def read_image_with_pil(image_directory):
try:
img = Image.open(image_directory)
return img
except IOError:
orc_logger.error('PIL can not read a file, maybe it is not a image')
raise IOError
def _validate_file_extension():
_, file_extension = os.path.splitext(image_directory)
if file_extension.upper() in VALID_EXTEND_FILE:
orc_logger.info('end validate input')
return file_extension
orc_logger.error(f"only allow file extend: {VALID_EXTEND_FILE}")
raise Exception
def get_string(input, output, verbose):
try:
if output:
orc_logger.set_log_name(output)
orc_logger.info(f'start time: {datetime.datetime.now()}')
orc_logger.info(f'receive input file: {input}')
orc_logger.info(f'receive output file: {output}')
type_file = validate_file(input)
if type_file == VALID_EXTEND_FILE[1]:
input = convert_pdf_to_img()
text = recognize_text_with_tesseract(input)
orc_logger.info(f'result = \n {text}')
except Exception as e:
orc_logger.error(e)
finally:
orc_logger.info(f'end time: {datetime.datetime.now()}')
def recognize_text_with_tesseract(image_directory):
try:
orc_logger.info(f'start recognize text with tesseract')
orc_logger.info(f'read image and convert it to gray')
img = cv2.imread(image_directory)
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
orc_logger.info(f'remove noise in image')
kernel = np.ones((1, 1), np.uint8)
img = cv2.dilate(img, kernel, iterations=1)
cv2.erode(img, kernel, iterations=1)
cv2.imwrite(TEMP_IMG, img)
image_detect = Image.open(TEMP_IMG)
orc_logger.info(f'start recognize text with tesseract')
result = pytesseract.image_to_string(image_detect)
orc_logger.info(f'end recognize text with tesseract')
return result.strip()
except Exception as e:
orc_logger.error(e)
def _validate_file_exist():
if not os.path.isfile(image_directory) or not os.access(
image_directory, os.R_OK):
orc_logger.error('file not exists or not readable')
raise Exception