Post-Processing the Recognition Results

• You have subscribed to Passport OCR by referring to Subscribing to an OCR Service.
• You have installed the OCR Python SDK by referring to Calling an OCR API Locally and have installed dependency packages by running the pip install xlsxwriter command.
• You have obtained the AK and SK by accessing the Access Keys page. On this page, you can create an access key or use an existing one. The AK and SK are contained in the credentials.csv file.

  

# -*- coding: utf-8 -*-
import base64
import xlsxwriter

from huaweicloudsdkcore.auth.credentials import BasicCredentials
from huaweicloudsdkocr.v1.region.ocr_region import OcrRegion
from huaweicloudsdkcore.exceptions import exceptions
from huaweicloudsdkocr.v1 import *
from huaweicloudsdkcore.http.http_config import HttpConfig


def recognize_id_card_request():
    try:
        request = RecognizeIdCardRequest()
        request.body = IdCardRequestBody(
            image=image_base64
        )
        response = client.recognize_id_card(request)
        return response
    except exceptions.ClientRequestException as e:
        print(e.status_code)
        print(e.request_id)
        print(e.error_code)
        print(e.error_msg)


def get_credential():
    return BasicCredentials(ak, sk)


def get_client():
    config = HttpConfig.get_default_config()
    config.ignore_ssl_verification = True
    return OcrClient.new_builder(OcrClient) \
        .with_credentials(credentials) \
        .with_region(OcrRegion.CN_NORTH_4) \
        .with_http_config(config) \
        .build()


def image_to_base64(imagepath):
    """
    Convert a local image into Base64 code.
    """
    with open(imagepath, "rb") as bin_data:
        image_data = bin_data.read()
    base64_data = base64.b64encode(image_data).decode("utf-8")
    return base64_data


def response_to_execl(save_file, data):
    """
    :param save_file: File name
    :param data: result data
    """

    # Process the result data returned after the API is called.
    keys_list = list(data["result"].keys())
    values_list = list(data["result"].values())

    options = {'in_memory': True}
    with xlsxwriter.Workbook(save_file, options) as workbook:
        worksheet = workbook.add_worksheet()
        worksheet.set_column('A1:A20', 23)
        worksheet.set_column('B1:B20', 100)
        worksheet.write_column('A1', keys_list)
        worksheet.write_column('B1', values_list)
    workbook.close()


if __name__ == '__main__':
    # Enter the AK and SK.
    ak = "Enter the AK."
    sk = "Enter the SK."

    # Init Auth Info
    credentials = get_credential()

    # Create OcrClient
    client = get_client()

    image_base64 = image_to_base64 (r"Image path, for example, D:\local\test.png")

    # request id card service
    response = recognize_id_card_request().to_dict()

    # Store the data in the Excel file.
    response_to_execl(r"Excel file path, for example, D:\local\test.xlsx", response)

• You have subscribed to Passport OCR by referring to Subscribing to an OCR Service.
• You have installed the OCR Python SDK by referring to Calling an OCR API Locally and have installed dependency packages by running the pip install fitz and pip install PyMuPDF==1.18.0 commands.
• You have obtained the AK and SK by accessing the Access Keys page. On this page, you can create an access key or use an existing one. The AK and SK are contained in the credentials.csv file.

  

# -*- coding: utf-8 -*-
import os
import base64
import fitz
import io
from PIL import Image
from glob import glob


from huaweicloudsdkcore.auth.credentials import BasicCredentials
from huaweicloudsdkocr.v1.region.ocr_region import OcrRegion
from huaweicloudsdkcore.exceptions import exceptions
from huaweicloudsdkocr.v1 import *
from huaweicloudsdkcore.http.http_config import HttpConfig


class CovertPdfToJpg:
    def __init__(self, file_path, save_root):
        self.file_path = file_path
        self.save_root = save_root

    @staticmethod
    def open_pdf(file):
        return fitz.open(file)

    @staticmethod
    def get_trans(doc, page, min_side=0, max_side=0, rotate=0.0):
        """ Create a scale object. """
        region = doc[page].rect
        scale = 1
        if max_side > min_side > 0:
            scale = min_side / min(region.width, region.height)
            if max(region.width, region.height) * scale > max_side:
                scale = max_side / max(region.width, region.height)
        trans = fitz.Matrix(scale, scale).preRotate(rotate)
        return trans

    def page2pix(self, doc, page, trans):
        """ Parse the current page as image data based on given parameters."""
        return doc[page].getPixmap(matrix=trans, alpha=False)

    def pdf_to_jpg(self, width=1024, height=1400):
        """ Convert PDF images to JPG images."""
        doc = self.open_pdf(self.file_path)
        save_dir = os.path.join(self.save_root)
        if not os.path.exists(save_dir):
            os.makedirs(save_dir)
        print("document", len(doc), doc.pageCount)
        for i in range(len(doc)):
            trans = self.get_trans(doc, i, width, height, rotate=0)
            try:
                pdf = self.page2pix(doc, i, trans)
            except:
                continue
            image = pdf.getPNGData()
            image = Image.open(io.BytesIO(image))
            print(os.path.join(
                save_dir, os.path.basename(self.file_path).replace('.pdf', '') + '_' + str(i + 1) + '.jpg'))
            image.save(
                os.path.join(save_dir, os.path.basename(self.file_path).replace('.pdf', '') + '_' + str(i + 1) + '.jpg'))
        return


def recognize_general_text_request():
    try:
        request = RecognizeGeneralTextRequest()
        request.body = GeneralTextRequestBody(
            image=image_base64
        )
        response = client.recognize_general_text(request)
        print(response)
    except exceptions.ClientRequestException as e:
        print(e.status_code)
        print(e.request_id)
        print(e.error_code)
        print(e.error_msg)


def get_credential():
    return BasicCredentials(ak, sk)


def get_client():
    config = HttpConfig.get_default_config()
    config.ignore_ssl_verification = True
    return OcrClient.new_builder(OcrClient) \
        .with_credentials(credentials) \
        .with_region(OcrRegion.CN_NORTH_4) \
        .with_http_config(config) \
        .build()


def url_to_base64(imagepath):
    with open(imagepath, "rb") as bin_data:
        image_data = bin_data.read()
    base64_data = base64.b64encode(image_data).decode("utf-8")
    return base64_data


if __name__ == '__main__':
    # Enter the AK and SK.
    ak = "Enter the AK."
    sk = "Enter the SK."

    # Init Auth Info
    credentials = get_credential()

    # Create OcrClient
    client = get_client()

    df_path = r"Local PDF file path, for example, D:\local\test.pdf"
    save_path = r"Path of the converted image, for example, D:\local"

    covert_pdf_to_jpg = CovertPdfToJpg(df_path, save_path)
    covert_pdf_to_jpg.pdf_to_jpg()

    jpgs = glob(os.path.join(save_path, "*.jpg"))
    for jpg in jpgs:
        image_base64 = url_to_base64(jpg)
        recognize_general_text_request()

If the error message "AttributeError: 'Document' object has no attribute 'pageCount'" is displayed, uninstall the PyMuPDF dependency package and install version 1.18.0 instead.