pdraw.py: Generate Deterministic, Procedural Artwork from Arbitrary Text

pdraw allows users to generate line art from arbitrary text. It has no dependencies other than the Python standard library, and works with Python 3.10+ (3.7+ should work but are untested).

Thanks to Numberphile for the idea for this script.

Installation

Either copy the pdraw.py file or git clone the repository.

$ git clone [email protected]:Sonictherocketman/pdraw.git

Then simply run the script.

$ python3 pdraw.py -i a-file.txt [-e for arbitrary text]

IMPORTANT NOTICE: I have not confirmed this, but it looks like versions of Python installed via Homebrew are not suitable for this script as they do not come with the necessary tkinter libraries bundled in. I had to install Python direct from python.org

Testing

There are no tests. ¯_(ツ)_/¯

Usage Notes

Encoding Text as Base 10 Numbers

pdraw natively supports drawing strings of base 10 digits. However, pdraw can also sketch arbitrary text. To enable the drawing of arbitrary text, add the --encode flag to your command. This will instruct pdraw to first convert each character into its base 10 form (using a silly conversion that could probably be a lot more efficient).

Faster Drawing

pdraw is meant to be interactive, and so it renders its output fairly frequently. To speed up the drawing process, add the -r/--refresh-rate flag with a larger cycle count.

Changing Draw Distance & Angles

pdraw can also draw with arbitrary draw distances and angles to suit your needs. The -d/--distance flag allows you to set a distance multiplier for each line, and the -a/--angle flag allows you to set the rotation angle between each line. The default is 90 which draws squares. Use -a 60 for hexagons, -a 120 for triangles, and -a 160 for cool sharp ridges.

Saving Drawings

pdraw can save drawings as .eps files (which can be opened in Preview.app for macOS as well as other PDF viewers). Those files can then be converted to PNG/JPG/etc. Simply supply the -o/--output flag with a path to the desired file.

Optionally, you can close pdraw when your drawing is complete and saved by supplying the -c/--close flag.

Example Usage

Below are some quick recipes to get you started.

Drawing Digits of Pi (in hexagons)

You can download the first billion digits of Pi here and start plotting them. Make sure to give pdraw an iteration count though, otherwise it will attempt to draw the entirety of those billion digits, which would probably not work and take forever.

$ ./pdraw -i digits-of-pi.txt

Note: pdraw also accepts an --offset parameter that tells pdraw how many bytes into the file to start reading. Use this to plot various sections of Pi (or other longer files)!

Drawing pdraw Itself!

Yup. pdraw can of course draw itself. Code is just text after all.

$ ./pdraw -i pdraw.py --encode

Visualizing Song Lyrics

One thing I've been having fun with is graphing arbitrary song lyrics. I even printed some for family members of their favorite songs.

$ ./pdraw -i peaceful-easy-feeling.txt --encode -a 120

Visualizing Randomness

pdraw can draw any digital data, and it can read from stdout! It cannot however deal with binary data, so you'll need to convert your random bytes into a string. Thankfully, we can do that easily on the command-line.

$ cat /dev/urandom | base64 | ./pdraw -n 5000 --encode

Note: Make sure to give pdraw an iteration count or it will draw forever!

About this Script

This script is based on a Numberphile Video about using procedural techniques to sketch the digits of Pi. Originally, I aimed to simply reproduce their work and play with the artwork I could generate from it. Eventually, that evolved into building a fully-fledged command-line application that could graph arbitrary digits, and since all data on a computer can be converted to numbers, I had effectively built a way to draw anything!

Thanks to Numberphile for putting out amazing videos and for inspiring me to write this script.

License

This script is available under the terms of the MIT License. See license.txt for more information.

Copyright © 2022 Brian Schrader