Modified Python Script for MS Computer Vision API (OCR)

do_ocr_ms.py

#! /usr/bin/env python3
# -*- coding: utf-8 -*-
import http.client, urllib.request, urllib.parse, urllib.error, base64

# https://dev.projectoxford.ai/docs/services/56f91f2d778daf23d8ec6739/operations/56f91f2e778daf14a499e1fc

import http.client, urllib.request, urllib.parse, urllib.error, base64
import os, json, argparse


api_key = os.environ['MSCV_API_KEY']
# print(api_key)

parser = argparse.ArgumentParser()
parser.add_argument("imagefile", help = "image file for ocr(png, jpg)")
args = parser.parse_args()
filename = args.imagefile

# print(filename)

headers = {
    # Request headers
    #'Content-Type': 'application/json',
    'Content-Type': 'application/octet-stream',
    'Ocp-Apim-Subscription-Key': api_key,
}

params = urllib.parse.urlencode({
    # Request parameters
    'language': 'unk',
    'detectOrientation ': 'true',
})

try:
    conn = http.client.HTTPSConnection('westus.api.cognitive.microsoft.com')
    img  = open(filename, 'rb').read()
    conn.request("POST", "/vision/v1.0/ocr?%s" % params, img, headers)
    response = conn.getresponse()

    data = response.read().decode( "utf-8") 
    json_dict = json.loads(data)

#    print(json_dict['regions'])
    for region  in json_dict['regions'] :
        print("region: {}".format(region['boundingBox']))

        for line in region['lines'] :
            print("line: {}".format(line['boundingBox']))
            for word in line['words'] :
                print(word['text'], end="")

            print("\n")
   # print("\n")

    conn.close()

except Exception as e:
    print("[Errno {0}] {1}".format(e.errno, e.strerror))