… instead

The HomepageExtractor was not able to correctly detect references to official webpages
when conveyed through templates placed in the External links section.
Only matches to Template:Official were considered and the URL extraction was failing
when the template property contained an ExternalLinkNode.

With this change:
- A map is provided to store info about official website templates, for different langs
- Template redirects are checked when looking for official websites
- URL is extracted also when the template property is an ExternalLinkNode

I was able to extract 550k homepages from enwiki-20130604 compared to about 505k items
without this patch (about 10% more).

Fix HomepageExtractor issues

When extracting quads from the extraction server, we should
deduplicate quads so that we get a similar result to that
produced by the extraction framework

Issue dbpedia#144: Deduplicate quads in the extraction server

Various enhancements for Live

…perty

- Decide whether to generate statistics by reading a JVM system property.
By doing that it is possible to create stats without having to edit the code
and recompile the framework.
- Add a Scala launcher in pom to start stats creation by setting the
appropriate JVM system property

Issue dbpedia#149: Read stats generation property from JVM system property

Set default values in case any of the HomepageExtractor required config does not exist
for a specific language (e.g. there is no official website template in itwiki, but
there is an official keyword used in external links section).

Not using English as default because it looks like a strong assumption which is likely
to be either wrong or disruptive.

HomepageExtractor config: return default empty values for undefined settings in a language

…ersand

Issue dbpedia#151: URL encode wikititle in sample page extraction