Search code examples
htmlrubyparsingcommentsnokogiri

Using ruby and nokogiri to parsing HTML using HTML comments as markers

How could I use ruby to extract information from a table consisting of these rows? Is it possible to detect the comments using nokogiri?

<!-- Begin Topic Entry 4134 --> 
    <tr> 
        <td align="center" class="row2"><image src='style_images/ip.boardpr/f_norm.gif' border='0'  alt='New Posts' /></td> 
        <td align="center" width="3%" class="row1">&nbsp;</td> 
        <td class="row2"> 
            <table class='ipbtable' cellspacing="0"> 
                <tr> 

<td valign="middle"><alink href='http://www.xxx.com/index.php?showtopic=4134&amp;view=getnewpost'><image src='style_images/ip.boardpr/newpost.gif' border='0'  alt='Goto last unread' title='Goto last unread' hspace=2></a></td> 

                    <td width="100%"> 
                    <div style='float:right'></div> 
                    <div> <alink href="http://www.xxx.com/index.php?showtopic=4134&amp;hl=">EXTRACT LINK 1</a>  </div> 
                    </td> 
                </tr> 
            </table> 
            <span class="desc">EXTRACT DESCRIPTION</span> 
        </td> 
        <td class="row2" width="15%"><span class="forumdesc"><alink href="http://www.xxx.com/index.php?showforum=19" title="Living">EXTRACT LINK 2</a></span></td> 
        <td align="center" class="row1" width='10%'><alink href='http://www.xxx.com/index.php?showuser=1642'>Mr P</a></td> 
        <td align="center" class="row2"><alink href="javascript:who_posted(4134);">1</a></td> 
        <td align="center" class="row1">46</td> 
        <td class="row1"><span class="desc">Today, 12:04 AM<br /><alink href="http://www.xxx.com/index.php?showtopic=4134&amp;view=getlastpost">Last post by:</a> <b><alink href='http://www.xxx.com/index.php?showuser=1649'>underft</a></b></span></td> 
    </tr> 
<!-- End Topic Entry 4134 -->
-->
Solution

  • You could implement a Nokogiri SAX Parser. This is done faster than it might seem at first sight. You get events for Elements, Attributes and Comments.

    Within your parser, your should rememeber the state, like @currently_interested = true to know which parts to rememeber and which not.