javascript - Split string into sentences - ignoring abbreviations for splitting - Stack Overflow

admin管理员组

文章数量:1318564

I'm trying to split this string into sentences, but I need to handle abbreviations (which have the fixed format x.y. as a word:

content = "This is a long string with some numbers 123.456,78 or 100.000 and e.g. some abbreviations in it, which shouldn't split the sentence. Sometimes there are problems, i.e. in this one. here and abbr at the end x.y.. cool."

I tried this regex:

content.replace(/([.?!])\s+(?=[A-Za-z])/g, "$1|").split("|");

But as you can see there are problems with abbreviations. As all the abbreviations are of the format x.y. it should be possible to handle them as a word, without splitting the string at this point.

"This is a long string with some numbers 123.456,78 or 100.000 and e.g.", 
"some abbreviations in it, which shouldn't split the sentence."
"Sometimes there are problems, i.e.", 
"in this one.", 
"here and abbr at the end x.y..",
"cool."

The result should be:

"This is a long string with some numbers 123.456,78 or 100.000 and e.g. some abbreviations in it, which shouldn't split the sentence."
"Sometimes there are problems, i.e. in this one.", 
"here and abbr at the end x.y..",
"cool."

I'm trying to split this string into sentences, but I need to handle abbreviations (which have the fixed format x.y. as a word:

content = "This is a long string with some numbers 123.456,78 or 100.000 and e.g. some abbreviations in it, which shouldn't split the sentence. Sometimes there are problems, i.e. in this one. here and abbr at the end x.y.. cool."

I tried this regex:

content.replace(/([.?!])\s+(?=[A-Za-z])/g, "$1|").split("|");

But as you can see there are problems with abbreviations. As all the abbreviations are of the format x.y. it should be possible to handle them as a word, without splitting the string at this point.

"This is a long string with some numbers 123.456,78 or 100.000 and e.g.", 
"some abbreviations in it, which shouldn't split the sentence."
"Sometimes there are problems, i.e.", 
"in this one.", 
"here and abbr at the end x.y..",
"cool."

The result should be:

"This is a long string with some numbers 123.456,78 or 100.000 and e.g. some abbreviations in it, which shouldn't split the sentence."
"Sometimes there are problems, i.e. in this one.", 
"here and abbr at the end x.y..",
"cool."

• javascript
• regex
• string

Share Improve this question Follow asked Jan 14, 2016 at 8:21 user3142695user3142695 17.4k5555 gold badges195195 silver badges375375 bronze badges 1

• What do you mean by string manipulation? – user3142695 Commented Jan 14, 2016 at 8:26

Add a ment  | 

2 Answers 2

Sorted by: Reset to default 7

The solution is to match and capture the abbreviations and build the replacement using a callback:

var re = /\b(\w\.\w\.)|([.?!])\s+(?=[A-Za-z])/g; 
var str = 'This is a long string with some numbers 123.456,78 or 100.000 and e.g. some abbreviations in it, which shouldn\'t split the sentence. Sometimes there are problems, i.e. in this one. here and abbr at the end x.y.. cool.';
var result = str.replace(re, function(m, g1, g2){
  return g1 ? g1 : g2+"\r";
});
var arr = result.split("\r");
document.body.innerHTML = "<pre>" + JSON.stringify(arr, 0, 4) + "</pre>";

Regex explanation:

• \b(\w\.\w\.) - match and capture into Group 1 the abbreviation (consisting of a word character, then . and again a word character and a .) as a whole word
• | - or...
• ([.?!])\s+(?=[A-Za-z]):
  • ([.?!]) - match and capture into Group 2 either . or ? or !
  • \s+ - match 1 or more whitespace symbols...
  • (?=[A-Za-z]) - that are before an ASCII letter.

Given your example, I have managed to achieve what you are after through the use of this expression: (?<!\..)[.?!]\s+ (example here).

This expression will look for period, question mark or exclamation mark characters which are not preceded by a character and a period.

You would then need to replace them with the | character and finally, you replace the | with .\n.

本文标签： javascriptSplit string into sentencesignoring abbreviations for splittingStack Overflow