Develop Reference

Javascript word boundary regex with "\w" for words with double byte characters

I do match of words in a text to retrieve the word offset begin and end. This normally works for both ascii and unicode texts when using an appropriate unicode-aware regex like '(?<=^|\\PL)$1(?=\\PL|$)'. When I have mixed text (like Korean and English here) there are some problems, while tokenizing:
function aggressive_tokenizer(text) {
// most punctuation
text = text.replace(/([^\w\.\'\-\/\+\<\>,&])/g, " $1 ");
// commas if followed by space
text = text.replace(/(,\s)/g, " $1");
// single quotes if followed by a space
text = text.replace(/('\s)/g, " $1");
// single quotes if last char
text = text.replace(/('$)/, " $1");
text = text.replace(/(\s+[`'"‘])(\w+)\b(?!\2)/g, " $2")
// periods before newline or end of string
text = text.replace(/\. *(\n|$)/g, " . ");
// replace punct
// ignore "-" since may be in slang scream
text = text.replace(/[\\?\^%<>=!&|+\~]/g, "");
text = text.replace(/[…;,.:*#\)\({}\[\]]/g, "");
// finally split remainings into words
text = text.split(/\s+/)
return text;
}
var text = "점점 더 깊이 끌려가"
var tokens = aggressive_tokenizer(text);
var seen = new Map();
var indexes = tokens.map(token => { // for each token
let item = {
"word": token
}
var pattern = '(?<!\\pL\\pM*)$1(?!\\pL)';
var escaped = token.replace(/[\-\[\]{}()*+?.,\\\^$|#\s]/g, "\\$&");
var wordRegex = new XRegExp(pattern.replace('$1', escaped), "g");
// calculate token begin end
var match = null;
while ((match = wordRegex.exec(text)) !== null) {
if (match.index > (seen.get(token) || -1)) {
var wordStart = match.index;
var wordEnd = wordStart + token.length - 1;
item.characterOffsetBegin = wordStart;
item.characterOffsetEnd = wordEnd;
seen.set(token, wordEnd);
break;
}
}
return item;
});
indexes.forEach(index => {
if (index.word != text.slice(index.characterOffsetBegin, index.characterOffsetEnd + 1)) {
console.log("NOT MATCHING!!! " + index.word + " : " + text.slice(index.characterOffsetBegin, index.characterOffsetEnd + 1))
} else {
console.log("\tMATCHED " + index.word + " : " + text.slice(index.characterOffsetBegin, index.characterOffsetEnd + 1))
}
});
<script src="https://unpkg.com/xregexp/xregexp-all.js"></script>
The problem is that I do some cleanup in tokenization like
text = text.replace(/([^\w\.\'\-\/\+\<\>,&])/g, " $1 ");
where \w is not unicode-aware, but if I replace it with \p{Alnum}:
text = text.replace(/([^\p{Alnum}\.\'\-\/\+\<\>,&])/g, " $1 ");
that it should be the equivalent for Unicode word, it does not work properly.
NOTE
Please note that I do use XRegExp to support Unicode regex in JavaScript.
UPDATE
According to the comments below, I have updated the code with the modified pattern regexp '(?<=^|\\PL)$1(?=\\PL|$)' by – Wiktor Stribiżew and replaced XRegExp with built-in RegExp, due to missing support for varied width lookbehind patterns (see comments).
This solution works better, but I have identified an additional case where the char offset begin and end cannot be matched for the given input text: "점점 더 깊이 끌려가" the output will have a missing offset / match for
{
"index": 2,
"word": "점"
}
function aggressive_tokenizer(text) {
// most punctuation
text = text.replace(/[^\w\.\-\/\+\<\>,&]/g, " $& ");
// commas if followed by space
text = text.replace(/(,\s)/g, " $1");
// single quotes if followed by a space
text = text.replace(/('\s)/g, " $1");
// single quotes if last char
text = text.replace(/('$)/, " $1");
text = text.replace(/(\s+[`'"‘])(\w+)\b(?!\2)/g, " $2")
// periods before newline or end of string
text = text.replace(/\. *(\n|$)/g, " . ");
// replace punct
// ignore "-" since may be in slang scream
text = text.replace(/[\\?\^%<>=!&|+\~]/g, "");
text = text.replace(/[…;,.:*#\)\({}\[\]]/g, "");
// finally split remainings into words
text = text.split(/\s+/)
return text;
}
var text = "점점 더 깊이 끌려가"
var tokens = aggressive_tokenizer(text);
var seen = new Map();
var indexes = tokens.map(token => { // for each token
let item = {
"word": token
}
var pattern = '(?<!\\pL\\pM*)$1(?!\\pL)';
var escaped = token.replace(/[\-\[\]{}()*+?.,\\\^$|#\s]/g, "\\$&");
var wordRegex = new RegExp(pattern.replace('$1', escaped), "g");
// calculate token begin end
var match = null;
while ((match = wordRegex.exec(text)) !== null) {
if (match.index > (seen.get(token) || -1)) {
var wordStart = match.index;
var wordEnd = wordStart + token.length - 1;
item.characterOffsetBegin = wordStart;
item.characterOffsetEnd = wordEnd;
seen.set(token, wordEnd);
break;
}
}
return item;
});
indexes.forEach(index => {
if (!index.characterOffsetBegin && !index.characterOffsetEnd) {
console.log("MISSING INDEXES " + index.word);
} else if (index.word != text.slice(index.characterOffsetBegin, index.characterOffsetEnd + 1)) {
console.log("NOT MATCHING!!! " + index.word + " : " + text.slice(index.characterOffsetBegin, index.characterOffsetEnd + 1))
} else {
console.log("\tMATCHED " + index.word + " : " + text.slice(index.characterOffsetBegin, index.characterOffsetEnd + 1))
}
});

find the match at end of the string using regex in javascript

I have a string for ex "adlkup.db.com" and I want to validate the string for ".com" at the end of the string.
var x = "adlkup.db.com";
So I am trying something like
/.com$/.test(x)
and the . is interpreting to some other regex which finds a single character, except newline or line terminator

A period in a regular expression matches any character.
To make it literal, you need to escape it:
/\.com$/.test('stackoverflow.com'); // true
/\.com$/.test('stackoverflowcom'); // false
Alternatively, as Racil Hilan points out in the comments, you can also use the .lastIndexOf() method in order to check:
var string = 'stackoverflow.com';
string.lastIndexOf('.com') === string.length - 4; // true
or using the .substr() method:
'stackoverflow.com'.substr(-4) === '.com'; // true

In ECMAScript 6 this is done with endsWith:
x.endsWith(".com");
There is a polyfill for old browsers.

After reading your comments, I think you can use this better than the regex:
var value1 = "adlkup.db.com";
var value2 = "adlkup.db.com.w3Schools";
var value3 = ".com";
document.write(value1 + " " + endWithCom(value1) + "<br/>");
document.write(value2 + " " + endWithCom(value2) + "<br/>");
document.write(value3 + " " + endWithCom(value3) + "<br/>");
function endWithCom(text){
if(text.length < 5)
return false;
return (text.substr(-4) == ".com");
}
And you can easily convert it to generic function so you can pass it any ending you want to check:
var value1 = "adlkup.db.com";
var value2 = "adlkup.db.com.w3Schools";
var value3 = ".com";
var value4 = "adlkup.db.org";
document.write(value1 + " " + endWithButNotEqual(value1, ".com") + "<br/>");
document.write(value2 + " " + endWithButNotEqual(value2, ".com") + "<br/>");
document.write(value3 + " " + endWithButNotEqual(value3, ".com") + "<br/>");
document.write(value4 + " " + endWithButNotEqual(value4, ".org") + "<br/>");
function endWithButNotEqual(text, ending){
if(text.length <= ending.length)
return false;
return (text.substr(-ending.length) == ending);
}

I need this code to check for digits only

var tel = "123457890";
if (tel.length != 10) {
console.log("Sorry, incorrect format.");
} else {
var areaCode = tel.substring(0,3);
var prefix = tel.substring(3,6);
var suffix = tel.substring(6,10);
console.log("(" + areaCode + ") " + prefix + "-" + suffix);
}
Everything about this works for me except I can't get it to check for digits only. When I troubleshoot it with suggestions and run it through http://repl.it/languages/JavaScript I run into error messages.
i.e. When I put a "w" in with a string of 10 numerical digits I want it to return "Sorry, incorrect format" as if I had put in the wrong amount of numbers, etc.

You can use regex:
var tel = "123457890";
if (!tel.match(/^\d{10}$/)) {
console.log("Sorry, incorrect format.");
} else {
var areaCode = tel.substring(0,3);
var prefix = tel.substring(3,6);
var suffix = tel.substring(6,10);
console.log("(" + areaCode + ") " + prefix + "-" + suffix);
}

Here's a fiddle:
http://jsfiddle.net/e1qq659g/
Here's a regex pattern:
http://regexr.com/3b77b
var regularExpression = /\d{10}/g;
(tel.match(regularExpression)) ? alert("All digits!") : alert("Not Digits!");
this will match the expression to the string in variable tel - if it does match it alerts "All Digits" - if it does not match it alerts "Not Digits!"

This works. It deletes all letters and checks if the length changes
var tel = "1234567890";
if (tel.length !== 10 || tel.length !== parseInt(tel).toString().length) {
document.write("Sorry, incorrect format.");
} else {
var areaCode = tel.substring(0,3);
var prefix = tel.substring(3,6);
var suffix = tel.substring(6,10);
document.write("(" + areaCode + ") " + prefix + "-" + suffix);
}

You can use RegExp both to check if it only contains digits and to extract each part:
var tel = document.getElementById("tel");
var result = document.getElementById("result");
tel.oninput = function(e){
//var match = tel.value.match(/^(\d{3})(\d{3})(\d{4})$/); // Basic RegExp
var match = tel.value
.replace(/\s/g, '') // Removing spaces before RegExp makes it simplier
.match(/^(\d{3}|\(\d{3}\))(\d{3})\-?(\d{4})$/);
if(match == null) {
result.innerHTML = "Sorry, incorrect format.";
return false;
}
// parseInt is optional
var areaCode = parseInt(match[1].replace(/[\(\)]/g,'')); // Just need this with the second RegExp
var prefix = parseInt(match[2]);
var subfix = parseInt(match[3]);
result.innerHTML = "Valida input: (" + areaCode + ") " + prefix + "-" + subfix + "\n"
+ "\n\tArea Code: " + areaCode
+ "\n\tPrefix: " + prefix
+ "\n\tSubfix: " + subfix
};
<input type="text" id="tel" placeholder="Enter a valid phone number"/>
<pre id="result"></pre>
This way you could change the RegExp to make that match a formatted input too, lets say (123)456-7890, (123) 456-7890, (123) 456 - 7890, (123)4567890... that's what the current code does.
If you comment the current RegExp and uncomment the first one it will work just as you was asking for, accepting only a 10 digit input.

swapping in javascript without split

text left - text right
How to swap right to left using jquery? str pos is not so good because the parttern is not always fix. It could be somethingleft-somethingright

What's wrong with?:
var string = "text left - text right";
var newString = string.split("-").reverse().join(" - ");

var before = "text left - text right";
var separator = " - ";
var regex = new RegExp("(.*)(" + separator + ")(.*)");
var replace = "$3$2$1";
var after = before.replace(regex, replace);
document.write(after);

You may use String.replace() and RegExp for this task;
var swap = function(str) {
return str.replace(/(.*)(\s?\-\s?)(.*)/, '$3$2$1');
};
var str = 'Left - Right';
alert('Before is: ' + str);
alert('After is: ' + swap(str));
str = 'Apples-Mangoes';
alert('Before is: ' + str);
alert('After is: ' + swap(str));

Add .html to a string Javascript?

I want to check if a string is ending with ".php" extension, if not I want to add .html at the end. I have already tried various "slice" methods without success.

You can use Regex for that
var string1 = "www.example.com/index";
var newString = !/\.php$/i.test(string1)? string1+".html": string1;
// newString = "www.example.com/index.html"

Use (yourstring + '.html').replace(/\.php\.html$/, '.php') to do that:
var str1 = 'one.php';
var str2 = 'two';
var str3 = '.php.three.php';
var str4 = '.php.hey';
console.log((str1 + '.html').replace(/\.php\.html$/, '.php')); // Prints one.php
console.log((str2 + '.html').replace(/\.php\.html$/, '.php')); // Prints two.html
console.log((str3 + '.html').replace(/\.php\.html$/, '.php')); // Prints .php.three.php
console.log((str4 + '.html').replace(/\.php\.html$/, '.php')); // Prints .php.hey.html

Or perhaps:
function appendHTML(string) {
var html = string;
if (string.lastIndexOf('.php') === (string.length - 4)) {
html += '.html';
}
return html;
}

Well, slice() works ok for this task.
var s = "myfile.php";
if (s.slice(-4) != ".php")
s = s.slice(0, -4) + ".html";

Use regular expression to solve your problem.
/.php$/ is a regular expression that checks to see if a string ends with '.php'
For more information read: http://www.w3schools.com/js/js_obj_regexp.asp
Example Code:
str = "http://abc.com";
str = ( /\.php$/.test( str ) ) ? str : str + '.html'; // this is the line you want.
str === "http://abc.com.html" // returns true

Try something like this
function isPHP(str)
{
return str.substring(str.length - 4) == ".php";
}
Then you could do
str = isPHP(str) ? str : str + ".html";