Re^17: Need help with WWW::Mechanize and Chrome cookies

[bakiperl]

Corion,
This code did the trick for image files but not for PDFs. I also noticed that it works only for simple urls in the page. It did not work for me when the images are displayed using JavaScript.
I was hoping to find a way to block the browser from displaying the documents by using something like the Content-Disposition option.

Content-Disposition: attachment; filename=$filename

This has worked very well with WWW::Mechanize.
Thank you.

[Corion]

Your problem statement seems to continously change. I don't see/understand how you can get from "image download using its URL" to "image displayed using Javascript" without changing the problem entirely.

Maybe you can come up with a small, self-contained program that reproduces your problem, and also shows the response you receive and also tell us what parts you have already investigated, instead of letting us guess. That way, we can more easily reproduce what you see and maybe suggest better approaches.

If an image gets created using Javascript, you will have to fetch the image data from the browser memory, most likely by using some Javascript to fetch that image data. You can run arbitrary Javascript on a page using ->evaluate_in_page.

[bakiperl]

Corion, forget the JavaScript for now. The goal is to use a single method to download all type of files (images, pdfs, CSVs etc..)
According to the documentations, the following method should be able to do it.

my @links = $mech->find_all_links(url_regex => qr/\.pdf/i);
my @urls = map { $_->url_abs } @links;
foreach (my $foo @urls )
{
   my $abs_path = "C:/path";
   $mech->set_download_directory( $abs_path );    
   $mech->get($foo);
}

Unfortunately this method works only with files that the browser normally does not load such as (.csv). If you use this method with images (.jpg etc...) or PDFs (.pdf) the documents will load in the browser and the download fails. This exact same task has worked with WWW::Mechanize using the following method

my $filename = "C:/path/filename";
my $foo = "http link";
$mech->get($foo, ':content_file'=>$filename);

The workaround that you have suggested below didn't work with PDFs and CSVs. it worked with images only.

my $response = $mech->get($url);
my $c = $mech->content; # Dummy request to initialize everything
open my $output, '>:raw', '/tmp/output.jpg';
print { $output } $response->decoded_content;

[Corion]

The following gets at the content of images displayed on a page:

#!perl
use strict;
use warnings;
use 5.012;
use WWW::Mechanize::Chrome;

use Log::Log4perl ':easy';

Log::Log4perl->easy_init($TRACE);


use File::Temp 'tempdir';
use Cwd;
my $tempdir = tempdir();
my $mech = WWW::Mechanize::Chrome->new(
    headless => 1,
    data_directory => $tempdir,
    download_directory => cwd(),
);

use Data::Dumper;

my $res = $mech->get('https://egp.rutgers.edu/cgi/wmc.pl');
say Dumper $mech->getResourceTree_future()->get;

my $link = $mech->xpath( '//a[text()="MY IMAGE"]', single => 1  );
$mech->click($link);
$mech->sleep(1);

my $resources = $mech->getResourceTree_future()->get;
my @images = grep { $_->{type} eq 'Image' } @{$resources->{resources}}
+;

my $image = $mech->getResourceContent_future( $images[0]->{url} )->get
+->{content};

open my $fh, '>:raw', 'test.jpg';
print $fh $image;

Note that you will need a way to find which image is the one you want.