iiifDLtoGRImg.py

#!/usr/bin/env python3
import urllib.request
import json
import sys
from PIL import Image
import requests
from io import BytesIO
import cv2
import numpy as np
import img2pdf
import argparse


# usage  - see help: $python iiifDL.py -h
# example $python iiifDL.py http://www2.dhii.jp/nijl/NIJL0018/099-0014/manifest.json -s 5 -e 10 
# requirement: Python3, Pillow & requests
# PIL save() with .jpg - default quality is '75'.
# Code written partly by Kiyonori Nagasaki


def getImgUrls(x):
    readObj = urllib.request.urlopen(x)
    data = json.loads(readObj.read().decode('utf-8'))
    # get Image Urls as a list
    iiif3 = 'http://iiif.io/api/presentation/3/context.json'
    arImageUrls = []
    if data['@context'] == iiif3:
        ver3 = True
        for seq in data['items']:
            for canvas in seq['items']:
                for image in canvas['items']:
                    arImageUrls.append(image['body']['id'])
    else:
        ver3 = False
        for seq in data['sequences']:
            for canvas in seq['canvases']:
                for image in canvas['images']:
                    arImageUrls.append(image['resource']['service']['@id'])
    return arImageUrls, ver3

if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument('url', help='the url of iiif manifest')
    parser.add_argument('-s', '--startPage', type=int, default=1, help='start page (default = 1)')
    parser.add_argument('-e', '--endPage', type=int, default=50, help='end page (default = 50)')
    args = parser.parse_args()

    url = args.url
    startPg = args.startPage
    endPg = args.endPage

    urlList, ver3 = getImgUrls(url)
    length = len(urlList)
    tLen = len(str(length)) + 1
    img_list = []

    print ('downloading...')
    for i in range(startPg - 1, endPg):
        if i == length:
            endtext = str(length) + ' is the last page'
            print (endtext)
            break
        localFilename = './img/' + str(startPg).zfill(tLen) + '.png'
        imageUrl = urlList[i]
        if (ver3):
            imageUrl = imageUrl
        else:
            imageUrl = imageUrl+'/full/full/0/default.jpg'
        response = requests.get(imageUrl)
        imgObj = Image.open(BytesIO(response.content))
        
        grau_img = cv2.cvtColor(np.float32(imgObj), cv2.COLOR_BGR2GRAY)
        cv2.imwrite(localFilename, grau_img)
        startPg = startPg + 1
        print (localFilename)