loanpy 1.0.1

Python framework for detecting old loanwords

Installation

$ python -m pip install loanpy

Documentation

Citation

Description

loanpy is a toolkit for historical linguists. It extracts sound changes from an etymological dictionary. It reconstructs hypothetical roots of modern L1 words. It creates hypothetical adaptions of L2 words into proto-L1. It searches for potential loanwords by first finding phonetic matches and then calculating their semantic similarity.

Data Sources

• dfhun_zaicz_backup.csv: data frame based on the Hungarian etymological dictionary (Zaicz 2006)

• dfgot_wikiling_backup.csv: data frame based on Wikiling

• dfgot_wiktionary_backup.csv: data frame based on Wiktionary

• dfuralonet.csv: data frame based on Uralonet

• substi.csv: Sound substitutions based on Information-theoretic causal inference of lexical flow (Dellert 2017)

• wordvectornames.xlsx: Names of pretrained word vector models from gensim-data

Dependencies

• gensim

• ipatok

• python-Levenshtein

• pandas

• see also requirements.txt

License

Academic Free License (AFL)

Git