Python Julia Plotter

A script to render Julia sets (the sets of points in the complex plane that don’t escape with the repeated application of a complex rational function) or grids of Julia sets with a comfortable command-line interface, enabled-by-default conversion to .png (magick required in path), and pretty-by-default HTML export.

Table of Contents

• Python Julia Plotter
• Table of Contents
• Why?
  • Low memory requirements
  • Flexibility
  • Grid rendering
• HTML Output Features
• General Usage
• Example Renders
• Arguments and options
  • Quick argument table
  • -f zₙ₊₁, --fn zₙ₊₁ (Default: z^2 + c)
  • -c constant (Default: 0 + 0i)
  • -a aspect, --aspect aspect (Default: 1)
  • -w width, --width width (Default: 500)
  • -i iterations, --iterations iterations (Default: 32)
  • -r c-range, --c-range c-range (Default: 1.5)
  • -n cell count, --cell-count cell count (Default: 1)
  • -e cx cy, --center cx cy (Default: 0 0)
  • -z zoom, --zoom zoom (Default: 1)
  • -s, --silent (Default: Off)
• License

Why?

Low memory requirements

Instead of storing the image data an array before writing it, it’s written one byte at a time. This means that rendering a 65536×65536 image takes up just as much (run-time) memory as rendering a 500×500 image. The raw .ppm output will take up 9 + ⌈log10(w)⌉ + ⌈log10(⌊w/a⌋)⌉ + 3w⌊w/a⌋ bytes (three per pixel plus a header, where w is the image width and a is the aspect ratio).

Flexibility

Any formula is accepted, with fairly versatile equation parsing.

Accepted features:

• Imaginary numbers in the form of (ex. 1i, 3.2i, .3i, etc.)
• Coefficients for variables and functions (2tan(z), 0.3c, 12 pi)
• Implicit parenthesis (ONLY if the variable passed to a function is unmodified; tanz and log10 z work, sin z^2 will parse as (sin z)^2)
• Implicit multiplication ((z - 2)(z + c), z(z - 3), 2c z, sinz tanz)
• Exponentiation (z^2, z^-1)

Grid rendering

julia.py contains a built-in system for rendering grids of Julia sets. Why? Because the same equation that makes this very interesting image:

./julia.py -f "(z-c)(z+2-0.5i)(z+0.5c)(z)" -c "0 + 0.9 i" -i 32 -w 2048 -a 1.0 -e 0.0 0.0 -z 1.0 -g 1.0 -u 30.0

(Note: Most of those arguments are superfluous, but are outputted in case defaults change. The actual rendering command was something much closer to ./julia.py -f "(z-c)(z+2-0.5i)(z+0.5c)(z)" -c "0 + 0.9 i" -w 2048)

Creates this very uninteresting image:

./julia.py -f "(z-c)(z+2-0.5i)(z+0.5c)(z)" -c "0.3 + 0.3 i" -i 32 -w 2048 -a 1.0 -e 0.0 0.0 -z 0.75 -g 1.0 -u 30.0

These two sets are in the same family of equations, but have a different constant c. With the -n cells option, julia.py will render a grid of cells × cells sets with c-values ranging from -range to range in the real axis, and -range·i to range·i in the imaginary axis, where the range is controlled with the -r range option. We may discover which constants are interesting and which are not with a preliminary render:

julia.py -f "(z-c)(z+2-0.5i)(z+0.5c)(z)" -n 11 -i 32 -w 2048 -a 1.0
-e 0 0 -z 0.5 -g 1.0 -u 30

(Aside: Although it might look it at first, the grid is not symmetrical over the real or imaginary axis.)

julia.py will then automatically open the HTML output, allowing us to click on an interesting cell and copy the command-line invocation to create a more detailed render.

HTML Output Features

At a glance, the HTML output contains everything you need to recreate a render, the easy way (by copying and pasting the given command-line invocation) or the hard way (by manually typing in the arguments in the table [do not do this]). There are two other notable features:

1. In renders with multiple cells (i.e. with -n > 1), clicking on a cell in the image will reveal the column, row, c-value, and a command-line invocation to render that cell in a larger image (by default to the same width as the render it came from). Note that this uses an image <map>, and is CSS-only (!) but also pixel-dependent; image-maps don’t resize to fit their respective images, so if the render is larger than your screen, it will be resized down, and Javascript will have to be enabled for the image map to resize correctly (easy/possible thanks to David J. Bradshaw’s imageMapResize.js)
2. In renders with one cell (i.e. with -n 1 or with -n omitted), with Javascript enabled, clicking on the render will update the shown command-line invocation with an updated center value (-e). By clicking on a region of the render you would like to see enlarged and increasing the zoom value (-z), more detailed renders can easily be created.

General Usage

usage: julia.py [-h] [-f zₙ₊₁] [-c constant] [-a aspect] [-w width]
                [-i iterations] [-r c-range] [-e center center]
                [-n cell count] [-z zoom] [-g gradient speed] [-u escape]

Note that if -f is omitted, it will default to z^2 + c, and if -c is omitted, it will default to 0 + 0i, but if both -f and -c are omitted, -f will default to z^2 + c and -c will default to random — this may be unexpected, but results in interesting-and-unpredictable-by-default renders.

Example Renders

./julia.py -f "(z^3)/(((z-2)^5)(z+0.5i)) + c" -c " 1.16667 + 0.166667i" -i 32 -w 2048 -a 1.0 -e 2.0 0.0 -z 0.5 -g 1.0 -u 30

./julia.py -c "0.285 + 0.01i" -z 0.75 -w 2048

Arguments and options

The following is an enumeration of most of the useful and common arguments. There are a few others and this readme may fall slightly out of date — view the most up-to-date and complete list with -h or --help.

Quick argument table

Arg	Long arg	What it controls	Default value
-f	--fn	Render function zₙ₊₁(z, c)	z^2 + c
-c	--constant	Render constant (c in -f)	0 + 0i
-a	--aspect	Image aspect ratio	1.0
-w	--width	Image width	500
-i	--iterations	Fractal iterations	32
-r	--c-range	Range of c-values when -n > 1	1.5
-n	--cell-count	Cell count	1
-e	--center	Render center	0 0
-z	--zoom	Image zoom	1.0
-s	--silent	Info output, shelling out	Off

-f zₙ₊₁, --fn zₙ₊₁ (Default: z^2 + c)

The Julia set's function for iteration. Enter random to generate a random complex rational function (P(z)/Q(z), where P and Q are complex polynomials of maximum degree 3 and 6, respectively).

-c constant (Default: 0 + 0i)

The constant c for the function zₙ₊₁(z, c). Enter random to select a random value for c.

-a aspect, --aspect aspect (Default: 1)

The output image's w/h aspect ratio. Ex.: -a 2 implies an image twice as wide as it is tall.

-w width, --width width (Default: 500)

The output image's width.

-i iterations, --iterations iterations (Default: 32)

The iterations to calculate the set to.

-r c-range, --c-range c-range (Default: 1.5)

The range of c values to use — only relevant if the cell count option is used to render a grid of sets; the c values for each sets will range from (c_r - crange, c_i - crange·i) to (c_r + crange, c_i + crange·i), where c_r and c_i are the real and imaginary components of the constant supplied with -c.

-n cell count, --cell-count cell count (Default: 1)

The number of rows and columns to render. A cell count of 1 will render a single set, and other values will render grids of Julia sets. The different values of c are determined by --c-range or -r. Note that -n 4 implies 4 rows and 4 columns, for a total of 16 actual cells (--row-and-column-count was considered slightly too verbose), perhaps a mildly confusing name.

-e cx cy, --center cx cy (Default: 0 0)

The coordinate the graph is centered around, entered as two floats separated by a space. (Not a comma! No parenthesis! Technically two separate arguments consumed by one option.)

-z zoom, --zoom zoom (Default: 1)

How zoomed in the render is. The distance between the center-point and the top / bottom of the rendered area is 1 / zoom. Larger values of will produce a more zoomed-in image, smaller values (<1) will produce a more zoomed-out image.

-s, --silent (Default: Off)

Don't log info, show progress, convert the .ppm to a .png, or open the file when finished. Equivalent to --no-open --no-convert --no-progress --no-info.

License

MIT / Expat; see license.txt.