def load_file(self, filepath: tuple):
phrase_tree_path = get_resource(filepath[0])
dep_tree_path = get_resource(filepath[1])
pf = TimingFileIterator(phrase_tree_path)
message_prefix = f'Loading {os.path.basename(phrase_tree_path)} and {os.path.basename(dep_tree_path)}'
for i, (dep_sent, phrase_sent) in enumerate(
zip(read_tsv_as_sents(dep_tree_path), pf)):
# Somehow the file contains escaped literals
phrase_sent = phrase_sent.replace('\\/', '/')
token = [x[1] for x in dep_sent]
pos = [x[3] for x in dep_sent]
head = [int(x[6]) for x in dep_sent]
rel = [x[7] for x in dep_sent]
phrase_tree = load_trees_from_str(phrase_sent, [head], [rel],
[token])
assert len(
phrase_tree
) == 1, f'{phrase_tree_path} must have on tree per line.'
phrase_tree = phrase_tree[0]
yield {
'FORM': token,
'CPOS': pos,
'HEAD': head,
'DEPREL': rel,
'tree': phrase_tree,
'hpsg': phrase_tree.convert()
}
pf.log(
f'{message_prefix} {i + 1} samples [blink][yellow]...[/yellow][/blink]'
)
pf.erase()
def load_file(self, filepath: str):
filepath = get_resource(filepath)
if os.path.isfile(filepath):
files = [filepath]
else:
assert os.path.isdir(
filepath), f'{filepath} has to be a directory of CoNLL 2012'
files = sorted(
glob.glob(f'{filepath}/**/*gold_conll', recursive=True))
timer = CountdownTimer(len(files))
for fid, f in enumerate(files):
timer.log(f'files loading[blink][yellow]...[/yellow][/blink]')
# 0:DOCUMENT 1:PART 2:INDEX 3:WORD 4:POS 5:PARSE 6:LEMMA 7:FRAME 8:SENSE 9:SPEAKER 10:NE 11-N:ARGS N:COREF
for sent in read_tsv_as_sents(f, ignore_prefix='#'):
sense = [cell[7] for cell in sent]
props = [cell[11:-1] for cell in sent]
props = map(lambda p: p, zip(*props))
prd_bio_labels = [
self._make_bio_labels(prop) for prop in props
]
prd_bio_labels = [self._remove_B_V(x) for x in prd_bio_labels]
prd_indices = [i for i, x in enumerate(sense) if x != '-']
token = [x[3] for x in sent]
srl = [None for x in token]
for idx, labels in zip(prd_indices, prd_bio_labels):
srl[idx] = labels
srl = [x if x else ['O'] * len(token) for x in srl]
yield {'token': token, 'srl': srl}
def bmes_to_flat(inpath, outpath):
with open(outpath, 'w', encoding='utf-8') as out:
for sent in read_tsv_as_sents(inpath):
chunks = get_entities([cells[1] for cells in sent])
chars = [cells[0] for cells in sent]
words = []
for tag, start, end in chunks:
word = ''.join(chars[start: end])
words.append(word)
out.write(' '.join(f'{word}/{tag}' for word, (tag, _, _) in zip(words, chunks)))
out.write('\n')
def ctb_pos_to_text_format(path, delimiter='_'):
"""
Convert ctb pos tagging corpus from tsv format to text format, where each word is followed by
its pos tag.
Args:
path: File to be converted.
delimiter: Delimiter between word and tag.
"""
path = get_resource(path)
name, ext = os.path.splitext(path)
with open(f'{name}.txt', 'w', encoding='utf-8') as out:
for sent in read_tsv_as_sents(path):
out.write(' '.join([delimiter.join(x) for x in sent]))
out.write('\n')
def make_con_txt(conll_file, output_file):
with open(output_file, 'w') as out:
for sent in read_tsv_as_sents(conll_file):
tree = []
pos_per_sent = []
for cell in sent:
if cell[0] == '#begin' or cell[0] == '#end':
continue
if len(cell) < 8:
print(cell)
filename, sentence_id, token_id, word, POS, parse, framefile, roleset, *_ = cell
parse = parse.replace('*', f'({POS} {word})')
tree.append(parse)
pos_per_sent.append(POS)
bracketed = ' '.join(tree)
out.write(bracketed)
out.write('\n')
def convert_conll03_to_json(file_path):
dataset = []
num_docs = [0]
def new_doc():
doc_key = num_docs[0]
num_docs[0] += 1
return {
'doc_key': doc_key,
'sentences': [],
'ner': [],
}
doc = new_doc()
offset = 0
for cells in read_tsv_as_sents(file_path):
if cells[0][0] == '-DOCSTART-' and doc['ner']:
dataset.append(doc)
doc = new_doc()
offset = 0
sentence = [x[0] for x in cells]
ner = [x[-1] for x in cells]
ner = hanlp.utils.span_util.iobes_tags_to_spans(ner)
adjusted_ner = []
for label, (span_start, span_end) in ner:
adjusted_ner.append(
[span_start + offset, span_end + offset, label])
doc['sentences'].append(sentence)
doc['ner'].append(adjusted_ner)
offset += len(sentence)
if doc['ner']:
dataset.append(doc)
output_path = os.path.splitext(file_path)[0] + '.json'
with open(output_path, 'w') as out:
for each in dataset:
json.dump(each, out)
out.write('\n')
