def get_languages(codes):
"""Turn some ISO2 language codes into ISO3 codes."""
supported = []
for code in list_to_alpha3(codes):
if code in LANGUAGES:
supported.append(code)
return '+'.join(sorted(supported))
def get_languages(self, languages):
if not hasattr(self, 'supported_languages'):
from tesserocr import get_languages
_, self.supported_languages = get_languages()
codes = set(['eng'])
for lang in list_to_alpha3(codes):
if lang in self.supported_languages:
codes.add(lang)
return '+'.join(sorted(codes))
def get_models(entity):
"""Iterate over the NER models applicable to the given entity."""
languages = entity.get_type_values(registry.language)
models = set()
for lang in list_to_alpha3(languages):
model = settings.NER_MODELS.get(lang)
if model is not None:
models.add(model)
for model in models:
yield _load_model(model)
def extract_text(self, data, languages=None):
"""Extract text from a binary string of data."""
codes = set(['eng'])
for lang in list_to_alpha3(codes):
if lang in self.supported_languages:
codes.add(lang)
languages = '+'.join(sorted(codes))
api = self.get_api(languages)
if languages != api.GetInitLanguagesAsString():
api.Init(lang=languages)
try:
# TODO: play with contrast and sharpening the images.
image = Image.open(BytesIO(data))
if not self.image_size_ok(image):
return
api.SetImage(image)
return api.GetUTF8Text()
except Exception as ex:
log.warning("Failed to OCR: %s", ex)
finally:
api.Clear()
def test_list(self):
assert 'srp' in list_to_alpha3('bs')
assert 'srp' not in list_to_alpha3('bs', synonyms=False)
assert 'deu' in list_to_alpha3(['bs', 'de'])
def test_list(self):
assert "srp" in list_to_alpha3("bs")
assert "srp" not in list_to_alpha3("bs", synonyms=False)
assert "deu" in list_to_alpha3(["bs", "de"])