web scraping tool or library that automatically finds text content without rules set - screen-scraping

Question

Is there a web scraping tool or library that auto-detects repeating HTML blocks and scrapes the text content inside the blocks, thus removing the need for human to manually input the rules - CSS selectors or xpath to find the content?
This is based on the assumptiom that modern content website is generated dynamically by server-side languages such as PHP or Python. The content is almost always rendered by a for loop in the template, hence the repeating HTML blocks can always be found. An example:
<div id="content">
<div class="blog entry">
<div class="title">
<h1>1st post</h2>
</div>
<div class="content">
<p>...</p>
</div>
</div>
<div class="blog entry">
<div class="title">
<h1>2nd post</h2>
</div>
<div class="content">
<p>...</p>
</div>
</div>
<div class="blog entry">
<div class="title">
<h1>3rd post</h2>
</div>
<div class="content">
<p>...</p>
</div>
</div>
</div>
Libraries like bautiful soap and scrapy rely on human to input the rules before the scraping can be carried out. They are not what I want.

Answer 1

Haven't used it, but heard about scrapely:
Unlike most scraping libraries, Scrapely doesn't work with DOM trees
or xpaths so it doesn't depend on libraries such as lxml or libxml2.
Instead, it uses an internal pure-python parser, which can accept
poorly formed HTML. The HTML is converted into an array of token ids,
which is used for matching the items to be extracted.
Scrapely extraction is based upon the Instance Based Learning
algorithm and the matched items are combined into complex objects
(it supports nested and repeated objects), using a tree of parsers,
inspired by A Hierarchical Approach to Wrapper Induction

Answer 2

You might want to look at my scraping library. It doesn't work automatically nor does it detect repeated parts. But it comes close, since it doesn't need rules at all and instead uses templates, which you can get directly from the html you have.
E.g. with your example above, the template to read all the posts in 2 arrays is:
<div id="content">
<div class="blog entry">
<div class="title">
<h1>{title:=.}</h1>
</div>
<div class="content">
<p>{content:=.}</p>
</div>
</div>*
</div>

Answer 3

You may try HTQL:
import htql;
a=htql.Browser();
p,b=a.goUrl('http://channel9.msdn.com/Blogs/Vector/Announcing-BUILD-2012');
htql.query(p, '&html_main_text');
p,b=a.goUrl('http://stackoverflow.com/questions/tagged/screen-scraping');
htql.query(p, '&html_main_text');