def start_element(name, attrs):
debug('<' + name + '>' + str(attrs))
if name == SENTENCE:
procSentStart()
elif CHUNK.match(name):
procChunkStart(name)
elif name == TOKEN:
try:
pos = attrs[POS]
except:
pos = "PUNCT"
try:
markedEvent = attrs[MARKEDEVENT]
except:
markedEvent = None
try:
lemma = attrs[LEMMA]
except:
lemma = None
procTokStart(pos, markedEvent, lemma)
elif name == CHUNKHEAD:
procChunkHeadStart()
elif name == CHUNKPOSS:
procPOSStart()
currentDoc.addDocNode(startElementString(name, attrs))
def start_element(name, attrs):
debug('<' + name + '>' + str(attrs))
if name == SENTENCE:
procSentStart()
elif CHUNK.match(name):
procChunkStart(name)
elif name == TOKEN:
try:
pos = attrs[POS]
except:
pos = "PUNCT"
try:
markedEvent = attrs[MARKEDEVENT]
except:
markedEvent = None
try:
lemma = attrs[LEMMA]
except:
lemma = None
procTokStart(pos,markedEvent,lemma)
elif name == CHUNKHEAD:
procChunkHeadStart()
elif name == CHUNKPOSS:
procPOSStart()
currentDoc.addDocNode(startElementString(name, attrs))
def start_element(name, attrs):
#logger.debug("\n============\nNAME: "+str(name)+"\n============\n")
#logger.debug("\nATTRS: "+str(attrs)+"\n")
if name == SENTENCE:
#print ">> SENTENCE"
procSentStart()
elif CHUNK.match(name):
#print ">> CHUNK:", name
procChunkStart(name)
elif name == TOKEN:
#print ">> TOKEN:", attrs
if attrs.has_key(POS):
pos = attrs[POS]
else:
pos = "PUNCT"
procTokStart(pos)
elif name == CHUNKHEAD:
#print ">> CHUNKHEAD"
procChunkHeadStart()
elif CHUNKVERBAL.match(name):
#print ">> CHUNKVERBAL:", name
procChunkVerbalStart()
elif name == CHUNKPOSS:
#print ">> CHUNKPOSS"
procPOSStart()
elif name == EVENT:
#print ">> EVENTS:", attrs
procEventStart(attrs)
elif name == INSTANCE:
#print ">> INSTANCE:", attrs
procInstanceStart(attrs)
elif name == TIMEX:
#print ">> TIMEX:", attrs
procTimexStart(attrs)
if name in EMPTY_TAGS:
#print ">> EMPTY TAG:", name, attrs
if name == 'MAKEINSTANCE':
# to avoid confusion with the lex.pos attribute.
# It is a bit hackish.
posVal = attrs[EPOS]
del attrs[EPOS]
attrs[POS] = posVal
currentDoc.addDocNode(emptyContentString(name, attrs))
elif name == TIMEML:
pass
else:
#print "OTHER:", name, attrs
currentDoc.addDocNode(startElementString(name, attrs))
def start_element(name, attrs):
# logger.debug("\n============\nNAME: "+str(name)+"\n============\n")
# logger.debug("\nATTRS: "+str(attrs)+"\n")
if name == SENTENCE:
# print ">> SENTENCE"
procSentStart()
elif CHUNK.match(name):
# print ">> CHUNK:", name
procChunkStart(name)
elif name == TOKEN:
# print ">> TOKEN:", attrs
if attrs.has_key(POS):
pos = attrs[POS]
else:
pos = "PUNCT"
procTokStart(pos)
elif name == CHUNKHEAD:
# print ">> CHUNKHEAD"
procChunkHeadStart()
elif CHUNKVERBAL.match(name):
# print ">> CHUNKVERBAL:", name
procChunkVerbalStart()
elif name == CHUNKPOSS:
# print ">> CHUNKPOSS"
procPOSStart()
elif name == EVENT:
# print ">> EVENTS:", attrs
procEventStart(attrs)
elif name == INSTANCE:
# print ">> INSTANCE:", attrs
procInstanceStart(attrs)
elif name == TIMEX:
# print ">> TIMEX:", attrs
procTimexStart(attrs)
if name in EMPTY_TAGS:
# print ">> EMPTY TAG:", name, attrs
if name == "MAKEINSTANCE":
# to avoid confusion with the lex.pos attribute.
# It is a bit hackish.
posVal = attrs[EPOS]
del attrs[EPOS]
attrs[POS] = posVal
currentDoc.addDocNode(emptyContentString(name, attrs))
elif name == TIMEML:
pass
else:
# print "OTHER:", name, attrs
currentDoc.addDocNode(startElementString(name, attrs))