scan.py

#!/usr/bin/env python3
import os
import subprocess
import tempfile
import sys
import shutil
import glob
import requests
from datetime import datetime

OUT_DIR = '/tmp/scan-inbox/'
PAPERLESS_DIR = '/docker/paperless/consume/'
NEXTCLOUD_URL = 'https://cloud.cloud/remote.php/dav/files/admin/Scan-Inbox/'
NEXTCLOUD_USER = 'nextcloud'
NEXTCLOUD_PASSWORD = 'nextcloud'

def upload(path: str, fname: str):
    print("Uploading...")
    headers = {'Content-type': 'application/octet-stream'}
    url = f"{NEXTCLOUD_URL}{fname}"
    auth = (NEXTCLOUD_USER, NEXTCLOUD_PASSWORD)
    r = requests.put(url, data=open(f"{path}{fname}", 'rb'), headers=headers, auth=auth)

os.makedirs(OUT_DIR, exist_ok=True)

with tempfile.TemporaryDirectory() as tmpdir:
    # print(f"Working dir: {tmpdir}")
    os.chdir(tmpdir)

    print("Scanning...")
    try:
        subprocess.run(['/usr/bin/scanimage',
                        '--resolution', '300',
                        '--source', 'ADF Duplex',
                        '--ald=yes',
                        '--swcrop=yes',
                        '--swskip', '10',
                        '--batch=scan_%03d.tif', '--format=tiff',
                        '--mode', 'Color',
                        '--device-name', 'fujitsu:ScanSnap S1500:2853',
                        '-y', '297', '-x', '210',
                        '--page-width', '210', '--page-height', '297',
                        '--sleeptimer', '1'], check=True)
    except subprocess.CalledProcessError as e:
        if e.returncode == 7:
            print("No document in the paper feed, exiting.")
            sys.exit(0)
        else:
            print("Unknown error in scanimage, exiting")
            sys.exit(e.returncode)

    print("Converting to PDF...")
    subprocess.run(['/usr/bin/tiffcp', *glob.glob('scan_*.tif'), 'output.tif'], check=True)
    fname_prefix = f"scan_{datetime.now().strftime('%Y%m%d-%H%M%S')}"
    fname = f"{fname_prefix}_src.pdf"
    output_pdf = f"{OUT_DIR}{fname}"
    with open(output_pdf, 'wb') as f:
        subprocess.run(['/usr/bin/tiff2pdf', 'output.tif', '-j', '-q', '60', '-p', 'A4'], stdout=f, check=True)

    paperless_file = f"{PAPERLESS_DIR}{fname_prefix}.pdf"
    shutil.copy(output_pdf, paperless_file) 
    # print(f"Output PDF: ${output_pdf}")

    final_fname = f"{fname_prefix}.pdf"
    final_pdf = f"{OUT_DIR}{final_fname}"
    subprocess.run(['/usr/bin/ocrmypdf',
        '-r', '-d', '--remove-background', 
        '-l', 'deu+eng',
        output_pdf, final_pdf], check=True)
    
    upload(OUT_DIR, final_fname)

    print("Done, cleaning up. Bye. 👋")