Open funderburkjim opened 6 years ago
In addition to providing reliable inflections, the code should be as straightforward as possible.
The code is derived from a previous project: pysanskritv1. and this previous python code was a translation of code written in Elisp (Emacs lisp). One flaw of this previous code is that it is often quite convoluted, and thus difficult to understand and improve. It is hoped that the current code will remedy this flaw to a great extent.
How are the generated forms proofed?
How proofed ?
Proofing the generated forms is certainly an important topic. Some of the possibilities:
https://gitlab.inria.fr/huet/Heritage_Resources/blob/master/XML/SL/SL_nouns.xml houses the inflected forms as per Gerard Huet's site. Also see http://sanskrit.inria.fr/xml.html and https://gitlab.inria.fr/huet/Heritage_Resources/tree/master/XML/SL for details of how the XML is organized.
@funderburkjim and @gasyoun We can parse the XML by Gerard and test the inflected forms in our database and his database.
Jim, what's your take? The pirate is back in our boat.
Here is a brief description of the directory structure of this MWinflect repository, based on current thinking.