QUERY: comparing website contents

[andrewwan1980]

I've got two websites, one original, the other based off the original.

I like to diff/compare the websites using diff automatic comparison tools to see what text/information has changed. The problem is, the HTML code and layout has been changed drastically so I can't do a straight text file compare. What am interested in is purely the raw content (paragraphs, sentences, etc.). The original site has no javascript, onmouseover hovers, etc. The new revamped website has javascript, onmouseover hovers, popups, etc.

How can I create a script (Perl? C++?) so that it extracts the main text BODIEs from both sites? I guess also have to specify starting & ending delimiters. Once extracted, it would need to convert < p ></ p > paragraph tags, and strip out < a onmouseover... > anchor links (while maintaining the word inbetween the anchor link ofcourse). The new website uses two spaces after each full stop while the old website uses 1 space. Will this matter?

Once we got the plain text, how to wrap the paragraphs after 80 characters per line... so that we can easily do file compares.



And please do not suggest copying and pasting the text into NotePad or Word. I said 'website' which means they contain dozens of html files (probably 100s). Plus, I like a script to automate this compare process so I can repeat the process in future and remind myself of diffs....

[PeejAvery]

[ moved ]

[PeejAvery]

This seems like it would be a process near to impossible. You stated that the websites are drastically different from each other. You would have to implement multiple combinations of regular expressions and algorithms to even start the process.

[andrewwan1980]

I need a tool to get me the substring between delimiters then 79char line wrap the result and then diff... for both oldsite/old1.htm and newsite/new1.htm

As for web crawling, old site is local, new site is online. But I rather hard code the URLs in a big list (mapping).

I think I'll use Perl (maybe Python), to:

1. for each item in mapping list
1.1 download newsite/html file
1.2 substring using newsite delimiters on newsite file
1.3 substring using oldsite delimiters on oldsite file
1.3 html2txt/hindent both oldsite & newsite file and line wrap 79char and put into 2 separate new folders (diff1, diff2).
1.4 repeat through mapping list

After that I can use Beyond Compare to compare the diff1 & diff2 folders. Hopefully both corresponding text files will be 79char line wrapped with whitespace down to 1 char (eliminating 2 or more consecutive spaces, and tab spaces). Also maintain carriage returns?