Negate link search in Mechanize

Anonymous Monk has asked for the wisdom of the Perl Monks concerning the following question:

I am trying to scrape some links using Mechanize and cannot find a way to negate the search argument. I need to find links that do not contain (for example) 'foo'. Tried super search, perlre & Mechanize docs with no help, but feel like I'm overlooking something obvious. Can anyone help?

# this finds the links I do not want:
my @links = $mech->find_all_links(url_regex => qr/foo/);

# these (as well as some really bizarre attempts) don't work:
my @links = $mech->find_all_links(url_regex => !qr/foo/);
my @links = $mech->find_all_links(url_regex !~ qr/foo/);
[download]

Comment on Negate link search in Mechanize Download Code

Replies are listed 'Best First'.
Re: Negate link search in Mechanize by Plankton (Vicar) on Nov 02, 2011 at 05:04 UTC
Why not just get all the links like so ... `... use WWW::Mechanize; my $mech = WWW::Mechanize->new(); $mech->get( $url ); my @links = $mech->links();` [download] ... then iterate over @links. Something like ... `for my $l (@links) { next if $l->url() =~ /foo/; push ( @non_foo_links, $l ); ...` [download]	[reply] [d/l] [select]
Re^2: Negate link search in Mechanize by Anonymous Monk on Nov 02, 2011 at 05:18 UTC
Why not just get all the links like so ... Because I'm lazy (a Perl virtue, you know). But thanks, I was too focused on letting Mechanize do the work and did overlook the obvious. Maybe time to get some sleep.	[reply]
Re: Negate link search in Mechanize by mwp (Hermit) on Nov 02, 2011 at 04:57 UTC
You might be over-thinking it. `my @links = grep $_->url !~ /foo/, $mech->links;` Update: I might be under-thinking it. :-) Update 2: ology points out Negation/Complement of a Regex, which leads into a twisty labyrinth of prior art, including Negating Regexes: Tips, Tools, And Tricks Of The Trade. mwp	[reply] [d/l]
Re: Negate link search in Mechanize by Anonymous Monk on Nov 02, 2011 at 13:34 UTC
`my @links = $mech->find_all_links(url_regex => qr/^(?!.*foo)/);` [download]	[reply] [d/l]