73 lines
1.6 KiB
JavaScript
73 lines
1.6 KiB
JavaScript
fs = require('fs');
|
|
|
|
//var dict = '/usr/share/dict/american-english';
|
|
var dict = '/usr/share/dict/italian';
|
|
//var dict = '/usr/share/dict/ngerman';
|
|
|
|
fs.readFile(dict, 'utf8', function(err, data) {
|
|
if (err) {
|
|
return console.log(err);
|
|
}
|
|
|
|
var lines = data.split('\n');
|
|
console.log(lines.length + ' words in dictionary');
|
|
|
|
var word = '';
|
|
var words = [];
|
|
for(var i=0; i<lines.length; i++) {
|
|
word = lines[i];
|
|
if (word.length > 6 && word.length < 10) {
|
|
//console.log('using ' + word);
|
|
words.push(word);
|
|
};
|
|
}
|
|
console.log('using ' + words.length + ' words out of the dictionary');
|
|
|
|
|
|
|
|
var wildcardWord = function(word) {
|
|
var words = [];
|
|
for(var i=1; i<word.length; i++) {
|
|
words.push(word.substr(0, i) + '.' + word.substr(i + 1));
|
|
}
|
|
return words;
|
|
};
|
|
|
|
var matchingWordlist = [];
|
|
for(var w=0; w<words.length; w++) {
|
|
var word = words[w];
|
|
var wildcardedWords = wildcardWord(word);
|
|
var matches = [];
|
|
|
|
for(var i=0; i<wildcardedWords.length; i++) {
|
|
wildcardedWord = wildcardedWords[i];
|
|
for(var j=0; j<words.length; j++) {
|
|
//console.log('match ' + words[j] + ' with ' + wildcardedWord);
|
|
|
|
if(word.length == words[j].length && words[j].match(wildcardedWord) && word != words[j]) {
|
|
matches.push(words[j]);
|
|
//console.log(word + ' -> ' + words[j]);
|
|
}
|
|
}
|
|
}
|
|
|
|
//add to dictionary?
|
|
if(matches.length >= 4) {
|
|
console.log('using ' + word);
|
|
var mw = {};
|
|
mw[word] = matches;
|
|
matchingWordlist.push(mw);
|
|
}
|
|
|
|
if(w % 5000 == 0) {
|
|
console.log('intermediary results...');
|
|
console.log(matchingWordlist);
|
|
}
|
|
}
|
|
|
|
|
|
console.log('final result');
|
|
console.log(JSON.stringify(matchingWordlist));
|
|
process.exit(0);
|
|
});
|