Initializing matrix panels to display things onto requires 3 main steps: 

• Create a Geometry object to define the size and shape of the panel(s)
• Create a framebuffer to hold display data
• Create the PioMatter instance.

Geometry

The Geometry object defines the size and shape of the panel(s) that will serve as the display. There 6 arguments that can be passed to configure it.

• width - The total width of the panel(s) in pixels.
• height - The total height of the panel(s) in pixels.
• n_addr_lines - The number of connected address lines. The number of pixels in the shift register is automatically computed from these values. Many panels use 4 as a standard. The 64x64 panels that we stock use 5 address lines. Set this value according to which panels you're using. Note that using 5 addr lines also requires soldering a jumper as described on this guide page.
• serpentine - Controls the arrangement of multiple panels when they are stacked in rows. If it is True, then each row goes in the opposite direction of the previous row. Default is True.
• rotation - controls the orientation of the panel(s). Must be one of the Orientation constants. Default is Orientation.Normal. Over valid values are Orientation.R180, Orientation.CW, and Orientation.CCW
• n_planes - Controls the color depth of the panel. This is separate from the framebuffer layout. Decreasing n_planes can increase FPS at the cost of reduced color fidelity. The default, 10, is the maximum value.
• n_temporal_planes - Controls the temporal dithering. Valid values are 0, 2, and 4. Cannot be higher than n_planes. The default is 0. Using larger values can increase the framerate in some cases experiment with your panels and application.
• map - A pixel/lanes mapping list. For Active3 configurations use simple_multilane_mapper(width, height, n_addr_lines, n_lanes). For single matrix bonnet/HAT this can be omitted.

Framebuffer

Next you need a framebuffer, this is a numpy array that will contain pixel data in the appropriate colorspace. When refreshed the matrix will show the data that is in this framebuffer on the display. Depending on what you're displaying it can be created in a few different ways.

# For displaying an image from PIL, canvas is the PIL.Image instance
framebuffer = np.asarray(canvas) + 0

# For displaying arbitrary data, start with zeros
framebuffer = np.zeros(shape=(geometry.height, geometry.width, 3), dtype=np.uint8)

# For displaying an image from PIL, canvas is the PIL.Image instance
framebuffer = np.asarray(canvas) + 0

# For displaying arbitrary data, start with zeros
framebuffer = np.zeros(shape=(geometry.height, geometry.width, 3), dtype=np.uint8)

PioMatter

Now the PioMatter instance can be initialized. Aside from the Geometry, and Framebuffer created above, it also accepts arguments for colorspace and pinout.

• colorspace - Controls the colorspace that will be used for data to be displayed.
  It must be one of the Colorspace constants. Which to use depends on what data 
  your displaying and how it is processed before copying into the framebuffer. Many of the examples use Colorspace.RGB888Packed. Other valid values are Colorspace.RGB565 and Colorspace.RGB888
• pinout - Defines which pins the panels are wired to. Different pinouts can support different hardware breakouts and panels with different color order. The value must be one of the Pinout constants Pinout.AdafruitMatrixBonnet, Pinout.AdafruitMatrixBonnetBGR, Pinout.AdafruitMatrixHat, or Pinout.AdafruitMatrixHatBGR. For the active3/Triple Matrix Bonnet use Pinout.Active3 or Pinout.Active3BGR
• framebuffer - A numpy.array buffer as noted above.
• geometry - A Geometry instance as noted above.

Example Initialization Code

Single matrix Bonnet/HAT:

geometry = piomatter.Geometry(width=64, height=32, n_addr_lines=4, rotation=piomatter.Orientation.Normal)

matrix_framebuffer = np.zeros(shape=(geometry.height, geometry.width, 3), dtype=np.uint8)

matrix = piomatter.PioMatter(colorspace=piomatter.Colorspace.RGB888Packed, pinout=piomatter.Pinout.AdafruitMatrixBonnet, framebuffer=matrix_framebuffer, geometry=geometry)

geometry = piomatter.Geometry(width=64, height=32, n_addr_lines=4, rotation=piomatter.Orientation.Normal)

matrix_framebuffer = np.zeros(shape=(geometry.height, geometry.width, 3), dtype=np.uint8)

matrix = piomatter.PioMatter(colorspace=piomatter.Colorspace.RGB888Packed, pinout=piomatter.Pinout.AdafruitMatrixBonnet, framebuffer=matrix_framebuffer, geometry=geometry)

Triple Matrix Bonnet:

pixelmap = simple_multilane_mapper(width, height, n_addr_lines, n_lanes)
geometry = piomatter.Geometry(width=width, height=height, n_addr_lines=n_addr_lines, n_planes=10, n_temporal_planes=4, map=pixelmap, n_lanes=n_lanes)
framebuffer = np.asarray(canvas) + 0  # Make a mutable copy
matrix = piomatter.PioMatter(colorspace=piomatter.Colorspace.RGB888Packed,
                             pinout=piomatter.Pinout.Active3,
                             framebuffer=framebuffer,
                             geometry=geometry)

pixelmap = simple_multilane_mapper(width, height, n_addr_lines, n_lanes)
geometry = piomatter.Geometry(width=width, height=height, n_addr_lines=n_addr_lines, n_planes=10, n_temporal_planes=4, map=pixelmap, n_lanes=n_lanes)
framebuffer = np.asarray(canvas) + 0  # Make a mutable copy
matrix = piomatter.PioMatter(colorspace=piomatter.Colorspace.RGB888Packed,
                             pinout=piomatter.Pinout.Active3,
                             framebuffer=framebuffer,
                             geometry=geometry)