WikiRenderer |  |
This is a small sample Forrest webapp (available for download here), which demonstrates Forrest's support for Wiki content by rendering a directory full of Wiki text files. While mainly done as an exercise and demonstration, having PDF renditions of Wiki content (including aggregations) may prove useful to some.
Wiki parsing
All the hard work of parsing is done by the Chaperon parser, using the wiki.grm grammar from Cocoon's samples. This allows us to convert Wiki syntax into Forrest's document-v12 format, which can then be rendered to HTML or PDF. Many thanks to Stephan Michels for making this possible!
Navigation
A directory, src/documentation/content/xdocs/wiki/, contains a selection of Wiki files from the Cocoon Wiki:
[wikirenderer ~]$ ls src/documentation/content/xdocs/wiki ApacheModProxy.cwiki DistributingCocoonApplications.cwiki RhinoWithContinuations.cwiki AuthWithTomcat.cwiki EXistInCocoon.cwiki SeparationOfLogicAndContent.cwiki BeginnerInstallTomcatUnix.cwiki ForrestProposal.cwiki TextFormattingRules.cwiki BeginnerInstallTomcatWindows.cwiki index.cwiki WoodySample.cwiki BeginnerSimpleWebappOrganisation.cwiki Jars2exclude.cwiki XMLFormXindice.cwiki BlocksDefinition.cwiki JBossDeployment.cwiki XMLFormXindiceOldVersion.cwiki book.xml ModularDatabaseActions.cwiki XSPSyntax.cwiki CommandLine.cwiki OlderNews.cwiki ConfiguringTheLogs.cwiki OpenOfficeGeneration.cwiki
A menu for this directory is auto-generated by overriding navigation.xmap, and defining this matcher:
<map:match pattern="wiki/book.xml">
<map:generate type="directory" src="content/xdocs/wiki">
<map:parameter name="dateFormat" value="yyyy-MM-dd hh:mm" />
</map:generate>
<map:transform src="resources/stylesheets/directory2book.xsl" />
<map:serialize type="xml"/>
</map:match>
The 'directory' generator generates an XML directory listing of the wiki/ directory. The directory2book.xsl stylesheet transforms this into Forrest's book.xml menu format. The result is that wiki/*.html URLs all have a menu with links to other Wiki pages.
Aggregation
Having a single page of all Wiki content is useful for printing or searching. This is achieved by overriding the forrest.xmap sitemap, which defines the XML for raw pages, and defining a wiki/combined.xml page source:
<map:match pattern="wiki/combined.xml">
<map:generate type="directory" src="content/xdocs/wiki">
<map:parameter name="dateFormat" value="yyyy-MM-dd hh:mm" />
</map:generate>
<map:transform src="resources/stylesheets/directory2cinclude.xsl" />
<map:transform type="cinclude"/>
<map:transform src="resources/stylesheets/docs2document.xsl" />
<map:transform type="idgen"/>
<map:serialize type="xml"/>
</map:match>
Once again, we use the DirectoryGenerator to create an XML listing of all files in the wiki/ directory. This is then transformed by directory2cinclude.xsl into food for the CIncludeTransformer, of the form:
<content>
<cinclude:include src="cocoon:/wiki/ApacheModProxy.xml"/>
<cinclude:include src="cocoon:/wiki/AuthWithTomcat.xml"/>
<cinclude:include src="cocoon:/wiki/BeginnerInstallTomcatUnix.xml"/>
<cinclude:include src="cocoon:/wiki/BeginnerInstallTomcatWindows.xml"/>
<cinclude:include src="cocoon:/wiki/BeginnerSimpleWebappOrganisation.xml"/>
<cinclude:include src="cocoon:/wiki/BlocksDefinition.xml"/>
...
</content>
The CIncludeTransformer then pulls in the content from each referenced pipeline to form one large XML file of all Wiki content.
The docs2document.xsl stylesheet then converts <document> tags to <section> tags, so the result is valid doc-v12 XML. The idgen transformer derives a #fragment-identifier for each section. These will be used in the mini-TOC at the top of the page. Title collisions are resolved to render each #identifier unique within the aggregated page.
Once the combined.xml matcher is defined, nothing else is required; the combined.html and combined.pdf pages can now be rendered.
