The repo provides a simple API where the user selects the lexicons to be loaded and then is given a function that will do the glossing and return the first gloss found as well as a code showing which lexicon returned the result if no gloss is found a default result can be supplied. There are two version of this. One using an Either Monad and one that simply returns the glosses.
The repo provides 5 different glossing options:
- Autenrieth's homeric lexicon
- LSJ
- Abbott-Smith NT Greek lexicon
- Short Defs from the Perseus Project
- Greek to Greek deffinitions from Rouse's A Greek Boy at Home.
- Dodson Greek Lexicon
These options are stored in a named tuple named Lexica:
Lexica.AUTENREITH, Lexica.LSJ, Lexica.ABBOTT, Lexica.SHORT, Lexica.ROUSE, Lexica.DODSON
To create a glossing function use either load_dicts_simple or load_dicts (the monadic version) and supply it a list of the lexicons to check. The order of the lexicons in the list determines the order in which they are checked.
The return values is a tuple of the shape (<deffinition found>, <lexicon>).
By default the lexicon value is a string, but you can also supply map_names=False to the load_dicts function and it will return the Lexica tuple value representing the lexicon in which the gloss/deff was found.
The following snippet creates a glossing function that tried Autenreith's lexicon first and then proceeds to check LSJ and the Perseus Short Defs. If it can't find anything, it returns a default value of "N/A".
glosser = load_dicts_simple([Lexica.AUTENREITH, Lexica.LSJ, Lexica.SHORT], default='N/A')
# print gloss/deff and lexicon code
result = glosser('καί')
print(f"{result[0]} (found in {result[1]})")
The following snippet does the same thing but uses the monadic result:
glosser = load_dicts([Lexica.AUTENREITH, Lexica.LSJ, Lexica.SHORT], default='N/A')
res = glosser('NOT A GREEK WORD')
assert res == Left('N/A')
res2 = glosser('καί')
assert typeof(res2) == Right
# print deff
print(res2.value[0])
# print lexicon used
print(res2.value[1])
Both the aut and lsj repos (included here) are from the Alpheios Project and are based on data taken from the Perseus Project (http://www.perseus.tufts.edu/).
The National Endowment for the Humanities provided support for entering these texts.
The Perseus data is released under a CC BY-SA 4.0 license
The gloss data in ShortdefsforOKLemma_perseus.txt is taken from https://github.com/helmadik/shortdefs which contains data taken from Perseus (perseus.uchicago.edu/greek.html) and Logeion (https://logeion.uchicago.edu/lexidium).
The Greek to Greek definitions found in the Rouse vocabulary data are from Rouse's A Greek Boy At Home's Vocabulary which can be found here
The Dodson lexicon is taken from here. It is in the public domain.
In order to comply with licensing of the source data, this work is also released under a CC BY-SA 4.0 license