another regex question

[Old Pedant]

I notice you using [^.] instead of [^\.\?\!] so I changed that:
[code]
/(?:^|\.|\?|\!)?[\w]([^\.\!\?]*happy[^\.\!\?]*)(?:\.|\?|\!|$)/

Different results, but still not right.

Also it would seem that (?:^|\.|\?|\!)? is doing nothing at all.

I removed it completely, to get this:
/[\w]([^\.\!\?]*happy[^\.\!\?]*)(?:\.|\?|\!|$)/
and got the same (wrong) results.

[AndrewGSW]

I will say that your (?:^|\.|\?|\!) seemed to have mostly worked. Surprised me.

Yes, I forgot a | in this expression. ^ means the beginning of the line when it is outside of a character class.

So this is a non-capturing group that will be either the beginning of a line (or the string), a full-stop, etc..

But, no, I'm not testing it as thoroughly as you are.

[AndrewGSW]

Well this works with the expressions xelawho proposed, but only if it starts with a Capital letter (the original request):

(?:^|\.|\;|\!|\?)\s?([A-Z][^.]*happy[^\.\!\?]*)(?:\.|\?|\!|$)

but again I'm not testing it beyond the sample phrases, so I won't be offended if it is dropped

Revision:

(?:^|\.|\;|\!|\?)\s?([A-Z][^\.\!\?]*happy[^\.\!\?]*)(?:\.|\?|\!|$)

[AndrewGSW]

I notice you using [^.] instead of [^\.\?\!] so I changed that

@Old Pedant - yes, the ! and ? weren't mentioned until recently

[007julien]

These two pages could maybe inspire a solution : Highlight a word and Select a Sentence

[Old Pedant]

Originally Posted by AndrewGSW

Yes, I forgot a | in this expression. ^ means the beginning of the line when it is outside of a character class.

I knew that. DOH. I was just completely misreading the intent of that sub-expression. Double DOH on me.

[xelawho]

looks to me like

(?:^|\.|\?|\!)?[\w]([^\.\!\?]*happy[^\.\!\?]*)(?:\.|\?|\!|$)

and

[\w]([^\.\!\?]*happy[^\.\!\?]*)(?:\.|\?|\!|$)

are the only one that fulfill the 4 requirements

(the rule is not that it must begin with a capital letter, but that if there is no capital letter preceding the keyword, it captures the string from the start)

thanks everybody!

[Old Pedant]

Originally Posted by xelawho

(the rule is not that it must begin with a capital letter, but that if there is no capital letter preceding the keyword, it captures the string from the start)

And what about a period that is *NOT* followed by a capital letter?

aardvarks whistle. dogs jump when they are happy

Given your rule as stated here, that would return the *entire* string.

Is that truly what you want?

[AndrewGSW]

I suppose my version without requiring a capital would be

(?:^|\.|\;|\!|\?)\s?([\w][^\.\!\?]*happy[^\.\!\?]*)(?:\.|\?|\!|$)

but all my sentences begin with capital letters (sic.) and end with a smile

Just for completeness (for anyone reading) to embed a string in a regex requires use of the RegExp constructor:

var re = new RegExp("^\\/" + resource + "\\/\\d+$");

[AndrewGSW]

Originally Posted by xelawho

looks to me like

(?:^|\.|\?|\!)?[\w]([^\.\!\?]*happy[^\.\!\?]*)(?:\.|\?|\!|$)

and

[\w]([^\.\!\?]*happy[^\.\!\?]*)(?:\.|\?|\!|$)

are the only one that fulfill the 4 requirements

(the rule is not that it must begin with a capital letter, but that if there is no capital letter preceding the keyword, it captures the string from the start)

thanks everybody!

Don't these versions drop the first letter?

Added: I think I would account for zero or more spaces at the beginning as well:

(?:^|\.|\;|\!|\?)\s*([\w][^\.\!\?]*happy[^\.\!\?]*)(?:\.|\?|\!|$)

and also if happy is the very first word:

(?:^|\.|\;|\!|\?)\s*([^\.\!\?]*happy[^\.\!\?]*)(?:\.|\?|\!|$)

[xelawho]

no - they both capture the full stop at the end of the preceding sentence.

[xelawho]

Originally Posted by Old Pedant

And what about a period that is *NOT* followed by a capital letter?

aardvarks whistle. dogs jump when they are happy

Given your rule as stated here, that would return the *entire* string.

Is that truly what you want?

Yes. This is to do spellchecks on content editable divs on a website for professional writers. If they're starting their sentences with lower case letters they have bigger problems than a few extra characters in their output.

maybe a little detail will help. The spellchecker scans the text nodes, so if the innerHTML is

I had a <b>great</b> day

it will scan 3 chunks:
1) I had a
2) great
3) day

Which is kind of meaningless there, but there could be an entire paragraph without formatting and so you just want to show a snippet. A sentence would be ideal, but by the time the spellchecker gets to the text it's all broken up into text nodes. So instead of piecing it back together, there are the four rules (rule 3 revised halfway through this thread):

- Start capturing from the closest word before the variable word that starts with a capital/uppercase.
- If there is no word that starts with a capital before the variable word, start capturing from the start of the string.
- Equally, if the part of the string after the variable word contains a full stop, exclamation mark or question mark, finish capturing at the punctuation.
- If not, capture until the end of the string.

Make sense?

[Old Pedant]

Well, I gave you my answer, using 4 regexps. Which I see no reason can't be combined into one. And then one more regexp eliminates not only any preceding [.?!] but also trims the string on the left side of any spaces.

Unless you are a fanatic about doing it all with a single regexp, I don't see what is wrong with that. What difference does it make if it's 1 regexp or 2 or 5?

[xelawho]

but you're right Andrew - neither of the ones that I thought were winners in #22 capture anything if the word comes at the start of the string

[AndrewGSW]

Originally Posted by xelawho

but you're right Andrew - neither of the ones that I thought were winners in #22 capture anything if the word comes at the start of the string

I haven't tested my most recent solution in great depth but I'll repeat it here:

and also if happy is the very first word:
Code:
(?:^|\.|\;|\!|\?)\s*([^\.\!\?]*happy[^\.\!\?]*)(?:\.|\?|\!|$)

All those escapes \, though, are not necessary in a character class - Added: I'm blaming OP for this (I wouldn't have included them )

(?:^|\.|\;|\!|\?)\s*([^.!?]*happy[^.!?]*)(?:\.|\?|\!|$)