def query(entity, attr):
soup = To.get_html_baidu("http://baike.baidu.com/item/" + entity)
basicInfo_block = soup.find(class_='basic-info cmn-clearfix')
if basicInfo_block == None:
# print 'info None'
return attr + "::找不到"
else:
info = get_info(basicInfo_block)
# for i in info:
# print i
# print info[i]
# print '-----------'
if info.has_key(attr.decode('utf8')):
# print 'has key'+attr.decode('utf8')
return info[attr.decode('utf8')]
else:
# print 'no key 进行同义词判断'
# 同义词判断
attr_list = T.load_baikeattr_name(
os.path.dirname(os.path.split(os.path.realpath(__file__))[0]) +
'/resources/Attribute_name.txt')
attr = T.load_synonyms_word_inattr(
attr,
os.path.dirname(os.path.split(os.path.realpath(__file__))[0]) +
'/resources/SynonDic.txt', attr_list)
if info.has_key(attr.decode('utf8')):
return info[attr.decode('utf8')]
else:
return attr + "::找不到"
def query(entity, attr):
entity_uri = 'http://baike.baidu.com/item/' + entity
result = '查询百科列表实体:' + entity_uri + '\n'
soup = To.get_html_baidu(entity_uri)
basicInfo_block = soup.find(class_='basic-info cmn-clearfix')
if basicInfo_block == None:
return result + entity + "-找不到\n"
else:
info = get_info(basicInfo_block)
if attr in info:
return info[attr]
else:
# 同义词判断
result += '属性' + attr + '-找不到\n'
attr_list = T.load_baikeattr_name(
os.path.dirname(os.path.split(os.path.realpath(__file__))[0]) +
'/resources/Attribute_name.txt')
attr = T.load_synonyms_word_inattr(
attr,
os.path.dirname(os.path.split(os.path.realpath(__file__))[0]) +
'/resources/SynonDic.txt', attr_list)
if attr in info:
return info[attr]
else:
return result + '同义属性' + attr + '-找不到\n'