class TestSerializations(unittest.TestCase):
"""
Tests for output serialization
"""
def setUp(self):
with open(os.path.join(resources, 'records.json'), 'r') as fp:
self.records = [rpred.ocr_record(**x) for x in json.load(fp)]
self.validator = HocrValidator('standard')
def test_vertical_hocr_serialization(self):
"""
Test vertical line hOCR serialization
"""
fp = StringIO()
fp.write(serialization.serialize(self.records, image_name='foo.png', writing_mode='vertical-lr', template='hocr'))
fp.seek(0)
report = self.validator.validate(fp, parse_strict=True)
self.assertTrue(report.is_valid())
def test_hocr_serialization(self):
"""
Test hOCR serialization
"""
fp = StringIO()
fp.write(serialization.serialize(self.records, image_name='foo.png', template='hocr'))
fp.seek(0)
report = self.validator.validate(fp, parse_strict=True)
self.assertTrue(report.is_valid())
def test_alto_serialization_validation(self):
"""
Validates output against ALTO schema
"""
fp = StringIO()
fp.write(serialization.serialize(self.records, image_name='foo.png', template='alto'))
doc = etree.fromstring(fp.getvalue().encode('utf-8'))
with open(os.path.join(resources, 'alto-4-0.xsd')) as schema_fp:
alto_schema = etree.XMLSchema(etree.parse(schema_fp))
alto_schema.assertValid(doc)
def test_abbyyxml_serialization_validation(self):
"""
Validates output against abbyyXML schema
"""
fp = StringIO()
fp.write(serialization.serialize(self.records, image_name='foo.png', template='abbyyxml'))
doc = etree.fromstring(fp.getvalue().encode('utf-8'))
with open(os.path.join(resources, 'FineReader10-schema-v1.xml')) as schema_fp:
abbyy_schema = etree.XMLSchema(etree.parse(schema_fp))
abbyy_schema.assertValid(doc)
def validate_hocr(self, fp):
fp.seek(0)
validator = HocrValidator('standard')
report = validator.validate(fp, parse_strict=True)
self.assertTrue(report.is_valid())
doc = etree.fromstring(fp.getvalue().encode('utf-8'))
ids = [x.get('id') for x in doc.findall('.//*[@id]')]
counts = Counter(ids)
self.assertEqual(counts.most_common(1)[0][1],
1,
msg='Duplicate IDs in hOCR output')
def main():
args = parser.parse_args()
validator = HocrValidator(args.profile,
skip_check=args.skip_check,
implicit_capabilities=args.implicit_capabilities)
failed = 0
for source in args.sources:
report = validator.validate(source,
parse_strict=args.parse_strict,
filename=args.filename)
failed += not report.is_valid()
if not args.silent:
print(report.format(args.format))
sys.exit(0 if not failed else 1)
class TestSerializations(unittest.TestCase):
"""
Tests for output serialization
"""
def setUp(self):
with open(os.path.join(resources, 'records.json'), 'r') as fp:
self.records = [rpred.ocr_record(**x) for x in json.load(fp)]
self.validator = HocrValidator('standard')
def test_vertical_hocr_serialization(self):
"""
Test vertical line hOCR serialization
"""
fp = StringIO()
fp.write(serialization.serialize(self.records, image_name='foo.png', writing_mode='vertical-lr', template='hocr'))
fp.seek(0)
report = self.validator.validate(fp, parse_strict=True)
self.assertTrue(report.is_valid())
def test_hocr_serialization(self):
"""
Test hOCR serialization
"""
fp = StringIO()
fp.write(serialization.serialize(self.records, image_name='foo.png', template='hocr'))
fp.seek(0)
report = self.validator.validate(fp, parse_strict=True)
self.assertTrue(report.is_valid())
def test_alto_serialization_validation(self):
"""
Validates output against ALTO schema
"""
fp = StringIO()
fp.write(serialization.serialize(self.records, image_name='foo.png', template='alto'))
doc = etree.fromstring(fp.getvalue())
print(fp.getvalue()[:2000])
with open(os.path.join(resources, 'alto-3-1.xsd')) as schema_fp:
alto_schema = etree.XMLSchema(etree.parse(schema_fp))
alto_schema.assertValid(doc)
def setUp(self):
with open(os.path.join(resources, 'records.json'), 'r') as fp:
self.records = [rpred.ocr_record(**x) for x in json.load(fp)]
self.validator = HocrValidator('standard')
def setUp(self):
with open(os.path.join(resources, 'records.json'), 'r') as fp:
self.records = [rpred.ocr_record(**x) for x in json.load(fp)]
self.validator = HocrValidator('standard')