Integrate RDR tagger

[zesch]

We should check whether we want to integrate this tagger
http://rdrpostagger.sourceforge.net

[reckart]

Just added that to my queue today :) The questions would IMHO rather be:

• why not add it?
• who volunteers?

[zesch]

@Horsmann Interested in adding a new tagger? :)

[reckart]

@zesch for curiosity: what makes this tagger particularly interesting to you?

[zesch]

Morphological tags for a lot of languages sound promising, but unsure what kind of performance is to be expected.

[reckart]

There seem to be a few quite exotic languages indeed :)

[daxenberger]

FYI: one of the developers has been doing an internship at UKP Lab a while ago ;-)

[reckart]

Really? I did check the publication history and it appears to be another person by the same name to me.

[daxenberger]

Really?

yes. See the bio.

[Horsmann]

@reckart how would you integrate a java application that is only released to a software repository as code but is not available as maven release. In this case, the developer uses a versioning schema but this is not reflected in the git-log so far I see. The latest version the developer distinguishes is from May, the last code changes are from September. How to integrate this properly into Core?

[reckart]

A third-party deploy to Maven central would be necessary.

banana-split (a compound splitter) was at the time also only released as a ZIP file on a website (I don't know it's status now). What I did was:

• ask the developers if they agree that I do a third-party upload to Maven Central (and propose a groupId/artifactId) or if they want to do it themselves: they said I could do it
• download the ZIP
• create a local Maven project using the source code from the ZIP
• asked Sonatype what to put as the SCM URL into the pom - we agreed to simply put the URL of the ZIP file
• build the JAR and source JAR from the local project
• uploaded them to Maven Central as a third-party upload

Then I used the resulting artifact in DKPro Core.

[reckart]

I have opened an upstream issue asking for the Java version of the tagger to be released to Maven Central: datquocnguyen/RDRPOSTagger#19