mudflat

Minimal universal dependency friendly little automated tagger

A coding system supporting PLOVER (of course): https://github.com/openeventdata/PLOVER; http://ploverdata.org

mudflat is basically a testbed for the FJOLTYNG coder, which will be a full-featured PLOVER coder based on universal dependency parses. FJOLTYNG will use a new set of JSON formats for both input and dictionaries, which is to say it will not be compatible with the input files of either mudflat or the TABARI/PETRARCH family of coders.

Files

All programs are Python 3.7 and open source under the MIT License.

mudflat.py

Main driver program

globals.py

Globals, adapted from PETRARCH-2 (https://github.com/openeventdata/petrarch2)

reader.py

Configuration and dictionary input adapted from PETRARCH-2; CoNNL input

coder.py

Coding routines

utilities.py

Small utility routines, with parts adapted from PETRARCH-2

CAMEO2PLOVER.txt

PLOVER event-mode-context equivalents to CAMEO codes. Definitely a first approximation.

extract_UD_parse.py

Utility program for experimenting with CoNLL-U format routines

mudflat_testdata_Mk1.txt

Records used to achieve "first light": these are actual reports, with the CoNLL-U parse generated by Stanford CoreNLP; actor and verb dictionaries are just the usual ones for PETRARCH-2

Documentation.md

Draft documentation, which is more a storage point for future elements of the documentation than a finished product