The CantusCorpus is a corpus of plainchant intended for
computational research. It is essentially a dump of
the Cantus database. The database was scraped
using its API, and converted to easy-to use CSV files. For many chants,
transcriptions in the Volpiano format are included. These can be loaded into
music21, a Python toolkit for computational
musicology, using the library chant21
.
Note: Even the latest version of the corpus will generally be out-dated, as the Cantus database is updated continuously. CantusCorpus is intended only for computational studies, where this is less of a problem. If you require up-to-date information, please do not use this corpus, but use Cantus directly.
> Download the latest release of the CantusCorpus
> Check out chant21 and the GregoBaseCorpus, a related plainchant corpus.
If you use this corpus in your research, please cite the Cantus Database as suggested on its website:
Cantus: A Database for Latin Ecclesiastical Chant -- Inventories of Chant Sources. Directed by Debra Lacoste (2011-), Terence Bailey (1997-2010), and Ruth Steiner (1987-1996). Web developer, Jan Koláček (2011-). Available from http://cantus.uwaterloo.ca/. [date accessed].
Further please cite the paper describing the CantusCorpus:
Forthcoming...
As Cantus is being updated continuously, we plan to occasionaly release new versions of the CantusCorpus as well. All of these will be versioned, can be downloaded from GitHub.
The CantusCorpus (the collection of .csv
files) is released under a
CC BY-NC-SA 4.0 license,
just like the Cantus Database itself. The Python code used to generate the
corpus (the code in the src/
directory) is released under an MIT license.
The CantusCorpus is created automatically after scraping the Cantus API. If you just want to use the corpus, you don't have to regenerate it yourself: simply download the latest release and you're good to go. But if you want to regenerate the corpus yourself, you can of course do so: read on....