# header file for published German lexicon

# Each line of the lexicon contains the following seven fields
# in the following order.  The first field contains a name, the second
# contains a regular expression that should match every instance of that
# field.  Fields are tab separated, here as well as in the lexicon.

# For full documentation of the lexicon, please read the file ge_lex.doc!

# Usage of this file:
#  simple syntax check:
#	parse-lex -s ge_lex.hdr ge_lex.01
#  output full "parse" of fields:
#	parse-lex ge_lex.hdr ge_lex.01
# The "parse-lex" tool is available via FTP from ftp.ldc.upenn.edu

HEADWORD	[-A-ZÄÖÜßa-zäöüé_]+

# morphological and stem information
# (Note that multiple entries in Celex create some complexity here.)
let $morphpiece=([-A-ZÄÖÜßa-zäöüé]+(\+[1-3]?[A-Z][a-zA-Z_]*)+)
MORPH	$morphpiece(//$morphpiece)*(\|\|$morphpiece(//$morphpiece)*)*

# pronunciation
let $pronpiece=([aeiouA@EIOUWwYy&pbtdkgGmnlrfvszSZjxhV\$]|([wa\$O]~))+
PRON	$pronpiece(\|\|$pronpiece)*
#PRON	[a-zA-Z@&\$~]+(\|\|[a-zA-Z@&\$~]+)*

# stress pattern(s): 1=stressed, 0=unstressed
STRESS	0*1[01]*(\|\|0*1[01]*)*

# was this word in the CELEX lexicon?
CELEX	[01]

# number of occurrences in training transcripts
TRAIN	[0-9]+

# number of occurrences in devtest transcripts
DEV	[0-9]+