在javascript中，我无法弄清楚如何使正则表达式只替换捕获而不是匹配

Question

The following function is meant to remove random articles (parts of speech) from the text. Eventually the percentages will be user-adjustable, and the regex more sophisticated to catch word boundaries better, etc. It is replacing (and about 50/50), but it's also squashing the spaces (which are matched, but not captured). I think I'm being really bone-headed here but I can't figure out the proper syntax... can anyone help?

function posArticles(t) {
   var text = t;
   var re = / (a|the|an) /g;    
   var rArray;

   text = text.replace(re, function(_, m) {
       if (Math.floor(Math.random()*101) < 50) return '';
       else return m;
   });

   return text;
}

I realize that this has to do with the positional/optional arguments to the anon function, but I can't figure out which is the match and which is the capture and so forth.

Answer 1

There are numerous ways you could do this, but I think your best bet is to use \\b – a zero-width match for a "word boundary." That guarantees that you're getting "the" and not "there" or whatever, but doesn't match the spaces around it.

Thus, use re = /\\b([Aa]n?|[Tt]he)\\b/;

Answer 2

I realize that this has to do with the positional/optional arguments to the anon function, but I can't figure out which is the match and which is the capture and so forth.

First argument passed to your callback function is whole match (ie: _ = ' the ' ). Next argument are your captured groups ( m = 'the' ). Callback function is replacing whole match, so if you are including spaces in your expresion, they will also be replaced.

Answer 3

if (Math.floor(Math.random()*101) < 50) return ' ';

返回空格而不是空字符串:)