I looked at several examples on how to use regex in JS but I can not seem to find the right syntax for what I need. Basically I have an array of words:
commonWords=["she", "he", "him", "liked", "i", "a", "an", "are"]
and a string:
'She met him where he liked to eat "the best" cheese pizza.'
Basically I want to use non-alphas and my array of commonWords as delimiters for extracting phrases. The above would yield somthing like this:
'met, where, to eat, the best, cheese pizza'
From the OP:
"Basically I want to use non-alphas and my array of commonWords as delimiters for extracting phrases."
This does both (unlike some other answers ;-) ). It returns either a string or an array.
var commonWords = ["she", "he", "him", "liked", "i", "a", "an", "are"];
var SourceStr = 'She met him where he liked to eat "the best" cheese pizza, didn\'t she, $%&#! Mr. O\'Leary?';
//--- Kill (most) non-alphas, and the keywords replace with tab.
var zRegEx = eval ('/([^0-9a-z\' ]+)|\\s*\\b(' + commonWords.join ("|") + ')\\b\\s*/ig');
var sPhraseList = SourceStr.replace (zRegEx, '\t');
//-- Trim empty results and leading and trailing delimiters.
sPhraseList = sPhraseList.replace (/ *\t+ */g, ', '). replace (/, ?, ?/g, ', ');
sPhraseList = sPhraseList.replace (/(^[, ]+)|([, ]+$)/g, '');
//-- Make optional array:
aPhraseList = sPhraseList.split (/, */g);
//-- Replace "console.log" with "alert" if you're not using Firebug.
console.log (SourceStr);
console.log (sPhraseList);
console.log (aPhraseList);
.
This returns:
"met, where, to eat, the best, cheese pizza, didn't, Mr, O'Leary"
and
["met", "where", "to eat", "the best", "cheese pizza", "didn't", "Mr", "O'Leary"]
Are you looking for something like this:
var commonWords=["she", "he", "him", "liked", "i", "a", "an", "are"];
var regstr = "\\b(" + commonWords.join("|") + ")\\b";
//regex is \b(she|he|him|liked|i|a|an|are)\b
var regex = new RegExp(regstr, "ig");
var str = 'She met him where he liked to eat "the best" cheese pizza.';
console.log(str.replace(regex, ""));
output
met where to eat "the best" cheese pizza.
split version:
var commonWords=["she", "he", "him", "liked", "i", "a", "an", "are"];
var regstr = "\\b(?:" + commonWords.join("|") + ")\\b";
var regex = new RegExp(regstr, "ig");
var str = 'She met him where he liked to eat "the best" cheese pizza.';
var arr = str.split(regex);
console.log(arr);// ["", " met ", " where ", " ", " to eat "the best" cheese pizza."]
for(var i = 0; i < arr.length; i++)
if(arr[i].match(/^\s*$/)) //remove empty strings and strings with only spaces.
arr.splice(i--, 1);
else
arr[i] = arr[i].replace(/^\s+|\s+$/g, ""); //trim spaces from beginning and end
console.log(arr);// ["met", "where", "to eat "the best" cheese pizza."]
console.log(arr.join(", "));// met, where, to eat "the best" cheese pizza.
split instead of replace, but it's similar enough. (ie, remove capturing group, and maybe empty tokens)This version is quite verbose, but works with “lazy” single and double quotes as well:
If array contains object (like indexOfObject) with a case-insensitive comparison flag:
if (!Array.prototype.containsObject) Array.prototype.containsObject = function (object, caseInsensitive) {
for (var i = 0; i < this.length; i++) {
if (this[i] == object) return true;
if (!(caseInsensitive && (typeof this[i] == 'string') && (typeof object == 'string'))) continue;
return (this[i].match(RegExp(object, "i")) != null);
}
return false;
}
Push object to the array if not empty:
if (!Array.prototype.pushIfNotEmpty) Array.prototype.pushIfNotEmpty = function (object) {
if (typeof object == 'undefined') return;
if ((object && object.length) <= 0) return;
this.push(object);
}
Canonicalizing strings:
function canonicalizeString (inString, whitespaceSpecifier) {
if (typeof inString != 'string') return '';
if (typeof whitespaceSpecifier != 'string') return '';
var whitespaceReplacement = whitespaceSpecifier + whitespaceSpecifier;
var canonicalString = inString.replace(whitespaceSpecifier, whitespaceReplacement);
var singleQuotedTokens = canonicalString.match(/'([^'s][^']*)'/ig);
for (tokenIndex in singleQuotedTokens) canonicalString = canonicalString.replace(singleQuotedTokens[tokenIndex], String(singleQuotedTokens[tokenIndex]).replace(" ", whitespaceReplacement));
var doubleQuotedTokens = canonicalString.match(/"([^"]*)"/ig);
for (tokenIndex in doubleQuotedTokens) canonicalString = canonicalString.replace(doubleQuotedTokens[tokenIndex], String(doubleQuotedTokens[tokenIndex]).replace(" ", whitespaceReplacement));
return canonicalString;
}
Have fun:
function getSignificantTokensFromStringWithCommonWords (inString, inCommonWordsArray) {
if (typeof inString != 'string') return [];
if (typeof (inCommonWordsArray && inCommonWordsArray.length) != 'number') return [];
var canonicalString = canonicalizeString(inString, "_");
var commonWords = [];
for (indexOfCommonWord in inCommonWordsArray) commonWords.pushIfNotEmpty(canonicalizeString(inCommonWordsArray[indexOfCommonWord], "_"));
var tokenizedStrings = canonicalString.split(" ");
for (indexOfToken in tokenizedStrings)
if (commonWords.containsObject(tokenizedStrings[indexOfToken], true))
tokenizedStrings[indexOfToken] = undefined;
var responseObject = [];
for (indexOfToken in tokenizedStrings)
if (typeof tokenizedStrings[indexOfToken] == 'string')
responseObject.push(tokenizedStrings[indexOfToken]);
for (indexOfTokenInResponse in responseObject)
if (typeof responseObject[indexOfTokenInResponse] == 'string')
responseObject[indexOfTokenInResponse] = String(responseObject[indexOfTokenInResponse]).replace("__", " ");
return responseObject;
}
getSignificantTokensFromStringWithCommonWords(inString, inCommonWordsArray) to work with this snippet.
'met, where, to eat, the best, cheese pizza'. "liked" is in the commonWords list.