Unless you are using someone-else's website to grab HTML content from, using Selenium may sound a bit cowboy-ish... 
For solid and repeatable migration outcome, the preference would be to have access to the content management system of the original site, e.g. database or some sort of extract of articles, and then writing a migration script/program which will copy articles' content across.
In most cases we just create a simple servlet web application which does this, deploy it to a target Liferay portal and then invoke the migration run by calling the URL of this servlet.
Using this approach a number of large (and multilingual) content migrations have been started and completed in a couple of days, from custom CMS, Domino, WordPress, etc.
I do not think that these tasks could be achieved using Selenium, does not matter how exciting the HTML hacking process could be for whoever would attempt doing it this way.
Just another relevant link: Web site WEB CONTENT(e.g. html) migration to Liferay
For solid and repeatable migration outcome, the preference would be to have access to the content management system of the original site, e.g. database or some sort of extract of articles, and then writing a migration script/program which will copy articles' content across.
In most cases we just create a simple servlet web application which does this, deploy it to a target Liferay portal and then invoke the migration run by calling the URL of this servlet.
Using this approach a number of large (and multilingual) content migrations have been started and completed in a couple of days, from custom CMS, Domino, WordPress, etc.
I do not think that these tasks could be achieved using Selenium, does not matter how exciting the HTML hacking process could be for whoever would attempt doing it this way.
Just another relevant link: Web site WEB CONTENT(e.g. html) migration to Liferay