if (walker.currentNode.isSameNode(end)) break
}
}
const subWalker = document.createTreeWalker(
end, NodeFilter.SHOW_ALL
)
while (subWalker.nextNode()) yield subWalker.currentNode
}