def test_extract_upc_more_than_one_labeler(self):
# Note 300436122 is not included (invalid length) even though the record
# indicates it's a UPC.
expected_upcs = [
'300676122304',
'032251004292',
'400910131502',
'300670675103',
'300670668150',
'300670675301',
'300670675400',
'300670675004',
'300670675127',
'300676255736',
'300670674106',
'300670674304',
'681131739276',
'681131739283',
'050428068595',
'050428046784',
'050428065747',
'050428579350',
'050428046791',
'050428065822',
'050428160282'
]
xml = open_data_file('upc-more-than-one-labeler.xml')
xml_dict = xmltodict.parse(xml)
extracted_upcs = extract.extract_upc_from_recall(xml_dict['recall-number'])
self.assertListEqual(sorted(expected_upcs), sorted(extracted_upcs) )
def test_extract_upc_with_spaces(self):
expected_upcs = [ '361958010115' ]
xml = open_data_file('upc-with-spaces.xml')
xml_dict = xmltodict.parse(xml)
extracted_upcs = extract.extract_upc_from_recall(xml_dict['recall-number'])
self.assertListEqual(expected_upcs, extracted_upcs,
'upc-with-spaces.xml')
def test_extract_upc_with_spaces_and_dashes(self):
expected_upcs = [ '306037039397' ]
xml = open_data_file('upc-with-spaces-and-dashes.xml')
xml_dict = xmltodict.parse(xml)
extracted_upcs = extract.extract_upc_from_recall(xml_dict['recall-number'])
self.assertListEqual(expected_upcs, extracted_upcs,
'upc-with-spaces-and-dashes.xml')
def map(self, key, value, output):
# These keys must exist in the JSON records for the annotation logic to work
logic_keys = [
'code-info',
'report-date',
'product-description'
]
val = json.loads(value)
if val['product-type'] == 'Drugs':
val['upc'] = extract.extract_upc_from_recall(val)
val['ndc'] = extract.extract_ndc_from_recall(val)
# There is no good ID for the report, so we need to make one
doc_id = self._hash(json.dumps(val, sort_keys=True))
val['@id'] = doc_id
val['@version'] = 1
# Only write out vals that have required keys and a meaningful date
if set(logic_keys).issubset(val) and val['report-date'] != None:
output.add(key, val)
else:
logging.warn('Docuemnt is missing required fields. %s',
json.dumps(val, indent=2, sort_keys=True))
def run(self):
output_dir = self.output().path
os.system('mkdir -p "%s"' % output_dir)
current_event_files = glob.glob(self.input()[0].path + '/*.json')
historic_event_files = glob.glob(self.input()[1].path + '/*.json')
all_files = current_event_files + historic_event_files
output_file = self.output().path + '/all_res.json'
out = open(output_file, 'w')
# These keys must exist in the JSON records for the annotation logic to work
logic_keys = [
'code-info',
'report-date',
'product-description'
]
for filename in all_files:
json_file = open(filename, 'r')
for row in json_file:
record = json.loads(row)
if record['product-type'] == 'Drugs':
record['upc'] = extract.extract_upc_from_recall(record)
record['ndc'] = extract.extract_ndc_from_recall(record)
# Only write out records that have required keys and a meaningful date
if set(logic_keys).issubset(record) and record['report-date'] != None:
out.write(json.dumps(record) + '\n')
def map(self, key, value, output):
def cleaner(k, v):
if k in RENAME_MAP:
k = RENAME_MAP[k]
if k in DATE_KEYS:
if not v:
return None
v = arrow.get(v, 'MM/DD/YYYY').format('YYYYMMDD')
return (k, v)
val = common.transform_dict(value, cleaner)
# These keys must exist in the JSON records for the annotation logic to work
logic_keys = ['code-info', 'report-date', 'product-description']
if val.get('product-type') == 'Drugs':
val['upc'] = extract.extract_upc_from_recall(val)
val['ndc'] = extract.extract_ndc_from_recall(val)
# There is not a decent ID for the report, so we need to make one
doc_id = self._hash(json.dumps(val, sort_keys=True))
val['@id'] = doc_id
val['@version'] = 1
# Only write out vals that have required keys and a meaningful date
if set(logic_keys).issubset(val) and val['report-date'] is not None:
output.add(doc_id, val)
else:
logging.warn('Docuemnt is missing required fields. %s',
json.dumps(val, indent=2, sort_keys=True))
def test_extract_upc_with_dashes(self):
# ensure the best by date "041913 12265 1" is not included
expected_upcs = [ '698997806158' ]
xml = open_data_file('upc-with-dashes.xml')
xml_dict = xmltodict.parse(xml)
extracted_upcs = extract.extract_upc_from_recall(xml_dict['recall-number'])
self.assertListEqual(expected_upcs, extracted_upcs,
'upc-with-dashes.xml')
def run(self):
output_dir = self.output().path
os.system('mkdir -p "%s"' % output_dir)
current_event_files = glob.glob(self.input()[0].path + '/*.json')
historic_event_files = glob.glob(self.input()[1].path + '/*.json')
all_files = current_event_files + historic_event_files
output_file = self.output().path + '/all_res.json'
out = open(output_file, 'w')
# These keys must exist in the JSON records for the annotation logic to work
logic_keys = ['code-info', 'report-date', 'product-description']
for filename in all_files:
json_file = open(filename, 'r')
for row in json_file:
record = json.loads(row)
if record['product-type'] == 'Drugs':
record['upc'] = extract.extract_upc_from_recall(record)
record['ndc'] = extract.extract_ndc_from_recall(record)
# Only write out records that have required keys and a meaningful date
if set(logic_keys).issubset(
record) and record['report-date'] != None:
out.write(json.dumps(record) + '\n')
def map(self, key, value, output):
# These keys must exist in the JSON records for the annotation logic to work
logic_keys = ['code-info', 'report-date', 'product-description']
for val in value['recall-number']:
if val['product-type'] == 'Drugs':
val['upc'] = extract.extract_upc_from_recall(val)
val['ndc'] = extract.extract_ndc_from_recall(val)
# Copy the recall-number attribute value to an actual field
# The recall-number is not a reliable id, since it repeats
val['recall-number'] = val['@id']
# There is no good ID for the report, so we need to make one
doc_id = self._hash(json.dumps(val, sort_keys=True))
val['@id'] = doc_id
val['@version'] = 1
# Only write out vals that have required keys and a meaningful date
if set(logic_keys).issubset(val) and val['report-date'] != None:
output.add(doc_id, val)
else:
logging.warn('Docuemnt is missing required fields. %s',
json.dumps(val, indent=2, sort_keys=True))
