Develop Reference

How to get neighbor character from selected text?

I have a string like this:
var comment = 'this is a test';
Assume this i is selected, Now I need to null (left side) and s (right side). How can I get them?
I can get selected text like this:
function getSelectionHtml() {
var html = "";
if (typeof window.getSelection != "undefined") {
var sel = window.getSelection();
if (sel.rangeCount) {
var container = document.createElement("div");
for (var i = 0, len = sel.rangeCount; i < len; ++i) {
container.appendChild(sel.getRangeAt(i).cloneContents());
}
html = container.innerHTML;
}
} else if (typeof document.selection != "undefined") {
if (document.selection.type == "Text") {
html = document.selection.createRange().htmlText;
}
}
return html;
}
var selected_text = getSelectionHtml();

document.addEventListener("click", function() {
var selection = window.getSelection();
// Check if there are any ranges selected.
if (selection.rangeCount > 0 && selection.type == "Range") {
// Text content of the element.
var text = selection.anchorNode.textContent;
// selection.anchorOffset is the start position of the selection
var before = text.substring(selection.anchorOffset-1, selection.anchorOffset);
// selection.extentOffset is the end position of the selection
var after = text.substring(selection.extentOffset, selection.extentOffset+1);
// Check if there are any letters before or after selected string.
// If not, change before and after to null.
before = before.length === 1 ? before : null;
after = after.length === 1 ? after : null;
console.log(before, after);
}
});
<div>this is a test</div>
To get two characters:
document.addEventListener("click", function() {
var selection = window.getSelection();
// Check if there are any ranges selected.
if (selection.rangeCount > 0 && selection.type == "Range") {
// Text content of the element.
var text = selection.anchorNode.textContent;
// selection.anchorOffset is the start position of the selection
var before = text.substring(selection.anchorOffset-2, selection.anchorOffset);
// selection.extentOffset is the end position of the selection
var after = text.substring(selection.extentOffset, selection.extentOffset+2);
// Check if there are any letters before or after selected string.
// If not, change before and after to null.
before = before.length >= 1 ? before : null;
after = after.length >= 1 ? after : null;
console.log(before, after);
}
});
<div>this is a test</div>

You don't need jQuery for this. All you need is window.getSelection(). This returns an object. You can get the surrounding text with window.getSelection().anchorNode.data, and get the index of the selection within that text using window.getSelection().anchorOffset. Putting this all together, we have
var selection = window.getSelection();
var selectionText = selection.toString();
var surroundingText = selection.anchorNode.data;
var index = selection.anchorOffset;
var leftNeighbor = surroundingText[index - 1];
var rightNeighbor = surroundingText[index + selectionText.length];
Note that you will get undefined instead of null when there is no neighbor character.
window.addEventListener("click", function(){
var selection = window.getSelection();
var selectionText = selection.toString();
var surroundingText = selection.anchorNode.data;
var index = selection.anchorOffset;
var leftNeighbor = surroundingText[index - 1];
var rightNeighbor = surroundingText[index + selectionText.length];
alert(leftNeighbor + " " + rightNeighbor);
});
<div>
this is a test
</div>

The problem with the other two solutions is they both work if the user selects from left to right, but not if they select right to left. As such, I use this modified version:
const selection = window.getSelection();
const text = selection.anchorNode.textContent;
let beforeChar;
let afterChar;
if (selection.anchorOffset < selection.focusOffset) {
beforeChar = text.substring(selection.anchorOffset, selection.anchorOffset-1);
afterChar = text.substring(selection.extentOffset, selection.extentOffset+1);
} else {
beforeChar = text.substring(selection.extentOffset, selection.extentOffset-1)
afterChar = text.substring(selection.anchorOffset, selection.anchorOffset+1);
}
console.log(beforeChar, afterChar);

How do I store a range into a database

I want to store user selection in a database in order to highlight them the next time user visits the page.
Here is the code I am using to add a span around the selected text.
window.rangeIntersectsNode = function(range, node) {
var nodeRange;
if (range.intersectsNode) {
return range.intersectsNode(node);
} else {
nodeRange = node.ownerDocument.createRange();
try {
nodeRange.selectNode(node);
} catch (e) {
nodeRange.selectNodeContents(node);
}
return range.compareBoundaryPoints(Range.END_TO_START, nodeRange) == -1 && range.compareBoundaryPoints(Range.START_TO_END, nodeRange) == 1;
}
}
window.getSelectedElementTags = function(win) {
var range, sel, elmlist, treeWalker, containerElement;
sel = win.getSelection();
if (sel.rangeCount > 0) {
range = sel.getRangeAt(0);
}
if (range) {
containerElement = range.commonAncestorContainer;
if (containerElement.nodeType != 1) {
containerElement = containerElement.parentNode;
}
treeWalker = win.document.createTreeWalker(
containerElement, NodeFilter.SHOW_TEXT, function(node) {
return rangeIntersectsNode(range, node) ? NodeFilter.FILTER_ACCEPT : NodeFilter.FILTER_REJECT;
}, false);
elmlist = [treeWalker.currentNode];
while (treeWalker.nextNode()) {
elmlist.push(treeWalker.currentNode);
}
return (elmlist);
}
}
window.wrapInSpan = function(range) {
var tempElement;
var selectedArr = getSelectedElementTags(window);
for (var i = 2; i < (selectedArr.length - 1); i++) {
tempElement = document.createElement("span");
tempElement.innerHTML = selectedArr[i].nodeValue;
selectedArr[i].parentElement.replaceChild(tempElement, selectedArr[i])
}
var sC = range.startContainer;
if (selectedArr.length == 2) {
if (sC.nodeType == 3) {
tempElement = document.createElement("span");
tempElement.innerHTML = sC.nodeValue.substring(range.startOffset, range.endOffset);
var anc1 = sC.splitText(range.startOffset);
var anc2 = (anc1.splitText(range.endOffset)).previousSibling;
anc2.parentElement.replaceChild(tempElement, anc2);
}
} else {
if (sC.nodeType == 3) {
tempElement = document.createElement("span");
tempElement.innerHTML = sC.nodeValue.substring(range.startOffset);
sC.parentElement.insertBefore(tempElement, sC.nextSibling);
sC.nodeValue = sC.nodeValue.replace(sC.nodeValue.substring(range.startOffset), "");
}
var eC = range.endContainer;
if (eC.nodeType == 3) {
tempElement = document.createElement("span");
tempElement.innerHTML = eC.nodeValue.substring(0, range.endOffset);
eC.parentElement.insertBefore(tempElement, eC);
eC.nodeValue = eC.nodeValue.replace(eC.nodeValue.substring(0, range.endOffset), "");
}
}
}
i am using the following code to get the details of the selection, but it in the console I get always get the variables as undefined.
var startNode=range.anchorNode;
var endNode=range.focusNode;
var startoffset=range.anchorOffset;
var endoffset=range.focusOffset;
var location=[startNode,endNode,startoffset,endoffset];
console.log(startNode);
return location;
I would greatly appreciate if some one can help me figure out how to store the range and to reload it. I am not looking to use rangy's library unless it is essentials. Also i am not sure if it will work in a chrome extension

how can I get the Character offset related to body with Range Object and window.getSelection()?

I'm writing a client-side application with Javascript, I'm using the following functions :
function creation() {
var userSelection;
if (window.getSelection) {
userSelection = window.getSelection();
}
else if (document.selection) { // should come last; Opera!
userSelection = document.selection.createRange();
}
var rangeObject = getRangeObject(userSelection);
var startOffset = rangeObject.startOffset;
var endOffset = rangeObject.endOffset;
var startCon = rangeObject.startContainer;
var endCon = rangeObject.endContainer;
var myRange = document.createRange();
myRange.setStart(startCon,rangeObject.startOffset);
myRange.setEnd(endCon, rangeObject.endOffset);
$('#result').text(myRange.toString());
}
function getRangeObject(selectionObject) {
if (selectionObject.getRangeAt) {
var ret = selectionObject.getRangeAt(0);
return ret;
}
else { // Safari!
var range = document.createRange();
range.setStart(selectionObject.anchorNode, selectionObject.anchorOffset);
range.setEnd(selectionObject.focusNode, selectionObject.focusOffset);
return range;
}
}
I need a way to know the character offset related to body element. I found a function with counts the character in an element :
function getCharacterOffsetWithin(range, node) {
var treeWalker = document.createTreeWalker(
node,
NodeFilter.SHOW_TEXT,
function (node) {
var nodeRange = document.createRange();
nodeRange.selectNode(node);
return nodeRange.compareBoundaryPoints(Range.END_TO_END, range) < 1 ?
NodeFilter.FILTER_ACCEPT : NodeFilter.FILTER_REJECT;
},
false
);
var charCount = 0;
while (treeWalker.nextNode()) {
charCount += treeWalker.currentNode.length;
}
if (range.startContainer.nodeType == 3) {
charCount += range.startOffset;
}
return charCount;
}

That looks like a function from one of my answers. I overcomplicated it a little; see this answer for a simpler function that works in IE < 9 as well: https://stackoverflow.com/a/4812022/96100. You can just pass in document.body as the node parameter. Also, please read the part in the linked answer about the shortcomings of this approach.
Here's a live demo: http://jsfiddle.net/PzQjA/

Remove All id Attributes from nodes in a Range of Fragment

Is there a way to remove the id attribute of every node in a range or fragment?
Update: I finally found out that the bug I'm struggling with is based on a <[script]> being included in a range, and therefore unexpectedly cloned, when a chrome user does a ctrl+a. My goal would be to remove any instance of <[script]> from the range (or doc fragment), such that it is not replicated when cloned.

You may be able to use a TreeWalker, which works in pretty much all the browers that Range works in.
function actOnElementsInRange(range, func) {
function isContainedInRange(el, range) {
var elRange = range.cloneRange();
elRange.selectNode(el);
return range.compareBoundaryPoints(Range.START_TO_START, elRange) <= 0
&& range.compareBoundaryPoints(Range.END_TO_END, elRange) >= 0;
}
var rangeStartElement = range.startContainer;
if (rangeStartElement.nodeType == 3) {
rangeStartElement = rangeStartElement.parentNode;
}
var rangeEndElement = range.endContainer;
if (rangeEndElement.nodeType == 3) {
rangeEndElement = rangeEndElement.parentNode;
}
var isInRange = function(el) {
return (el === rangeStartElement || el === rangeEndElement ||
isContainedInRange(el, range))
? NodeFilter.FILTER_ACCEPT : NodeFilter.FILTER_SKIP;
};
var container = range.commonAncestorContainer;
if (container.nodeType != 1) {
container = container.parentNode;
}
var walker = document.createTreeWalker(document,
NodeFilter.SHOW_ELEMENT, isInRange, false);
while (walker.nextNode()) {
func(walker.currentNode);
}
}
actOnElementsInRange(range, function(el) {
el.removeAttribute("id");
});

yes: http://api.jquery.com/removeAttr/

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
}