If you don't need JavaScript support, I'd use Mojo. Firstly an experiment to find the stuff:

#!/usr/bin/perl

use strict;
use warnings;
use Mojo::DOM;
use feature 'say';

my $html = '<div class="grouped-item  product-purchase-wrapper-1">One<
+/div><div class="grouped-item  product-purchase-wrapper-7117">7117</d
+iv>';

my $dom = Mojo::DOM->new( $html );

# find each div with a class beginning grouped-item product-purchase-w
+rapper
foreach my $div ( $dom->find('div[.class^=i"grouped-item product-purch
+ase-wrapper"]')->each ){
  say $div->text;
}
[download]

Prints:

One
7117
[download]

Getting it from some live site:

#!/usr/bin/perl

use strict;
use warnings;
use Mojo::UserAgent;
use feature 'say';
 
my $ua  = Mojo::UserAgent->new;
my $url = 'https://urlgoeshere';

my $dom = $ua->get( $url )->res->dom;

foreach my $div ( $dom->find('div[.class^=i"grouped-item product-purch
+ase-wrapper"]')->each ){
  say $div->text;
}
[download]

See Mojo::DOM, Mojo::UserAgent, Super search for more mojo goodness.

Update: if you do need JavaScript support I'd suggest automating Chrome using WWW::Mechanize::Chrome, and using the xpath method.

Probably there are two components you need to think about: fetching the page and processing it, and a multitude of ways to go about the task. The old school way to go about this might be to use LWP::Simple to fetch the page and HTML::Tree to parse and manipulate the contents. A more modern approach might be to use one of the WWW::Mechanize::xxx suite of modules such as WWW::Mechanize::Firefox.

"...such as WWW::Mechanize::Firefox."

Just an FYI, this has been difficult for some time due to the dependency on an old, unsupported Firefox. See Important Notice.

Yes, changes to Firefox caused problems! I use some Firefox plug-ins that aren't compatible with current releases (like SQLite Mgr). So I installed Waterfox and use it to run these older Firefox plug-ins. Whether or not Mechanize::Firefox would work with Waterfox, I don't know. It might. A more promising avenue is probably Mechanize::Chrome. I am having trouble getting that critter built so I can't say how well it works or doesn't work.

Update: Default cpan install of version 0.60 hangs on one test. However, Corion helped me and..
cpan CORION/WWW-Mechanize-Chrome-0.61.tar.gz appears to work to force a 0.61 installation. Some test cases don't pass but I don't think that matters. At least installation of the 0.61 version does not "hang" and it gets back to a command prompt.

As another addendum, I don't recommend Waterfox for browsing. I current just use it to run an SQLite tool that I like and know how to use. I've never bothered to fiddle with the command line tool for SQLite. The GUI tool gives the info that I need and I can test out SQL commands. Once I have a working SQLite command, implementing that in Perl code is fairly straightforward - the magic of the Perl DBI!

I would start by reading Cpan Mech Docs and then take a look at some Mech examples. Then I would start "hacking" and experimenting and see how far you can get with the base Mech module. If you are using a public, heavily trafficked web site, then show us the URL.

Also be aware of the potential impact that your code could have on the target web site. I have one application that "beats up" one web site pretty good. But I have agreement with the site owner about what hours and what days my application can run. This is an important consideration if you are going to retrieve a lot of data.

Update: s/Java/Javascript/; #Completely different things!

'Perl controls the browser and has the browser execute the Java code'

JavaScript, not Java.

Thanks for the correction!

• Bruce Gray - Refactoring and Readability: Crouching Regex, Hidden Structures (youtube, TPC in the cloud 2020)
• Web::Scraper
• (browser automation) Firefox::Marionette
• (browser automation) Playwright