You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
95 lines
2.6 KiB
95 lines
2.6 KiB
from collections import defaultdict
|
|
import logging
|
|
|
|
from luscenje_struktur.msd_translate import MSD_TRANSLATE
|
|
|
|
|
|
class WordCompressed:
|
|
def __init__(self, text, collocation, dependency_tree):
|
|
self.text = text
|
|
self.collocation = collocation
|
|
self.dependency_tree = dependency_tree
|
|
|
|
|
|
class WordMsdOnly:
|
|
def __init__(self, msd):
|
|
self.msd = msd
|
|
self.lemma = None
|
|
self.text = None
|
|
|
|
def most_frequent_text(self, _):
|
|
return None
|
|
|
|
|
|
class WordDummy:
|
|
def __init__(self, msd, lemma, text):
|
|
self.msd = msd
|
|
self.lemma = lemma
|
|
self.text = text
|
|
|
|
def most_frequent_text(self, word_renderer):
|
|
return word_renderer.render(self.lemma, self.msd)
|
|
|
|
|
|
class Word:
|
|
def __init__(self, lemma, msd, wid, text, do_msd_translate, fake_word=False, previous_punctuation=None):
|
|
self.lemma = lemma
|
|
self.msd = MSD_TRANSLATE[msd] if do_msd_translate else msd
|
|
self.id = wid
|
|
self.idi = None
|
|
self.text = text
|
|
self.glue = ''
|
|
self.previous_glue = '' if previous_punctuation is None else previous_punctuation
|
|
self.fake_word = fake_word
|
|
|
|
self.links = defaultdict(list)
|
|
|
|
last_num = self.id.split('.')[-1]
|
|
if last_num[0] not in '0123456789':
|
|
last_num = last_num[1:]
|
|
self.int_id = int(last_num)
|
|
|
|
assert None not in (self.id, self.lemma, self.msd)
|
|
|
|
@staticmethod
|
|
def from_xml(xml, do_msd_translate):
|
|
lemma = xml.get('lemma')
|
|
msd = Word.get_msd(xml)
|
|
wid = xml.get('id')
|
|
text = xml.text
|
|
return Word(lemma, msd, wid, text, do_msd_translate)
|
|
|
|
@staticmethod
|
|
def get_msd(comp):
|
|
d = dict(comp.items())
|
|
if 'ana' in d:
|
|
return d['ana'][4:]
|
|
elif 'msd' in d:
|
|
return d['msd']
|
|
else:
|
|
logging.error(d)
|
|
raise NotImplementedError("MSD?")
|
|
|
|
@staticmethod
|
|
def pc_word(pc, do_msd_translate):
|
|
pc.set('lemma', pc.text)
|
|
pc.set('msd', "N" if do_msd_translate else "U")
|
|
return Word.from_xml(pc, do_msd_translate)
|
|
|
|
@staticmethod
|
|
def fake_root_word(sentence_id):
|
|
wid = sentence_id
|
|
return Word('', '', wid, '', False, True)
|
|
|
|
def add_link(self, link, to):
|
|
self.links[link].append(to)
|
|
|
|
def get_links(self, link):
|
|
if link not in self.links and "|" in link:
|
|
for l in link.split('|'):
|
|
self.links[link].extend(self.links[l])
|
|
|
|
return self.links[link]
|
|
|
|
def most_frequent_text(self, word_renderer):
|
|
return word_renderer.render(self.lemma, self.msd) |