Script Python pour aspirer les documents numérisés de la Zentral- und Landesbibliothek de Berlin. Extrait les pages en JPG et assemble un PDF. Les URL soumis doivent être du type : https://digital.zlb.de/viewer/image/16054086_1911/12/, c'est à dire l'adresse quand vous lisez une page depuis la visionneuse.

import os
import re
from PIL import Image # "pip install Pillow" to install package
import requests # "pip install requests" to install package
import shutil
import mimetypes
from bs4 import BeautifulSoup # "pip install beautifulsoup4" to install package
headers = {'User-Agent': 'Mozilla/5.0'}

# functions
def downloadFile(url, fileName):
    url = url.replace("\\","")
    response = requests.get(url, stream=True)
    content_type = response.headers['content-type']
    extension = mimetypes.guess_extension(content_type)
    fileName += extension
    with open(fileName, 'wb') as out_file:
        shutil.copyfileobj(response.raw, out_file)
        del response

def getRecordName(url):
    reg = re.match(r"https?:\/\/digital.zlb.de\/viewer\/image\/([A-Za-z0-9_]*)\/", url)
    if reg:
        return reg.group(1)
    else:
        print("Erreur : format d'URL incorrect. Exemple correct https://digital.zlb.de/viewer/image/16054086_1911/9/LOG_0006/")
        return None

def getMaxPagesFromRecord(record):
    url = "https://digital.zlb.de/viewer/image/"+record+"/"
    response = requests.get(url, headers=headers)
    if response:
        reg = re.search(r"rft\.tpages=(\d+)&", str(response.text))
        if reg:
            return int(reg.group(1))
        else:
            print("Impossible de trouver le nombre maximal de page automatiquement.")
            return int(input("Nombre de pages à extraire :"))
    else:
        print("Erreur : la page ZLB n'est pas chargeable.")
        return None

def findImageFromUrl(url,resolution):
    response = requests.get(url, headers=headers)
    if response:
        soup = BeautifulSoup(response.text,"html.parser")
        metaImage = soup.findAll("meta", {"property" : "og:image"})
        img = metaImage[0].get("content")
        img = img.replace("/300,/","/"+str(resolution)+",/")
        return img
    else:
        print("Erreur : la page ZLB n'est pas chargeable.")


#input
inputUrl = input("Adresse de la page ZLB à aspirer :")
resolution = int(input("Résolution (1300 recommandé) :"))

#get record infos
record = getRecordName(inputUrl)
if record:
    maxPages = getMaxPagesFromRecord(record)

#download pages as JPG
if record and maxPages:
    i = 1
    while i < (maxPages+1):
        print("Page",i,"sur",maxPages)
        url = "https://digital.zlb.de/viewer/image/"+record+"/"+str(i)+"/"
        img = findImageFromUrl(url,resolution)
        downloadFile(img, record+"_"+str(i).zfill(8))
        i += 1

#assemble PDF
filesInFolder = os.listdir()
imagesInFolder = []
imagesRGBInFolder = []
for file in filesInFolder:
    ext = os.path.splitext(file)[1]
    imageExtensions = [".jpg",".png",".gif",".jpeg",".svg"]
    if ext in imageExtensions:
        imagesInFolder.append(file)
        img = Image.open(file)
        if img.mode == "RGBA":
            img = img.convert(RGB)
        imagesRGBInFolder.append(img)

if imagesRGBInFolder:
    imagesRGBInFolder[0].save(record+".pdf", save_all=True, quality=85, append_images=imagesRGBInFolder[1:])

#delete JPG
    for file in imagesInFolder:
        os.remove(file)

Pour éviter des erreurs de mémoire avec les documents de beaucoup de pages, il faut utiliser une installation 64bits de Python.