Extract values from html table with a RegEx in bash/Perl

Question

I want to monitor my oki printer with munin, so I am trying to adapt this plugin to my printer.

The table of pages in my printer http server is:

<table width="560" border="0" cellspacing="2" cellpadding="3">
    <tr class="sub_item_color">
        <td  class="normal" width="200" align="right" valign="bottom" rowspan="2">Media Size</td>
        <td  class="normal" width="90" align="left">Color</td>
        <td  class="normal" width="90" align="left">Color</td>
        <td  class="normal" width="90" align="left">Mono</td>
        <td  class="normal" width="90" align="left">Mono</td>
    </tr>
    <tr class="sub_item_color">
        <td  class="normal" width="90" align="left">A3/Tabloid</td>
        <td  class="normal" width="90" align="left">A4/Letter</td><td  class="normal" width="90" align="left">A3/Tabloid</td>
        <td  class="normal" width="90" align="left">A4/Letter</td>
    </tr>
    <tr class="sub_item_color">
        <td  class="normal" width="200" align="left">Total Impressions</td>
        <td  class="normal" width="90" align="right">21906</td>
        <td  class="normal" width="90" align="right">33491</td>
        <td  class="normal" width="90" align="right">2084</td>
        <td  class="normal" width="90" align="right">4460</td>
    </tr>
    <tr class="sub_item_color">
        <td  class="normal" width="200" align="left">Total A4/Letter Impressions</td>
        <td  class="normal" colspan="2" align="center"><b>Color:77303</B></td>
        <td  class="normal" colspan="2" align="center"><b>Mono:8628</B></td>
    </tr>
</table>

That munin script is doing this:

infopage=`wget -q -O - http://root:$password@$destination/printer/printerinfo_top.htm | perl -p -e 's/\n/ /m'`
echo tray1.value    `echo $infopage | perl -p -e 's/^.+Tray\ 1\ Page\ Count\:\ \<\/TD\>\<TD\ WIDTH\=\"94\"\>([0-9]+)\<.+$/$1/'`

How I could get the total impressions?

Answer 1

Solution implemented as a Unix filter like in the question, only much more readable and declarative thanks to XPath.

#!/usr/bin/env perl
use 5.010;
use strictures;
use HTML::TreeBuilder::XPath qw();
use List::Util qw(sum);
my $tree = HTML::TreeBuilder::XPath->new;
$tree->parse_content(<>);
say sum map { s[.*:][]; $_ } $tree->findnodes_as_strings('//table/tr/td[@colspan=2]/b');

---

wget -q -O - http://… | perl sum-total-impressions.pl