File size: 3,847 Bytes
207e462 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 |
"""
Example data reader for the annotated SHINRA-5LDS
Created May 2, 2024.
Copyright Hassan S. Shavarani
"""
import json
import datetime
from zipfile import ZipFile
from tqdm import tqdm
from collections import namedtuple
ENELabel = namedtuple('ENELabel', ['level_id', 'name'])
class ENEAnnotation:
def __init__(self, ann):
self.ene_id = ann['ene_id']
self.representative_name = ann['representative_name']
self._id = int(ann['_id'])
self.original_level = ann['original_level']
self.values = {x['level']: ENELabel(x['level_id'], x['name']) for x in ann['entities']}
class SHINRAAnnotation:
def __init__(self, annotation):
self.lang = annotation['lang'] # this value is always equal to 'ja' since all original annotations have been assigned in Japanese
self.annotator = annotation['annotator']
self.annotation_date = datetime.datetime.strptime(annotation['date'], '%Y-%m-%d %H:%M:%S')
self.raw_flag = annotation['raw_flag']
self.type = annotation['type']
self.annotation_confidence = annotation['probability_score']
self.labels = ENEAnnotation(annotation['ene_id'])
class SHINRAArticle:
def __init__(self, article):
self.title = article['title']
self.content = article['content']
self.url = article['url']
self.page_id = article['page_id']
self.links = article['links']
self.summary = article['summary']
self.categories = article['categories']
class SHINRARecord:
def __init__(self, record):
self.annotation_id = record['annotation_id']
self.ene_level_ids = {
"0": record['ene_level_ids']['level0'],
"1": record['ene_level_ids']['level1'],
"2": record['ene_level_ids']['level2'],
"3": record['ene_level_ids']['level3']
}
self.ene_label_verified = record['ene_label_verified']
self.annotations = [SHINRAAnnotation(x) for x in record['annotations'] if x['annotator'] == 'HAND']
self.articles = {x['language']: SHINRAArticle(x) for x in record['annotation_articles']}
@property
def latest_annotation(self):
return max(self.annotations, key=lambda x: x.annotation_date)
class SHINRA5LDS:
"""
Size of the SHINRA-5LDS dataset (considering only the hand annotated records):
ja: 118635 records
en: 52445 records
fr: 34432 records
de: 29808 records
fa: 14058 records
"""
def __init__(self, zip_file_path, lang):
self.zip_file_path = zip_file_path
self.data = None
self.lang = lang
def _load_data(self):
zip_ref = ZipFile(self.zip_file_path, 'r')
self.data = zip_ref.open('data.jsonl')
def __iter__(self):
if self.data is None or len(self.data) == 0:
self._load_data()
return self
def __next__(self):
try:
while True:
next_line = self.data.readline()
if not next_line:
self.data.close()
self.data = None
raise IndexError
record = SHINRARecord(json.loads(next_line))
if not record.annotations: # the annotation is machine generated
continue
if self.lang in record.articles:
break
return record.articles[self.lang], record.latest_annotation
except IndexError:
raise StopIteration
except Exception as e:
print(f"Error: {e}")
return self.__next__()
if __name__ == "__main__":
for article, annotation in tqdm(SHINRA5LDS('SHINRA-5LDS.zip', 'en')):
print('='*80)
print(article.title, annotation.labels.values)
print('='*80)
|