Preparing a Corpus for the Sketch Engine: Overview

To prepare a corpus for the Sketch Engine, we must

• Prepare the data, including both
  • SkE/PrepareText [the text]
  • SkE/PrepareHeaders [the header information] (if any)
• Prepare a SkE/CorpusConfig [corpus configuration file]
• Run the encodevert program.

(Here we assume a running SketchEngine installation.)

This will give us a corpus which can be queried to give a range of concordances and lists. If, in addition, word sketches are required we must also

• Prepare a grammatical relations definitions (gramrels) file: see SkE/CorpusQuerying
• Run the mkws.sh script.

This will also prepare the thesaurus which requires no additional inputs. It takes the word sketch database as input.