def load_tikhonov_dict(path):
result = []
with open(path, 'r') as p:
for line in p:
word, parse = line.strip().split('\t')
result.append(parse_word(word, parse))
return result
def gen_multi_lexeme_parse(lemma, lemma_parse, multform, multform_tail):
parsed_lemma = parse_word(lemma, lemma_parse)
multform_parse = parsed_lemma.copy()
if multform == "люди":
return "люд:ROOT/и:END"
if len(lemma) == len(multform):
for i in range(len(multform)):
multform_parse.set_letter(i, multform[i])
return str(multform_parse)
tail_parts, tail_len = parse_tail_parts(multform_tail)
if len(multform) > len(lemma):
if len(tail_parts
) == 1 and multform[:-len(tail_parts[0].text)] == lemma:
new = True
for letter in tail_parts[0].text:
multform_parse.append_letter(letter, MorphemeLabel.END, new)
new = False
return str(multform_parse)
new = True
for i in range(len(multform) - tail_len):
if i < len(multform_parse):
if multform_parse.get_label(i) != 'END':
multform_parse.set_letter(i, multform[i])
else:
multform_parse.set_letter(i, multform[i])
multform_parse.set_label(i, MorphemeLabel.SUFF)
new = False
else:
multform_parse.append_letter(multform[i], MorphemeLabel.SUFF,
new)
new = False
while len(multform_parse) > len(multform) - tail_len:
multform_parse.pop_letter()
current_tail_pos = 0
while current_tail_pos < len(tail_parts) - 1:
if tail_parts[current_tail_pos].text == 'ь':
new = False
label = multform_parse.get_last_label_value()
else:
label = MorphemeLabel.SUFF
new = tail_parts[current_tail_pos].text[0] != 'ь'
for letter in tail_parts[current_tail_pos].text:
multform_parse.append_letter(letter, label, new)
new = False
current_tail_pos += 1
if current_tail_pos < len(tail_parts):
label = MorphemeLabel.END
if not tail_parts[current_tail_pos].new:
label = MorphemeLabel.SUFF
new = True
for letter in tail_parts[current_tail_pos].text:
multform_parse.append_letter(letter, label, new)
new = False
return str(multform_parse)
return str(lemma_parse)
def gen_parse2(common_part, tail, lemma, lemma_parse, is_gerund=False):
parsed_lemma = parse_word(lemma, lemma_parse)
common_parse = parsed_lemma.copy()
is_polden = lemma == "полдень"
has_postfix = False
if parsed_lemma.morphemes[-1].label == MorphemeLabel.POSTFIX:
has_postfix = True
tail_parts, tail_len = parse_tail_parts(tail)
postfix_part = None
if has_postfix:
postfix_part = tail_parts[-1]
tail_parts.pop()
while len(common_parse) > len(common_part):
common_parse.pop_letter()
if len(tail_parts) > 0:
for i in range(len(common_parse)):
if common_parse.get_label(i) == 'END':
common_parse.set_label(i, MorphemeLabel.SUFF)
first_is_prefix = False
if len(common_parse) == 0 or common_parse.get_last_label_value(
) == MorphemeLabel.PREF:
first_is_prefix = True
if not tail_parts:
return str(common_parse)
current_pos = len(common_parse)
if current_pos >= len(lemma):
if len(tail_parts) > 1:
raise Exception("Unknown case")
new = True
for letter in tail_parts[0].text:
common_parse.append_letter(letter, MorphemeLabel.END, new)
new = False
if has_postfix:
new = True
for letter in postfix_part.text:
common_parse.append_letter(letter, MorphemeLabel.POSTFIX, new)
new = False
return str(common_parse)
if parsed_lemma.get_label(current_pos) != 'ROOT':
complete_root = True
elif lemma[current_pos] == 'ь' and (
len(parsed_lemma) == current_pos + 1
or parsed_lemma.get_label(current_pos + 1) != 'ROOT'):
if tail_parts[0].text[0] != 'ь' or len(tail_parts[0].text[0]) > 1:
complete_root = True
else:
complete_root = False
else:
complete_root = False
current_tail_pos = 0
if first_is_prefix:
if not len(common_parse) == 0 and not tail_parts[current_tail_pos].new:
for letter in tail_parts[current_tail_pos].text:
common_parse.append_letter(letter, MorphemeLabel.PREF, False)
current_tail_pos += 1
if current_tail_pos < len(tail_parts):
new = True
for letter in tail_parts[current_tail_pos].text:
common_parse.append_letter(letter, MorphemeLabel.ROOT, new)
new = False
current_tail_pos += 1
if current_tail_pos < len(
tail_parts
) and not complete_root and not tail_parts[current_tail_pos].new:
for letter in tail_parts[current_tail_pos].text:
common_parse.append_letter(letter, MorphemeLabel.ROOT, False)
current_tail_pos += 1
while current_tail_pos < len(tail_parts) - 1:
new = tail_parts[current_tail_pos].text[0] != 'ь'
label = MorphemeLabel.SUFF
if is_polden:
label = MorphemeLabel.ROOT
for letter in tail_parts[current_tail_pos].text:
common_parse.append_letter(letter, label, new)
new = False
current_tail_pos += 1
if current_tail_pos < len(tail_parts):
label = MorphemeLabel.END
if not tail_parts[current_tail_pos].new and not tail_parts[
current_tail_pos].text == 'ь':
label = MorphemeLabel.SUFF
if tail_parts[current_tail_pos].text in ('л', 'вши', 'ши', 'в', 'ну',
'ек', 'ин', 'ен',
'н') or is_gerund:
label = MorphemeLabel.SUFF
new = True
for letter in tail_parts[current_tail_pos].text:
common_parse.append_letter(letter, label, new)
new = False
if has_postfix:
new = False
for letter in postfix_part.text:
common_parse.append_letter(letter, MorphemeLabel.POSTFIX, new)
new = False
if len(common_parse) != len(common_part) + tail_len:
raise Exception("Incorrect parse generated")
return str(common_parse)