Question

I’m trying to parse an html output file from a program called mirdeep. I have gotten the script to the point where the data has been extracted into an array of arrays and I am trying to extract the specific information that I want from the appropriate columns. For some odd reason, I keep getting output that shouldn’t be possible.

Bascially it should be pulling the counts from the mirbase miRNAs that were detected by mirdeep AND and the undetected ones, but leave the stat table and novel section alone.

However, the counts currently being pulled for the undetected section are strangely getting stored incorrectly. (Because of the way they were output, Some of the so called “star” and “mature” data were put into separate lines instead of being in the same line, so I am trying to recombine them. I’m fairly certain that an error in the recombination is the source of the problem.

I was checking my progress with print statements and noticed that some of the data that should only have either a mature count or a star count have both, and that the second number seems to be always equal to the mature count of the row below in the html file, which doesnt make sense to me. Can somebody help me track down whats gone wrong? This is my first time using packages for object orientation extensively, so there could be any kind of error in here.

Heres the script and an example data file

Just created an account, for some reason I can’t install HTML::TableExtract;

Running make install make test had returned bad status, won’t install without force Failed during this command: GAAS/HTML-Parser-3.68.tar.gz : make_test NO MSISK/HTML-TableExtract-2.10.tar.gz : make_test NO

seems my perl installation was screwey, i was trying to use bioperl and for some reason the bums use fink, which installed its OWN copy of perl instead of providing the libraries for the main system installation of perl. Does anybody know how to make fink play nice with the main MAC perl installation (& cpan)? When I use the main perl it works, but I can’t install the module to the older fink perl, which i have to use to get access to bioperl. Wonderful isn’t it.

P.S. The thing that bothers me the most about this, is that its not happening to every data item, and the ones that it IS happening to are mostly ones that should not be getting two values in the first place. It seems somewhat pseudorandom, but deterministic in that it happens to the same items every time.

Answer 1

That is hundreds of lines of code. If you are just trying to extract the table out of the HTML, use HTML::TableExtract.

For example:

#!/usr/bin/perl

use strict; use warnings;
use HTML::TableExtract;
use YAML;

my $te = HTML::TableExtract->new(
    headers => [
        '',
        'novel miRNAs',
        'known miRBase miRNAs',
        '',
        '',
    ],
    slice_columns => 0,
);

$te->parse_file('t.html');

for my $table ( $te->tables ) {
    for my $row ( $table->rows ) {
        print Dump $row;
    }
}