Corpus Data

CHAT Format

For a corpus dataset to be useful for modeling work beyond search queries, its source data has to be available in a machine-readable format. For Cantonese, several corpora that meet this criterion are those from CHILDES and TalkBank, thanks to research on Cantonese language acquisition in recent years. More generally, given the nature of Cantonese, many of its corpora are transcribed data from naturalistic speech. For these reasons, PyCantonese adopts the CHAT corpus format from CHILDES and TalkBank. CHAT is widely used, well-documented, and rich for linguistic annotations. PyCantonese uses the Python library PyLangAcq to parse CHAT data files. For a primer on CHAT, please see here.

Built-in Data

Currently, PyCantonese comes with one built-in corpus, the Hong Kong Cantonese Corpus (HKCanCor; license: CC BY), via the function hkcancor():

>>> import pycantonese
>>> hkcancor = pycantonese.hkcancor()
>>> hkcancor.n_files()  # number of data files
58
>>> len(hkcancor.words()) # number of words as segmented from all the utterances
153654

HKCanCor is word-segmented and annotated for both Jyutping romanization and part-of-speech tags.

The original HKCanCor source files are in an XML format. They have been converted to CHAT for incorporation into PyCantonese. On the format conversion, please consult this readme.

CHILDES and TalkBank Data

For corpora other than HKCanCor, PyCantonese provides the function read_chat() to read in Cantonese data in the CHAT format.

read_chat() is designed to be able to read in CHAT data from a URL that points to a ZIP file containing .cha CHAT files. The availability of Cantonese CHAT data from CHILDES and TalkBank means that it is possible to conveniently obtain and work with such data right from your own Python code, without having to manually download or unzip anything.

Note

All publicly available CHILDES and TalkBank datasets are associated with the CC BY-NC-SA 3.0 license.

As of May 2022, the following Cantonese-related datasets are available from CHILDES and TalkBank (in alphabetical order):

  • Child Heritage Chinese Corpus

    >>> url = "https://childes.talkbank.org/data/Biling/CHCC.zip"
    >>> corpus = pycantonese.read_chat(url)
    >>> corpus.n_files()
    190
    >>> len(corpus.words())
    533877
    
  • Guthrie Bilingual Corpus

    >>> url = "https://childes.talkbank.org/data/Biling/Guthrie.zip"
    >>> corpus = pycantonese.read_chat(url)
    >>> corpus.n_files()
    36
    >>> len(corpus.words())
    84233
    
  • HKU-70 Corpus

    >>> url = "https://childes.talkbank.org/data/Chinese/Cantonese/HKU.zip"
    >>> corpus = pycantonese.read_chat(url)
    >>> corpus.n_files()
    70
    >>> len(corpus.words())
    178270
    
  • Lee-Wong-Leung Corpus

    >>> url = "https://childes.talkbank.org/data/Chinese/Cantonese/LeeWongLeung.zip"
    >>> corpus = pycantonese.read_chat(url)
    >>> corpus.n_files()
    161
    >>> len(corpus.words())
    1177307
    
  • Leo Corpus

    >>> url = "https://childes.talkbank.org/data/Biling/Leo.zip"
    >>> corpus = pycantonese.read_chat(url)
    >>> corpus.n_files()
    54
    >>> len(corpus.words())
    223415
    
  • Paidologos Corpus: Cantonese

    >>> url = "https://phonbank.talkbank.org/data/Chinese/Cantonese/PaidoCantonese.zip"
    >>> corpus = pycantonese.read_chat(url)
    >>> corpus.n_files()
    160
    >>> len(corpus.words())
    16730
    
  • Yip-Matthews Bilingual Corpus

    >>> url = "https://childes.talkbank.org/data/Biling/YipMatthews.zip"
    >>> corpus = pycantonese.read_chat(url)
    >>> corpus.n_files()
    501
    >>> len(corpus.words())
    1949480
    

Custom Data

If you have a Cantonese corpus in the CHAT format in your local drive and would like to use PyCantonese to handle it, read_chat() takes a path that can be a ZIP file, a local directory, or a single CHAT file.

If more fine-grained control is needed when reading data, please check out CHATReader, particularly the following classmethods:

Since PyCantonese uses PyLangAcq for CHAT data reading and parsing under the hood, PyCantonese’s read_chat() and CHATReader function the same way as their counterparts in PyLangAcq. For more on reading CHAT data in general, please see PyLangAcq’s documentation.