Tabulae A build system for parsing Latin


There are three main phases to building a corpus-specific parser and working with its output.

  1. you write (or modify existing) plain-text tables defining rules and vocabulary for your parser.
  2. tabulae’s build packages read your tables, then write and compile SFST code.
  3. tabulae’s analysis packages can read the output of the SFST parser, and give you structured objects.

This documentation accordingly falls into three parts;

  1. Building a corpus-specific dataset
  2. Compiling and using a parser
  3. Analyzing the output of the parser