Files
dotfiles/Userdata/Scripts/rippers/trauner_ripper/ripper.py
T
2020-05-05 10:43:39 +02:00

76 lines
3.0 KiB
Python

import requests
import urllib.request
import xml.etree.ElementTree as ET
import subprocess
from pathlib import Path
import os
maxpages = 396
ebooknr = 584
contentdir = "./pages" # ./pages dir
pagedirtemplate = contentdir + "/{}" # ./pages/1 pagenr dir
subdirtemplate = pagedirtemplate + "/{}" # ./pages/1/img pagenr, "img" dir
svgsavetemplate = pagedirtemplate + "/{}.{}" # ./pages/1/1.svg pagenr, pagenr, "svg" file
imgsavetemplate = subdirtemplate + "/{}.{}" # ./pages/1/img/1.png pagenr, "img", imgnr, "png" file
# https://a.trauner-digibox.com/ebook/584/1/1.svg
# ebook number, page number, page number, sxtension
svgurltemplate = "https://a.trauner-digibox.com/ebook/{}/{}/{}.{}"
# https://a.trauner-digibox.com/ebook/584/1/img/1.png
# ebook number, page number, img kind, img number, extension
imgurltemplate = "https://a.trauner-digibox.com/ebook/{}/{}/{}/{}.{}"
if not os.path.exists(contentdir):
os.makedirs(contentdir)
ebpb="220831512%2c584%2c201dohn3vdp9%2c18390%2c0%2c0%2c%2c48+%7b983+1572032732+32D669F6A2D0D3447F9B730529C1F4A33C0883EB%7d";
for pagenr in range(1, maxpages + 1):
cookies = {
'digi4s': '18390%2c1571906836+%7b575+0+7D8ED830D87F0B2B8E7F0F98F4DC2E03C6BF66E8%7d',
'ebpb' : '220831512%2c584%2c201dohn3vdp9%2c18390%2c0%2c0%2c%2c48+%7b983+1572032732+32D669F6A2D0D3447F9B730529C1F4A33C0883EB%7d'
}
headers = {'User-Agent': 'Mozilla/5.0 (Linux; Android 5.1.1; SM-G928X Build/LMY47X) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.83 Mobile Safari/537.36'}
print(pagenr)
if not os.path.exists(pagedirtemplate.format(pagenr)):
os.makedirs(pagedirtemplate.format(pagenr))
if not os.path.exists(subdirtemplate.format(pagenr, "img")):
os.makedirs(subdirtemplate.format(pagenr, "img"))
if not os.path.exists(subdirtemplate.format(pagenr, "shade")):
os.makedirs(subdirtemplate.format(pagenr, "shade"))
response = requests.get(svgurltemplate.format(ebooknr, pagenr, pagenr, "svg"), headers=headers, cookies=cookies)
if (response.status_code != 404):
with open(svgsavetemplate.format(pagenr, pagenr, "svg"), 'wb') as f:
f.write(response.content)
imagecount = 0
with open(svgsavetemplate.format(pagenr, pagenr, "svg")) as f:
contents = f.read()
imagecount = contents.count("image")
for imgnr in range(0, imagecount + 1):
print("+{}".format(imgnr))
response = requests.get(imgurltemplate.format(ebooknr, pagenr, "img", imgnr, "png"), headers=headers, cookies=cookies)
if (response.status_code != 404):
with open(imgsavetemplate.format(pagenr, "img", imgnr, "png"), 'wb') as f:
f.write(response.content)
response = requests.get(imgurltemplate.format(ebooknr, pagenr, "shade", imgnr, "png"), headers=headers, cookies=cookies)
if (response.status_code != 404):
with open(imgsavetemplate.format(pagenr, "shade", imgnr, "png"), 'wb') as f:
f.write(response.content)
subprocess.call(["rsvg-convert", "-f", "pdf", "-o", "{}.pdf".format(pagenr), svgsavetemplate.format(pagenr, pagenr, "svg")])