Banner Generator

Procedural SVG banner with chaos-attractor backgrounds, SVGO optimization, and automatic dark-mode derivation.

Module

scripts/banner.py — ~1878 lines. Heavy dependency on svgwrite and numpy.

scripts/banner_patterns.py — PatternType enum extracted for zero-dep import.

Entry point

from scripts.banner import generate_banner, BannerConfig

cfg = BannerConfig(
    title="Wyatt Walsh",
    subtitle="Software Engineer & AI Enthusiast",
    output_path=Path(".github/assets/img/banner.svg"),
    width=1200,
    height=630,
    optimize_with_svgo=True,
    dark_mode=False,
)
generate_banner(cfg=cfg)

generate_banner is called twice by the CLI: once for dark_mode=False (light), once for dark_mode=True (dark). The dark output path is derived automatically by inserting -dark before the extension.

BannerConfig fields

Field	Type	Default	Notes
`title`	`str`	`"Wyatt Walsh"`	Main heading
`subtitle`	`str`	`"Software Engineer & AI Enthusiast"`
`output_path`	`Path`	`.github/assets/img/banner.svg`
`width`	`int`	`1200`	Canvas width (px)
`height`	`int`	`630`	Canvas height (px)
`optimize_with_svgo`	`bool`	`True`	Run SVGO after writing
`dark_mode`	`bool`	`False`	Toggle dark palette
`pattern_density`	`float`	`2.0`	Background pattern density
`layer_count`	`int`	`4`	Attractor layer count
`make_responsive`	`bool`	`True`	Add CSS viewBox responsiveness

PatternType

Defined in scripts/banner_patterns.py. Live variants:

Value	Description
`LORENZ`	Lorenz strange attractor
`NEURAL`	Neural-network node graph
`FLOW`	Perlin flow field
`MICRO`	Micro-geometry lattice
`AIZAWA`	Aizawa attractor

Dead variants (draw functions not yet implemented): REACTION, CLIFFORD, FLAME, PDJ, IKEDA.

The active pattern is chosen automatically based on the metrics hash. All five live patterns may appear in generated output.

SVGO optimization

If optimize_with_svgo=True and svgo is on PATH, the generated SVG is piped through SVGO to reduce file size. Falls back silently if SVGO is unavailable.

Output files

File	Description
`banner.svg`	Light-mode banner
`banner-dark.svg`	Dark-mode banner (auto-derived path)

Generative Art

Lorenz/neural/flow-field SVGs that share PatternType with the banner

Configuration

BannerSettings reference in config.yaml