Layout Gallery¶

A visual comparison of the graph-layout algorithms py2max can drive, all applied to the same sample patch (a small two-voice synth). Each image is generated by scripts/gen_layout_gallery.py, which repositions the boxes with a given backend and renders the result through py2max's own SVG exporter, so the layouts are directly comparable.

Regenerate the gallery with:

make gallery
# or: uv run --extra graph python scripts/gen_layout_gallery.py

The layout engines are optional dependencies and are not imported by the core library. Install them with pip install "py2max[graph]" (or uv sync --extra graph). Three backends are supported:

graph-hola - a wrapper for the adaptagrams HOLA algorithm (Human-like Orthogonal Network Layout).
graph-layout - pure-Python constraint-based (COLA) and force-directed / geometric layouts.
ogdf-py - Python bindings for the OGDF C++ library (layered, force-directed, planar).

Using these as py2max layout managers¶

Beyond the gallery script, every engine below is available as a first-class layout manager. Pass layout="graph:<algo>" and call optimize_layout() once the patch is built -- graph layouts need the whole graph, so they apply on optimize rather than as each box is added:

from py2max import Patcher

# HOLA orthogonal layout, applied directly to the patch
p = Patcher("synth.maxpat", layout="graph:hola")
osc = p.add_textbox("cycle~ 440")
flt = p.add_textbox("lores~ 800 0.7")
dac = p.add_textbox("ezdac~")
p.add_line(osc, flt)
p.add_line(flt, dac)
p.optimize_layout()   # graph:* layouts take effect here
p.save()

# ...or an OGDF layered layout on the same patch
p = Patcher("synth.maxpat", layout="graph:ogdf-sugiyama")
# ...add boxes and lines...
p.optimize_layout()

Available selectors (graph:<name>): hola, cola, sugiyama, fruchterman-reingold, kamada-kawai, spectral, circular, shell, ogdf-sugiyama, ogdf-fmmm, ogdf-planarization -- each corresponds to one of the images below. See GraphLayoutManager in py2max/layout/external.py.