aaajiao-viz 0.6.2

ASCII art visualization toolkit for PNG, GIF, and MP4 output.

viz - ASCII Art Visualization Toolkit

PNG/GIF/MP4 ASCII art. Variable resolution (default 1080x1080, up to 3840px), custom palettes, kaomoji, procedural effects, emotion-driven styles. Pure Python 3 + Pillow.

Install

pip install aaajiao-viz

Upgrade an existing install:

pip install -U aaajiao-viz

Verify the CLI is available:

viz --version
viz capabilities --format json

Quick Start

# Generate from emotion
viz generate --emotion euphoria --seed 42

# AI integration via stdin JSON
echo '{"headline":"BTC $95K","emotion":"euphoria","metrics":["ETH: $4.2k"]}' | viz generate

# Animated GIF
echo '{"emotion":"panic","video":true}' | viz generate

# Discover all options
viz capabilities --format json

How It Works

viz is an installed CLI for procedural ASCII image generation. AI decides what to express; VIZ decides how it looks.

emotion/text  -->  VAD vector  -->  grammar  -->  SceneSpec  -->  Engine  -->  output
                   (continuous)     (stochastic)   (full spec)    (auto-      (PNG/GIF/MP4,
                                                                  scaled)     variable res)

Everything is driven by the VAD emotion model (Valence-Arousal-Dominance). 26 named emotions map to points in continuous 3D space. The grammar system samples visual choices weighted by emotion — same emotion + different seed = different output.

Combinatorial space: 17 effects x 86 variants x 9 transforms x 7 postfx x 6 masks x 73 gradients x 8 decorations x bg_fill combos x continuous params = effectively infinite.

Commands

Command	Purpose
generate	Render visualization (variable resolution, PNG/GIF/MP4)
convert	Convert image to ASCII art
capabilities	Output full inventory as JSON for AI discovery

Input (stdin JSON, all fields optional)

{
  "emotion": "euphoria",
  "headline": "BTC $95K",
  "metrics": ["ETH: $4.2k", "SOL: $300"],
  "vocabulary": {"particles": "$€¥₿↑↓", "kaomoji_moods": ["euphoria", "excitement"]}
}

Emotion drives all visual choices. vocabulary overrides specific visual assets (particles, kaomoji moods, decoration chars). Full field reference in docs/ai-integration.md.

Output

{"status": "ok", "results": [{"path": "media/viz_20260203_120000_s42.png", "seed": 42, "format": "png"}], "emotion": "euphoria", "resolution": [1080, 1080]}

Director Mode

Grammar auto-selects everything from emotion. Override any dimension for precise control:

viz generate --emotion euphoria --seed 100 \
  --effect plasma --variant warped \
  --transforms kaleidoscope:segments=6 \
  --postfx vignette:strength=0.5 scanlines:spacing=4 \
  --composition radial_masked

Details in docs/composition.md.

Project Structure

viz.py                          # Single CLI entry
lib/                            # Shared: kaomoji, content, box-drawing chars, vocabulary
procedural/
  engine.py                     # Render orchestrator (auto-scaled buffer -> output)
  effects/                      # 17 pluggable effects + 86 structural variants
  transforms.py                 # 9 domain transforms (mirror, kaleidoscope, tile, ...)
  postfx.py                     # 7 buffer-level post-FX (vignette, scanlines, ...)
  masks.py                      # 6 spatial masks (radial, noise, sdf, ...)
  flexible/
    emotion.py                  # VAD continuous emotion space (26 anchors)
    grammar.py                  # Stochastic visual grammar
    pipeline.py                 # Main pipeline orchestrator
docs/                           # Detailed documentation
skills/viz-ascii-art/           # AgentSkills integration

Documentation

Doc	Content
ai-integration.md	Start here. JSON protocol, capabilities, examples
usage.md	CLI args, all commands
flexible.md	VAD model, grammar, SceneSpec
composition.md	Transforms, masks, PostFX, variants
effects.md	17 effects with params
kaomoji.md	22 moods, 336 faces
box_chars.md	73 gradients, 37 charsets
rendering.md	Engine internals

Development

git clone https://github.com/aaajiao/VIZ.git
cd VIZ
pip install -e ".[dev]"
pytest tests/ -v
python -m build

Published package: aaajiao-viz on PyPI

Runtime dependency: Pillow>=9.0.0. All math is pure Python stdlib (no NumPy). MP4 output requires system FFmpeg.

License

MIT - see LICENSE

Reference

• play.ertdfgcvb.xyz - Inspiration
• play.core - Core reference