Strip all HTML tags except links

Question

I am trying to write a regular expression to strip all HTML with the exception of links (the <a href and </a> tags respectively. It does not have to be 100% secure (I am not worried about injection attacks or anything as I am parsing content that has already been approved and published into a SWF movie).

The original "strip tags" regular expression I'm using was <(.|\n)+?>, and I tried to modify it to <([^a]|\n)+?>, but that of course will allow any tag that has an a in it rather than one that has it in the beginning, with a space.

Not that it should really matter, but in case anyone cares to know I am writing this in ActionScript 3.0 for a Flash movie.

Answer 1

<(?!\/?a(?=>|\s.*>))\/?.*?>

Try this. Had something similar for p tags. Worked for them so don't see why not. Uses negative lookahead to check that it doesn't match a (prefixed with an optional / character) where (using positive lookahead) a (with optional / prefix) is followed by a > or a space, stuff and then >. This then matches up until the next > character. Put this in a subst with

s/<(?!\/?a(?=>|\s.*>))\/?.*?>//g;

This should leave only the opening and closing a tags

Answer 2

I keep going on about it, but there's no way I can recommend regexr too often. It's fantastic for testing this type of things.

Answer 3

In general there are problems with this approach. Regexes are best for 'flat' text matches - nested data pushes regex engines into areas for which they are not designed. General HTML parsing needs a parser not a regex engine (Google for the difference between regular and context-free languages if you want the full technical details).

It is easy to strip out all tags by replacing /</ and />/ with the empty string or their entity equivalents but selectively filtering HTML using regexes will be vulnerable to a wide range of accidental or malicious inputs breaking things.

Answer 4

Here you go:

{<(?!i|b|h[1-6]|/i|/b|/h[1-6][\s|>|/])[^>]*>}

Answer 5

How about

<[^a](.|\n)+?>

?

Answer 6

strip_tags() does this.

Here, I am including all <a><p><font><b><i><sup> tags and outputting a tidied version:

cat input.htm | tr -d '\n' | php -r '$input=fgets(STDIN); echo strip_tags($input,"<a><p><font><b><i><sup>");' | tidy -i -wrap 0 -o output.htm