def if_azure(self):
ground_truth, xml_status = self.compute_ground_truth()
if xml_status != 0: return xml_status, 0
try:
subscription_key, endpoint = os.environ[
'COMPUTER_VISION_SUBSCRIPTION_KEY'], os.environ[
'COMPUTER_VISION_ENDPOINT']
ocr_url = endpoint + "vision/v2.1/ocr"
headers = {
'Ocp-Apim-Subscription-Key': subscription_key,
'Content-Type': 'application/octet-stream'
}
params = {'detectOrientation': 'true'}
response = requests.post(ocr_url,
headers=headers,
params=params,
data=self.content)
response.raise_for_status()
lines_info = [
region["lines"] for region in response.json()["regions"]
]
detected_str = extract_clean_str(''.join([word_info['text'] for line in lines_info \
for word_metadata in line for word_info in word_metadata["words"]]))
self.detected_str = detected_str
score = jaccard_similarity(detected_str, ground_truth)
return 0, score
except:
return -4, 0
def if_gc(self):
ground_truth, xml_status = self.compute_ground_truth()
if xml_status != 0: return xml_status, 0
try:
vision_client = vision.ImageAnnotatorClient()
image = vision.types.Image(content=self.content)
text_detection_response = vision_client.text_detection(image=image)
detected_str = extract_clean_str(text_detection_response.full_text_annotation.text)
self.detected_str = detected_str
score = jaccard_similarity(detected_str, ground_truth)
return 0, score
except:
return -4, 0
def if_aws(self):
ground_truth, xml_status = self.compute_ground_truth()
if xml_status != 0: return xml_status, 0
try:
imgobj = {'Bytes': self.content}
client = boto3.client('rekognition', region_name='us-east-1')
response = client.detect_text(Image=imgobj)
detected_str = extract_clean_str(''.join([txt['DetectedText'] for txt in response['TextDetections'] \
if txt['Type']=='WORD']))
self.detected_str = detected_str
score = jaccard_similarity(detected_str, ground_truth)
return 0, score
except:
return -4, 0