def from_text(cls, text):
input_lines = [l.strip() for l in text.splitlines() if l.strip() != '']
parse_entity = False
parse_event = False
script = cls(input_lines[0])
for line in input_lines[1:]:
if line == 'Entities:':
assert (not parse_entity) and (not parse_event)
parse_entity = True
continue
elif line == 'Events:':
assert parse_entity and (not parse_event)
parse_event = True
parse_entity = False
continue
if parse_entity:
entity = Entity.from_text(line.partition('\t')[2])
script.add_entity(entity)
elif parse_event:
event = Event.from_text(line.partition('\t')[2])
script.add_event(event)
else:
raise RuntimeError(
'cannot parse EventScript from: {}'.format(text))
return script
def from_text(cls, text):
if len(text.splitlines()) == 1:
# Empty document: error in event extraction
return cls(text.strip(), [], [])
sections = split_sections((l.strip() for l in text.splitlines()),
["Entities:", "Events:"])
# First comes the doc name
doc_name = sections[0][0].strip()
# Then a whole section giving the entities, matched by entity_line_re
entity_lines = [
line.partition('\t')[2] for line in sections[1] if line
]
entities = [Entity.from_text(line) for line in entity_lines]
# Then another section giving the events, matched by event_line_re
event_lines = [line.partition('\t')[2] for line in sections[2] if line]
events = [Event.from_text(line) for line in event_lines]
return cls(doc_name, entities, events)
