def format_gf_lexicon_entry(word, rektion):
verbtype, lin = get_lin(word, rektion)
funname = gfutils.get_funname(word, verbtype)
print('{0}\t{1} = mk{2} {3} ;'.format(word, funname, verbtype, lin))
def format_gf_lexicon_entry(word, rektion):
verbtype,lin = get_lin(word, rektion)
funname = gfutils.get_funname(word, verbtype)
print('{0}\t{1} = mk{2} {3} ;'.format(word, funname, verbtype, lin))
except KeyError:
pass
return text # leave as is
return re.sub("&#?\w+;", fixup, text)
def get_forms(word):
if word in lemma_to_forms:
return lemma_to_forms[word]
return '"' + word + '"'
def is_illegal(word):
return (word in illegal) or not (word in lemma_to_forms)
args = gfutils.get_args()
lemma_to_forms = gfutils.get_lemma_to_forms(args.forms)
line_number = 0
for line in sys.stdin:
line_number += 1
line = line.strip()
fields = line.split(':')
word = unescape(fields[0])
parts = word.split(' ')
verb = parts[-1]
if re.match('^[a-zõäöüšž]+$', verb) and not is_illegal(verb):
funname = gfutils.get_funname(word, 'V')
print('{0}\t{1} = mkV "{2}" (mkV {3}) ;'.format(word, funname, ' '.join(parts[0:-1]), get_forms(verb)))
else:
print("Warning: line " + str(line_number) + ": ignoring: " + word, file=sys.stderr)
args = gfutils.get_args()
lemma_to_forms = gfutils.get_lemma_to_forms(args.forms)
# The input line can have the Filosoft tag after the '//' sign (which we ignore).
# The word itself can contain spaces to denote compound word segment borders.
line_number = 0
for line in sys.stdin:
line_number += 1
line = line.strip()
word = re.sub(' //.*', '', line)
word = word.strip()
parts = word.split(' ')
word = re.sub(' ', '', word)
if re.match('^[a-zõäöüšž]+$', word) and not is_illegal(word):
funname = gfutils.get_funname(word, args.pos)
found_word = False
entry = ''
for i in range(0, len(parts)):
lemma = ''.join(parts[i:])
if lemma in lemma_to_forms:
if i == 0:
entry = 'mkN %s' % (lemma_to_forms[lemma])
else:
# The word is a compound word
prefix = ''.join(parts[:i])
# We do not accept single character prefixes.
# TODO: These might make sense, depending on how they are coded (ekiri vs e-kiri)
# TODO: What about 'apriori'?
if len(prefix) == 1:
continue
def format_gf_lexicon_entry(word, rektion):
verbtype, lin = get_lin(word, rektion)
funname = gfutils.get_funname(word, verbtype)
print "%s\t%s = mk%s %s ;" % (word, funname, verbtype, lin)
#!/usr/bin/env python
# -*- coding: utf-8 -*-
# @author Kaarel Kaljurand
# @version 2013-10-01
import sys
import re
import gfutils
for line in sys.stdin:
line = line.strip()
word = re.sub(' //.*', '', line)
word = word.strip()
funname = gfutils.get_funname(word, 'Adv')
print '%s\t%s = mkAdv "%s" ;' % (word, funname, word)
