This repo contains various datasets in Georgian for NLP or other purposes. These are entire text of "The Knight with the Panther skin" vefxistyaosani.txt, Georgian aphorisms aforizmebi.txt, first and last names of Georgian poets and writers poetswriters.txt, baby names in Georgian names.csv (© kids.ge), and full Georgian Alphabet anbani.csv with corresponding descriptions of the letters as it appears in Unicode.

Some of these datasets were fed to Neural Networks (char-rnn by Andrej Karpathy) to generate fake data, such as fake-aforizmebi.txt, fake-names.txt trained on Georgian (origin) subset, fake-poetswriters.txt.

The literature manifest distinguishes complete texts from entries that still need a trustworthy transcription or title correction. Wikisource-derived transcriptions retain revision-level provenance and are distributed under CC BY-SA 4.0; the underlying works are public domain in Georgia. Run node scripts/import-literature.mjs to reproducibly refresh the pinned texts.

Here are some of the resources you might like.

Fake Georgian text and names generation is supported by anbani.js - a multifunctional Javascript library for working with Georgian Alphabet. Read more about the package here [anbani / anbani.js]

npm install anbani

var anbani = require('anbani')

anbani.core.convert("ანბანი", "მხედრული", "ასომთავრული")
// 'ႠႬႡႠႬႨ'

anbani.lorem.names(3)
// ['დამერ გაშვითელი', 'სიბო ყორთელია', 'გიმოლ ვაწოშვილი']

anbani.lorem.sentences(10)
// 'მოეხვიდეს სიტირენ გიშიხარნი. წეითო გამიზრიან, ჰქონთავისთან გემრუფენ, უკრთებოდემნი მესმანცა მყივნე.'

For other awesome Georgian datasets, visit [bumbeishvili / awesome-georgian-datasets]

Datasets are available freely for non-commercial purposes only. For commercial purposes, contact the corresponding source.