Home

Corpus Kaamelott is a collection of screenplays from the French TV show Kaamelott. They originally result from the scraping of the Hypnoweb website. All screenplays have been tokenized, lemmatized and annotated with automatic methods:

• number of screenplays: 400
• number of sentences: 39.217
• number of tokens: 392.142
• multiple formats (v0.6: text, tagged, XML-TEI)
• character set: UTF-8

The corpus is distributed along with many resources:

• a directory of actors;
• a directory of characters in Kaamelott;
• a list of all the episodes transcribed on Hypnoweb;
• a collection of metadata about the original screenplays scraped from Hypnoweb;
• a list of named entities;
• a lexicon of French slang used in the screenplays;
• a map between the POS-tags used in the corpus and several other tagsets.

There is also a custom Python class designed to work with NLTK: KaamelottCorpusReader.