Stopwords filtering

Benchmark created on November 15, 2013

Description

Setup

var text = "They were talking about shoes, red shoes, that were found underneath a train bound for London, England this afternoon supposing the hooligans don't have their way; and as long as the innovative Australian lad wasn't going on and on about some sort of chap and a snat!";
    
    var stopwords = ["a", "about", "above", "above", "across", "after",
    "afterwards", "again", "against", "all", "almost", "alone", "along",
    "already", "also","although","always","am","among", "amongst",
    "amoungst", "amount",  "an", "and", "another", "any","anyhow","anyone",
    "anything","anyway", "anywhere", "are", "around", "as",  "at", "back",
    "be","became", "because","become","becomes", "becoming", "been",
    "before", "beforehand", "behind", "being", "below", "beside",
    "besides", "between", "beyond", "bill", "both", "bottom","but",
    "by", "call", "can", "cannot", "cant", "co", "con", "could",
    "couldnt", "cry", "de", "describe", "detail",  "do", "does", "done",  "down",
    "due", "during", "each", "eg", "eight", "either", "eleven","else",
    "elsewhere", "empty", "enough", "etc", "even", "ever", "every",
    "everyone", "everything", "everywhere", "except", "few", "fifteen",
    "fify", "fill", "find", "fire", "first", "five", "for", "former",
    "formerly", "forty", "found", "four", "from", "front", "full",
    "further", "get", "give", "go", "had", "has", "hasnt", "have",
    "he", "hence", "her", "here", "hereafter", "hereby", "herein",
    "hereupon", "hers", "herself", "him", "himself", "his", "how",
    "however", "hundred", "ie", "if", "in", "inc", "indeed",
    "interest", "into", "is",  "it", "its", "itself", "keep",
    "last", "latter", "latterly", "least", "less", "ltd", "made",
    "many", "may", "me", "meanwhile", "might", "mill", "mine",
    "more", "moreover", "most", "mostly", "move", "much", "must", "my",
    "myself", "name", "namely", "neither", "never", "nevertheless", "next",
    "nine", "no", "nobody", "none", "noone", "nor", "not", "nothing", "now",
    "nowhere", "of", "off", "often", "on", "once", "one", "only", "onto",
    "or", "other", "others", "otherwise", "our", "ours", "ourselves", "out",
    "over", "own","part", "per", "perhaps", "please", "put", "rather", "re",
    "same", "see", "seem", "seemed", "seeming", "seems", "serious", "several",
    "she", "should", "show", "side", "since", "sincere", "six", "sixty", "so",
    "some", "somehow", "someone", "something", "sometime", "sometimes", "somewhere",
    "still", "such", "system", "take", "ten", "than", "that", "the", "their", "them",
    "themselves", "then", "thence", "there", "thereafter", "thereby", "therefore",
    "therein", "thereupon", "these", "they", "thickv", "thin", "third", "this",
    "those", "though", "three", "through", "throughout", "thru", "thus", "to",
    "together", "too", "top", "toward", "towards", "twelve", "twenty", "two",
    "un", "under", "until", "up", "upon", "us", "very", "via", "was", "way", "we",
    "well", "were", "what", "whatever", "when", "whence", "whenever", "where",
    "whereafter", "whereas", "whereby", "wherein", "whereupon", "wherever",
    "whether", "which", "while", "whither", "who", "whoever", "whole", "whom",
    "whose", "why", "will", "with", "within", "without", "would", "yet", "you",
    "your", "yours", "yourself", "yourselves", "the",
    // contractions?
    "didnt", "doesnt", "dont", "isnt", "wasnt", "youre", "hes", "ive", "theyll",
    "whos", "wheres", "whens", "whys", "hows", "whats", "were", "shes", "im", "thats"
    ];
    
    var stopwordsRE = new RegExp(stopwords.join("|"), "gi");
    
    function filterIO(text) {
        var words = text.split(' ');
        var keywords = [];
        for (var i = 0; i < words.length; i++) {
                var word = words[i].toLowerCase().replace(/[^a-zA-Z]/, '');
                if (stopwords.indexOf(word) === -1 && keywords.indexOf(word) === -1) {
                        keywords.push(word);
                }
        }
    
        return keywords;
    }
    
    function filter(text) {
        var cleansed_string = text;
    
        // Split out all the individual words in the phrase
        var words = cleansed_string.match(/[^\s]+|\s+[^\s+]$/g);
    
        var keywords = [];
    
        // Review all the words
        for (x=0; x < words.length; x++) {
                // For each word, check all the stop words
                for (y=0; y < stopwords.length; y++) {
                        // Get the current word
                        var word = words[x].replace(/\s+|[^a-z]+/ig, "");       // Trim the word and remove non-alpha
    
                        // Get the stop word
                        var stopword = stopwords[y];
    
                        // If the word matches the stop word, remove it from the keywords
                        if (word.toLowerCase() == stopword) {
                                // Build the regex
                                regex_str = "^\\s*"+stopword+"\\s*$";           // Only word
                                regex_str += "|^\\s*"+stopword+"\\s+";          // First word
                                regex_str += "|\\s+"+stopword+"\\s*$";          // Last word
                                regex_str += "|\\s+"+stopword+"\\s+";           // Word somewhere in the middle
                                regex = new RegExp(regex_str, "ig");
    
                                // Remove the word from the keywords
                                cleansed_string = cleansed_string.replace(regex, " ");
                        }
                }
        }
    
        var keywordsFound = cleansed_string.split(' ');
        for (var i = 0; i < keywordsFound.length; i++) {
                if (keywords.indexOf(keywordsFound[i]) === -1) {
                        keywords.push(keywordsFound[i]);
                }
        }
    
        return keywords;
    }

Test runner

var filteredText = filter(text)
 

var filteredText = filterIO(text);

Revisions

You can edit these tests or add more tests to this page by appending /edit to the URL.

• Revision 1: published on November 15, 2013