Generate PDF report with HTML data

Question

Hi All,
I am using BI Publisher to generate my reports but I stumbled on a big problem...
I have a table with a column TASK_DESCRIPTION. This column is a CLOB containing a simple HTML coded page. When a user works with the application he sees a nicely formatted page but when I generate a report I get all the html tags in it (< html > < h4 > ...etc...)
Is it possible to somehow convert the source data from the table (which is html code) into readable formatted text when generating the report?
ANY ideas are very welcome!
Regards,
Pawel.

Answer 1

Yes you can do this with an XSL template converting HTML to Formatting objects.
There are a few catches however. First, you must be sure your markup-fragment is valid XHTML, that is all tags must be closed. Luckily the builtin standard HTML-editor does this.
I wasn't able to call BI-publisher from within Apex without the XHTML-fragment being escaped, but i managed to create a query on a table containing the fragment within BI-publisher itself. This leaves the fragment intact, producing a valid XML-document.
This is very important because the XSL attached to the RTF-template must be able to match and convert the XHTML-tags to FO.
Basic setup :
Create a datasource (query) within BIP.
Create the RTF-template and add an extra field (at the top) containing : <?import:file:///C:\xhtml-to-xslfo.xsl?> or whatever location the XSL-file resides.
Now, layout your report using the wizard and change the field containing the XHTML-fragment into <xsl:apply-templates select="TASK_DESRIPTION"/>
There are a few XSL-templates out there converting XHTML to FO. One I found particularly useful : http://www.ibm.com/developerworks/library/x-xslfo2app/xhtml-to-xslfo.xsl
You might want to change the .xsl so it matches tags case insensitive (eg. <xsl:template match="a|A"> )
Another caveat is the html-entities for special characters. When the XML from the datasource is parsed by BIP it doesn't recognise entities like &Agrave ; .
This can be solved by creating a pl/sql function converting these characters to numeric entities like &#192 ;. and use this function in the select statement of your datasource. see http://www.w3schools.com/tags/ref_entities.asp
Now you can call the report from Apex using an URL to the report definiton (see BIP docs for URL syntax and parameter passing)
Not a 'really' integrated solution but I had some nice results.
It would be nice though if we could specify a .dtd containing the references for HTML-entities and a sort of flag wheter the report column should be escaped or not, from within Apex. This would allow us to embed XML-fragments within the XML that Apex produces, resulting in a valid XML doc. Using XSL, you can then match the tags of the XML-fragment and convert it to FO.
Good luck !
Maarten