PDF417 decoder for python

Small python project that can decode, to an extent, 2D PDF417 barcodes from a PNG image.

I wrote this because I wanted to know what was in the barcode in my emailed driver's license and I couldn't find another OSS package to do this.

Some caveats:

• yes it works on PNG images, but it won't work on photos -- they have to be screenshots of a high-quality source (no, screenshot of a photo doesn't count. you can't wish for more wishes)
• only has been tested on a single screenshot I took of my driver's license
• doesn't fully implement the standard -- in particular, doesn't know how to do byte / numeric encodings, just text
• I sourced the spec from wikipedia and this pdf and copied and pasted the lookup tables from OCR -- they may be incomplete or wrong

That said, it has lots of asserts and has worked at least once.

Usage

# help text / flags
./decoder.py -h
# parse a PNG file
./decoder.py barcode-screenshot.png

PDF417 lookup tables

There are some CSVs in here with PDF417 lookup tables.

• cluster-{index}-bs.csv have bsbsbsbs (bar-space) lookups for clusters (0, 3, 6)
• text-codes.csv has the lookup for converting hi & low text codes to ascii chars / state commands

Contributions welcome

In particular:

• tests. can be unit tests of individual routines or link it up to a barcode generator
• more fully decode the standard, or use another library to decode the standard once codewords have been parsed
• bugfixes as always
• link to a better command-line barcode reader so I can archive this repo
• support non-screenshot images (i.e. photos), preferably by using an external library to do the image scanning
• use standard intermediate formats for barcode scans if those exist
• git grep todo