The zdatrtok module is based on a top down parser whose purpose is to preprocess a given DATR theory by building look-up tables and representing DATR equations in a decision tree structure.
As noted in connection with the zdatr command-line interface, the resulting token file contains all necessary information for the inference engine and therefore can be used for repeated query calls without the original DATR source.
Typing zdatrtok on the command line should produce the following help page:
This is ZDATR tokenizer, a part of ZDATR version 2.0,
(C) Sep 12 1998 Universitaet Bielefeld
Usage:
zdatrtok [-o OUTFILE] [-L LOGFILE] [ -Q QRYFILE ][ -D DECFILE ] [-s] DTRFILE
DTRFILE is the name of input DATR-theory
-n Do not assume '#load std.' by default
-o OUTFILE Name of the output token file,
if the outfile name is '-' the output will be
written to stdout instead of a file (to pipe it to zdatrinf).
-D DECFILE Name of the declaration file where to write #show and #hide,
default is DTRFILE.dec
-Q QRYFILE Name of the query file where to write #query,
default is DTRFILE.dec
-L LOGFILE Name of the output log-file (default DTRFILE.log)
-u Ignore 'undefined symbol' error
-s Silent mode