Regex group without capturing

Question

I created a regex to parse the src-attribute (hosting the images now on cloudfront) and alt-attribute from images. I'm substituting them to turn them into Markdown.

Some are embedded into an attribute. I created an optional group for the surrounding (<a href.*)? Since I use backreferences to create my Markdown I don't want these optional groups to be captured. How can I do this?

This is my regex:

/(<a href.*)?<img.*?src="http:\/\/www.example.com\/uploads(.*?)" alt="(.*?)".*?\/?>(<\/a>)?/gm

This is my substituion:

![$2](http://example.cloudfront.net/images$1)

[Don't parse HTML with regex!](http://stackoverflow.com/a/1732454/418066) — Biffen, Mar 15 '15 at 15:09
In general I agree with @Biffen - it's a bad idea to parse html with regex, unless it's a known subset/source etc. The keywords you're looking for here to find results are "non capturing groups" — James Thorpe, Mar 15 '15 at 15:11
I agree with @Biffen. But if you want group wthout capture use this syntax: (?: group content ) — tutankhamun, Mar 15 '15 at 15:12
@Biffen I'm migrating my blog. Therefore I'm manually migrating the posts (to clean of dust etc.). I'm just using this regex to make the cumbersome image conversion a tad easier before pasting into my post editor. — Hedge, Mar 15 '15 at 15:19
Either answer the question so I can mark it as answered or I will delete it within a few hours (so people don't get the wrong idea about regex'ing HTML) — Hedge, Mar 15 '15 at 15:32
I don't see why you need to match the enclosing `` element. Those parts of your regex are optional, so they don't narrow down the possible matches at all. Have you tried it without them? — Alan Moore, Mar 15 '15 at 15:36
@Hedge I think this question should be deleted. Parse html by regex is a doubtful idea and regex syntax documented good. — tutankhamun, Mar 15 '15 at 15:43

score 0 · Accepted Answer · answered Mar 15 '15 at 16:01

0

Agreed that parsing HTML with regex is not a nice idea. But since you are wanting to hand out some reputation points I'll be a cherry picker... :-)

/(?:<a href.*)?<img.*?src="http:\/\/www.example.com\/uploads(.*?)" alt="(.*?)".*?\/?>(?:<\/a>)?/gm

I will state again here at the end for clarity that parsing HTML with regex is not a good idea.

answered Mar 15 '15 at 16:01

Peter Bowers

3,063
1
10
18

I'm not using that regex programmatically but just to convert HTML to markdown by hand. – Hedge Mar 15 '15 at 16:20
Works for me - I'm almost at my 1000 level with your little boost! :-) – Peter Bowers Mar 15 '15 at 16:27

Regex group without capturing

1 Answers1