The KrdWrd Project ran from 2008 to 2011. The mission statement was
Provide tools and infrastructure for acquisition, visual annotation, merging and storage of web pages as parts of bigger corpora.
Develop a classification engine that learns to automatically annotate pages, and provide visual tools for inspection of results.
Basically, it was an infrastructure for research into web page cleaning. A good overview can be found in the paper and an extensive description in the master's thesis (both, see further down).
The CANOLA Corpus is a visually annotated English web corpus for training classification engines to remove boiler plate on unseen web pages. It was harvested, annotated and evaluated by the tools and infrastructure of the KrdWrd Project.
The corpus consists of 216 files (Web pages) - 208 of which constitute the main corpus.
See https://github.com/krdwrd/doc_CANOLA/releases/download/v1.1/canola.pdf for more information.