def clean_list_record(self, record): norm_dict_space(record, 'legal_name', 'address_raw', 'license', 'dba', 'start_date', 'application_date') record['dba'], record['business_type'] = record['dba'].rsplit(' - ', 1) record['start_date'] = parse_date(record['start_date'], '%m/%d/%Y') record['application_date'] = parse_date(record['application_date'], '%m/%d/%Y') # Remove the "Suite/Apt" or "Floor" clause from the address, if it exists. record['address_raw'] = record['address_raw'].replace(' ,', ',') m = re.search(r'^(.*?), (?:Suite/Apt|Floor):.*$', record['address_raw']) if m: record['address_raw'] = m.group(1) return record
def clean_list_record(self, record): if record['last_inspection_date'].lower() == 'not available': raise SkipRecord('No inspection available') else: record['last_inspection_date'] = parse_date(record['last_inspection_date'], '%m/%d/%Y') if record['aka']: record['aka'] = list_aka_re.findall(record['aka'])[0] else: record['aka'] = '' norm_dict_space(record, 'name', 'dba', 'address') record['result'] = record['result'].replace(' ', '').strip() record['city_id'] = int(record['city_id']) # Remove the trailing ZIP code from the address, if it exists. m = re.search(r'(.*?)\s+\d\d\d\d\d$', record['address']) if m: record['address'] = m.group(1) record['address'] = clean_address(record['address']) return record
def clean_list_record(self, record): if record['last_inspection_date'].lower() == 'not available': raise SkipRecord('No inspection available') else: record['last_inspection_date'] = parse_date( record['last_inspection_date'], '%m/%d/%Y') if record['aka']: record['aka'] = list_aka_re.findall(record['aka'])[0] else: record['aka'] = '' norm_dict_space(record, 'name', 'dba', 'address') record['result'] = record['result'].replace(' ', '').strip() record['city_id'] = int(record['city_id']) # Remove the trailing ZIP code from the address, if it exists. m = re.search(r'(.*?)\s+\d\d\d\d\d$', record['address']) if m: record['address'] = m.group(1) record['address'] = clean_address(record['address']) return record