topic sed/awk/perl help in Operating System - HP-UX

sed/awk/perl help

allanm77 — Tue, 24 Jan 2012 01:25:10 GMT

Hi All,

I have a following piece of output from a wget command and I want to strip the html tags.

    <td>INACCESSIBLE</td>
    <td>hostname:port <a title="inspect" href="dest/hostname:port">[I]</a> <a title="debugger" href="http://hostname:port/debug">[D]</a> <a title="browser" href="http://hostname/Browser.jsp/Url=http%3A//hostname%3A8501/oa/ww&amp;serviceUrl=http%3A//hostname%3A1port/service">[SB]</a></td>

Want the output something like this -

INACCESSIBLE

hostname:port

I have been successful with a bunch of awk/sed but it would be great if it can be done in a single command.

Thanks,

Allan

Re: sed/awk/perl help

Dennis Handly — Tue, 24 Jan 2012 04:40:03 GMT

>Want the output something like this:

>hostname:port

Don't you want something like this?:

hostname:port [I] [D] [SB]

Or do you just want to ignore anything in the <a ...> ... </a> blocks?

Re: sed/awk/perl help

allanm77 — Tue, 24 Jan 2012 06:04:48 GMT

Yes, that would do too if the html works out and points to the right links.

Thanks,
Allan.

Re: sed/awk/perl help

allanm77 — Tue, 24 Jan 2012 20:06:55 GMT

Dennis, perl would do too but sed / awk is preferable

thx.

Re: sed/awk/perl help

H.Merijn Brand (procura — Wed, 25 Jan 2012 08:17:17 GMT

I'd say not to try to get into html digging using regular expressions (with whatever language). You'll end up changing them all the time, as the html changes. Been there done that.

There are several fine ways to "convert" HTML to something else. If you want just text, use the proper tools (e.g. lynx):

$ lynx -dump file.html
$ lynx -dump http://some.host.com/index.htm

If you want to use a scripting language, WWW::Mechanize, LWP, LWP::Simple, LWP::UserAgent, and HTML::TreeBuilder are your friends in perl:

#!/usr/bin/perl

use strict;
use warnings;

use LWP::Simple;
use HTML::TreeBuilder;

my $site = get ("http://h30499.www3.hp.com/t5/Languages-and-Scripting/bd-p/itrc-150");
my $tree = HTML::TreeBuilder->new;
$tree->parse_content ($site) or die "Cannot parse as HTML\n";

# Print the whole page formatted
print $tree->as_HTML (undef, "  ", {});

# Print all <a> tags pointing to something with scripting in it
for ($tree->look_down (_tag => "a", href => qr{scripting}i)) {
    print "A: ", $_->as_text, "\t=> ", $_->attr ("href"), "\n";
    }

Re: sed/awk/perl help

allanm77 — Wed, 25 Jan 2012 19:44:36 GMT

Thanks Merjin!

But the problem is that the wget is part of a script , so if there is a Perl one-liner or sed/awk combination that is what is preferable.

Allan.

Re: sed/awk/perl help

allanm77 — Wed, 25 Jan 2012 23:29:36 GMT

Used sed to fix this.

sed 's/<[^>]*>//

Thanks All.

Re: sed/awk/perl help

H.Merijn Brand (procura — Wed, 25 Jan 2012 23:47:06 GMT

The "get ()" in the perl example is almost exactly what wget does. My perl scriplet is just an example. Modify to your hearts content.
One-lines to analyze HTML? Forget it! (it works once, but you'll end up with very very long lines to stick to a single line.