Add example pipelines

[johann-petrak]

Initially:

• Train Topic Model: ideally this would also make use of the example pipelines from stringannotation for token filtering by stopwords and corpusstats for token filtering by tfidf, but how?
• Apply Topic model