def parse_odin_inst(string, corpus = None, idnum=None):
"""
Method to parse and create an IGT instance from odin-style text.
"""
# Start by looking for the doc_id, and the line range.
doc_re = re.search('doc_id=(\S+)\s([0-9]+)\s([0-9]+)\s(.*)\n', string)
docid, lnstart, lnstop, tagtypes = doc_re.groups()
if idnum is not None:
id = gen_item_id('i', idnum)
elif corpus:
id = corpus.askIgtId()
else:
corpus = XigtCorpus()
id = 'i{}'.format(len(corpus))
inst = Igt(id = id, attributes={'doc-id':docid,
'line-range':'%s %s' % (lnstart, lnstop),
'tag-types':tagtypes})
# Now, find all the lines
lines = re.findall('line=([0-9]+)\stag=(\S+):(.*)\n?', string)
# --- 3) Create a raw tier.
rt = Tier(id = RAW_ID, type=ODIN_TIER_TYPE, attributes={STATE_ATTRIBUTE:RAW_STATE}, igt=inst)
for lineno, linetag, linetxt in lines:
l = Item(id = ask_item_id(rt), text=linetxt, attributes={'tag':linetag, 'line':lineno}, tier=rt)
rt.append(l)
inst.append(rt)
basic_processing(inst)
return inst
def xc_load(path, mode=FULL, do_basic_processing=False):
f = open(path, 'r', encoding='utf-8')
xc = xigtxml.load(f, mode=mode)
if do_basic_processing:
for inst in xc:
basic_processing(inst)
return xc