pyglottography

Programmatic curation of Glottography datasets

Installation

Install via pip from PyPI:

pip install pyglottography

Note

We use GDAL's ogr2ogr command to convert between GeoJSON and GeoPackage formats. Thus, some functionality of pyglottography requires a working GDAL installation.

Curating Glottography datasets with pyglottography

Bootstrapping a new dataset

pyglottography provides a cldfbench project template, which can be used with the cldfbench new command:

cldfbench new --template glottography

Providing the raw data

The cldfbench workflow uses data in a project's raw directory - enriched with information from etc - to create a CLDF dataset in the cldf directory. By default, pyglottography expects input data as follows:

• Geo-data, i.e. shapes for languoid areas, is expected in a GeoJSON file raw/dataset.geojson. Each feature in thie GeoJSON file should have a unique value for the id property.
• Metadata about the shapes is expected in a CSV file etc/features.csv. This file must have an id column with values corresponding to the feature ids in the geo-data.

While metadata could be read entirely from the properties object of features in the GeoJSON file, pyglottography looks up the metadata in a different file to allow for more transparent curation. Since the Glottolog language catalog is released in a new version about twice a year, it is necessary to be able to recreate a Glottography dataset with updated Glottocodes. With the raw data setup as implemented in pyglottography, this only requires changes in etc/features.csv, which can easily be tracked with versioning software such as git.

Running the CLDF creation

cldfbench makecldf cldfbench_<dsid>.py