yonyon100 yonyon100 - 2 months ago 11
Perl Question

Handling GET errors in WWW::Mechanize

I'm using a script that scrapes data from a website using WWW::Mechanize and it's all working great, except for the website itself. Sometimes it just doesn't respond for a short moment and for a given

my $url = 'http://www.somesite.com/more/url/text'
I will have this error on
$mech->get($url)
:

Error GETing http://www.somesite.com/more/url/text: Can't connect to www.somesite.com:443 at ./trackSomesite.pl line 34.


This error is something that occurs once in a while with no recognizable pattern and from my experience with the website I'm dealing with, it's because of server instabilities.

I want to be able to know specifically that this error occurred and not other errors like
Too many requests
.
My question is how can I get my script to handle this error and not die?

Answer

Either wrap your $mech->get(...) requests in an eval block or use autocheck => 0, then check the $mech->status code and/or $mech->status_line to decide what to do.

Here is an example:

#!/usr/bin/env perl

use WWW::Mechanize;

use constant RETRY_MAX => 5;

my $url = 'http://www.xxsomesite.com/more/url/text'; # Cannot connect

my $mech = WWW::Mechanize->new( autocheck => 0 );

my $content = fetch($url);

sub fetch {
    my ($url) = @_;

    for my $retry (0 .. RETRY_MAX-1) {
        my $message = "Attempting to fetch [ $url ]";
        $message .= $retry ? " - retry $retry\n" : "\n";
        warn $message;

        my $response = $mech->get($url);
        return $response->content() if $response->is_success();

        my $status = $response->status;
        warn "status = $status\n";

        if ($response->status_line =~ /Can['']t connect/) {
            $retry++;
            warn "cannot connect...will retry after $retry seconds\n";
            sleep $retry;
        } elsif ($status == 429) {
            warn "too many requests...ignoring\n";
            return undef;
        } else {
            warn "something else...\n";
            return undef;
        }
    }

    warn "giving up...\n";
    return undef;
}

Output

Attempting to fetch [ http://www.xxsomesite.com/more/url/text ]
status = 500
cannot connect...will retry after 1 seconds
Attempting to fetch [ http://www.xxsomesite.com/more/url/text ] - retry 1
status = 500
cannot connect...will retry after 2 seconds
Attempting to fetch [ http://www.xxsomesite.com/more/url/text ] - retry 2
status = 500
cannot connect...will retry after 3 seconds
Attempting to fetch [ http://www.xxsomesite.com/more/url/text ] - retry 3
status = 500
cannot connect...will retry after 4 seconds
Attempting to fetch [ http://www.xxsomesite.com/more/url/text ] - retry 4
status = 500
cannot connect...will retry after 5 seconds
giving up...
Comments