Palladian Nodes for KNIME Workbench version 1.7.0.v201807041014 by palladian.ws; Philipp Katz, Klemens Muthmann, David Urbansky.
This node provides different content extraction algorithms, which allow extracting textual content from web pages, discarding irrelevant elements like navigation, ads, footers, headers, etc. The extractors are generally optimized for typical news article and blog post web pages. Currently, three algorithms are provided:
"Readability [...] takes a crack at wiping out all that junk so you can have a more enjoyable reading experience. [...] its
success rate is pretty respectable (we'd guess over 90% of web sites are handled properly)". Readability operates on the
document's DOM tree. Basically, it assigns all elements a score for their contents. Metrics for the scoring are length of their
text content, number of commas and link density. Also,
id names are taken into consideration; for
example, elements with class name
sidebar contain unlikely actual content in contrast to elements with class
The Palladian content extractor extracts clean sentences from (English) texts. That is, short phrases are not included in the output. Consider Readability for general content. The main difference is that this class also finds sentences in comment sections of web pages.
To use this node in KNIME, install Palladian Nodes for KNIME Workbench from the following update site: