pydigitize is a simple command-line based tool to scan/archive documents.

It does the following steps:

• Scan a document with any scanner that supports SANE (ADF supported)
• Straightening and cleaning of scanned documents
• Run OCR on PDF so that it becomes searchable
• Generate PDF/A file for archival
• Add keywords to the PDF file

Because you don't want to type the same arguments for every piece of paper that you scan, pydigitize supports profiles: A profile pre-configures settings like the output directory, the resolution, whether to run OCR, additional PDF keywords, etc. You can create a profile for all your invoices for example. Then every time you get an invoice, you scan it with ./scan.py -p invoice, done.

• Python 3.x
• OCRmyPDF 3.x
• libtiff 4.x
• sane 1.x
• unpaper

See ./scan.py --help.

If you want to use profiles, create a profiles.toml file in the current directory.

For every profile you can specify the following parameters:

• path: The output directory
• name: Set a string that will be included in every filename in slugified form
• ocr: Whether to run OCR, straightening and cleanup on the scanned document
• keywords: List of keywords that will be added to PDF Metadata

You can also create sub-profiles. They inherit the settings from the parent.

Example:

[bill]
path = "/home/user/bills/"
name = "bill"
ocr = true
keywords = ["bill"]

[bill.dentist]
name = "dentist"
keywords = ["bill", "dentist"]

[drawing]
path = "/home/user/drawings/"
ocr = false

Then pass the name of the profile to the scan.py command using the -p parameter.

./scan.py -p bill.dentist

You can of course override your parameters:

./scan.py -p bill -n amazon

If you want to scan a specific number of pages, use the -c argument.

pydigitize will prompt you to confirm before scanning every page. This is very useful for example when scanning double-sided documents on a scanner that does not have a duplex unit, or when scanning a document partially in the ADF and partially on the flatbed.

If you don't want manual confirmation, but just want the scanner to scan as fast as it can, use the --nowait argument.