Automatically generate tags from strings with javascript

Question

I need to -automatically- generate tags for a text string. In this case, I'll use this string:

var text = 'This text talks about loyalty in the Royal Family with Príncipe Charles';

My current implementation, generates the tags for words that are 6+ characters long, and it works fine.

words = (text).replace(/[^a-zA-Z\s]/g,function(str){return '';});
words = words.match(/\w{6,}/g);
console.log(words);

This will return:

["loyalty","Family","Prince","Charles"]

The problem is that sometimes, a tag should be a specific set of words. I need the result to be:

["loyalty","Royal Family","Príncipe Charles"]

That means, that the replace/match code should test for:

words that are 6 characters long (or more); and/or
if a set of words starts with an uppercase letter, those words should be joined together in the same array element. It doesn't matter if some of the words are less than 6 characters long - but at least one of them has to be 6+, e.g.: "Stop at The UK Guardián in London" should return ["The UK Guardián", "London"]

I'm obviously having trouble in the second requirement. Any ideas? Thanks!

georg · Accepted Answer · 2012-07-12 14:14:18Z

7

var text = 'This text talks about loyalty in the Royal Family with Prince Charles. Stop at The UK Guardian in London';

text.match(/(([A-Z]\w*\s*){2,})|(\w{6,})/g)

will return

["loyalty", "Royal Family ", "Prince Charles", "The UK Guardian ", "London"]

To fulfill the second requirement, it's better to run another regexp over the matches found:

var text = 'This is a Short Set Of Words about the Royal Family'

matches = text.match(/(([A-Z]\w*\s*){2,})|(\w{6,})/g)
matches.filter(function(m) {
    return m.match(/\w{6,}/)
});

answered Jul 12, 2012 at 14:14

georg

216k57 gold badges324 silver badges401 bronze badges

Sign up to request clarification or add additional context in comments.

6 Comments

gen_Eric Over a year ago

This seems to work, but it will also match 'I Am Cool', which is not a match as none of the words have >= 6 characters.

gen_Eric Over a year ago

+1, good job with that update. This seems to work just as the OP wants :-)

Andres SK Over a year ago

great solution! just one important thing, the solution should consider special characters. For example, "Princé Hermione" is returning ["Hermione"]; and "superhábilmente" is returning ["superh","bilmente"]

georg Over a year ago

@andufo: that's true. \w, \d and friends are not unicode-aware in javascript (what a shame!)

georg Over a year ago

@andufo: you could replace \w with an explicit character class, like [\wéáè]

|

Vivian River · Accepted Answer · 2012-07-12 14:16:17Z

Okay, here's an idea. This is probably not the very best way to do this, but it might be a good start for you.

In order matching strings like Royal Family and Prince Charles, or perhaps even The United Kingdom, you could write a regex that looks for a succession of words starting with a capital letter in succession.

This might look like this: (A-Z(a-z){5,}* )+

You could then use the replace function to generate a new string with the matches removed and then use your original regex to match single words of a minimum length.

Update: In response to the comment about the other users answer, I have added the {5,} modifier to indicate a capital letter followed by five or more lower case letters and a space, one or more times.

Collectives™ on Stack Overflow

Automatically generate tags from strings with javascript

2 Answers 2

6 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

6 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related