def quantify(analysis):
edge_marker = [('', '')]
pos_tags = edge_marker + analysis.pos_tags() + edge_marker
num_tags = len(pos_tags) - 2 # Remove 2 edge markers
count = 0
for i in range(len(pos_tags)):
if is_proper_noun(pos_tags[i]) and \
not is_proper_noun(pos_tags[i-1]) and \
not is_proper_noun(pos_tags[i+1]):
count += 1
return {"single_naming": float(count) / num_tags}
def quantify(analysis):
edge_marker = [('', '')]
pos_tags = edge_marker + analysis.pos_tags() + edge_marker
num_tags = len(pos_tags) - 2 # Remove 2 edge markers
count = 0
for i in range(len(pos_tags)):
if is_proper_noun(pos_tags[i]) and \
not is_proper_noun(pos_tags[i-1]) and \
not is_proper_noun(pos_tags[i+1]):
count += 1
return { "single_naming": float(count) / num_tags }
def quantify(analysis):
"""Quantify explicit naming."""
proper_nouns = sum(1 for x in analysis.pos_tags() if is_proper_noun(x))
pronouns = sum(1 for x in analysis.tokens() if x in PRONOUNS_LIST)
return { "explicit_naming": 3.0 * (float(pronouns) / proper_nouns) }
def quantify(analysis):
"""Quantify mean multiple naming."""
if analysis.lang == 'en':
from translationese.utils import is_proper_noun
elif analysis.lang == 'zh':
from translationese.utils import is_proper_noun_zh as is_proper_noun
else:
print('language "{}" not implemented yet for single naming'.format(analysis.lang))
exit()
num_proper_noun_runs = 0
num_proper_noun_tokens = 0
currently_in_run = False
# pos_tags = nltk.pos_tag(analysis.case_tokens())
# pos_tags = [word_tag_pair[1] for word_tag_pair in analysis.pos_tags()]
pos_tags = analysis.pos_tags()
for token in pos_tags:
if is_proper_noun(token):
currently_in_run = True
num_proper_noun_tokens += 1
elif currently_in_run:
# Run just ended
currently_in_run = False
num_proper_noun_runs += 1
if num_proper_noun_runs == 0:
result = 0
else:
result = float(num_proper_noun_tokens) / num_proper_noun_runs
return { "mean_multiple_naming": result }
def quantify(analysis):
"""Quantify explicit naming."""
proper_nouns = sum(1 for x in analysis.pos_tags() if is_proper_noun(x))
pronouns = sum(1 for x in analysis.tokens() if x in PRONOUNS_LIST)
return {"explicit_naming": 3.0 * (float(pronouns) / proper_nouns)}
def quantify(analysis):
"""Quantify single naming."""
if analysis.lang == 'en':
from translationese.utils import is_proper_noun
elif analysis.lang == 'zh':
from translationese.utils import is_proper_noun_zh as is_proper_noun
else:
print('language "{}" not implemented yet for single naming'.format(
analysis.lang))
exit()
edge_marker = [('', '')]
pos_tags = edge_marker + analysis.pos_tags() + edge_marker
num_tags = len(pos_tags) - 2 # Remove 2 edge markers
count = 0
for i in range(len(pos_tags)):
if is_proper_noun(pos_tags[i]) and \
not is_proper_noun(pos_tags[i-1]) and \
not is_proper_noun(pos_tags[i+1]):
count += 1
return {"single_naming": float(count) / num_tags}
def quantify(analysis):
"""Quantify explicit naming."""
if analysis.lang == 'en':
from translationese.pronouns import PRONOUNS as PRONOUNS_LIST
from translationese.utils import is_proper_noun
elif analysis.lang == 'zh':
from translationese.pronouns import PRONOUNS_ZH as PRONOUNS_LIST
from translationese.utils import is_proper_noun_zh as is_proper_noun
else:
print('language "{}" not implemented yet for explicit naming'.format(
analysis.lang))
exit()
proper_nouns = sum(1 for x in analysis.pos_tags() if is_proper_noun(x))
pronouns = sum(1 for x in analysis.tokens() if x in PRONOUNS_LIST)
return {"explicit_naming": 3.0 * (float(pronouns) / proper_nouns)}
def quantify(analysis):
num_proper_noun_runs = 0
num_proper_noun_tokens = 0
currently_in_run = False
pos_tags = nltk.pos_tag(analysis.case_tokens())
for token in pos_tags:
if is_proper_noun(token):
currently_in_run = True
num_proper_noun_tokens += 1
elif currently_in_run:
# Run just ended
currently_in_run = False
num_proper_noun_runs += 1
if num_proper_noun_runs == 0:
result = 0
else:
result = float(num_proper_noun_tokens) / num_proper_noun_runs
return { "mean_multiple_naming": result }