interjection_skip = False
cnt_brackets_opening = 0
cnt_brackets_closing = 0
continue
else:
line = line.replace('<interjection_begin>',
'').replace('<interjection_end>', '')
if line:
interjection_text.append(line)
interjection_length += 1
continue
if current_speaker is not None and not endend_with_interjection:
if interjection_complete:
interjection_complete = None
text = []
line = helper.cleans_line(line)
text.append(line)
continue
else:
current_role = current_role.strip()
line = helper.cleans_line(line)
text.append(line)
continue
if s is not None:
if not endend_with_interjection:
if ":* " in line:
line = line.split(':* ', 1)[-1]
elif ":" in line:
line = line.split(':', 1)[-1]
cnt_brackets_opening = 0
cnt_brackets_closing = 0
else:
line = line.replace('<interjection_begin>',
'').replace('<interjection_end>', '')
if line and not line.isspace():
interjection_text.append(line)
interjection_length += 1
continue
if current_speaker is not None and not endend_with_interjection:
if interjection_complete:
interjection_complete = None
text = []
if line and not line.isspace() and not INTERJECTION_END.search(
line):
line = helper.cleans_line(line)
text.append(line)
continue
else:
current_role = current_role.strip()
line = helper.cleans_line(line)
if line and not line.isspace():
text.append(line)
continue
if s is not None:
if ":* " in line:
line = line.split(':* ', 1)[-1]
elif ":" in line:
line = line.split(':', 1)[-1]
}
table.insert(speech_dict)
#import pdb; pdb.set_trace()
sub += 1
interjection = True
interjection_text = []
if interjection:
cnt_brackets_opening += line.count('(')
cnt_brackets_closing += line.count(')')
#import pdb; pdb.set_trace()
if INTERJECTION_END.search(
line
) and cnt_brackets_opening <= cnt_brackets_closing or CHAIR_MARK.match(
line):
if current_speaker is not None:
interjection_text.append(helper.cleans_line(line))
interjection_text = [
i + ' ' if not i.endswith('-') else i
for i in interjection_text
]
interjection_text = ''.join(interjection_text)
# removes whitespace duplicates
interjection_text = re.sub(' +', ' ', interjection_text)
# removes whitespaces at the beginning and end
interjection_text = interjection_text.strip()
interjection_text = re.sub('-(?=[a-z])', '',
interjection_text)
if debug:
speech = pd.DataFrame({
'speaker': [current_speaker],