smix-annotate

Screenshot annotation primitives + PNG compression for smix. See RFC docs/ai-guide/rfc-v0.3.0-annotated-screenshots.md.

v0.3.0 scope

• 5 primitives: circle, arrow, text, box, line
• 3 compression presets: fast, balanced (default), aggressive (via oxipng)
• Named + hex + rgba color palette, plus semantic colors (EXPECTED, ACTUAL, HINT, ERROR, SUCCESS)
• Position: absolute pixel + normalized 0..1
• Text rendering via ab_glyph (font supplied by caller — v0.3.5 bundles Inter + Noto Sans SC)

Library usage

use smix_annotate::{Annotator, Annotation, Color, Position, Compression};

let png = std::fs::read("input.png")?;
let out = Annotator::new(&png)?
    .add(Annotation::circle(Position::pixel(100, 100))
        .color(Color::RED)
        .radius(40))
    .compression(Compression::BALANCED)
    .render()?;
std::fs::write("output.png", out)?;

Non-goals (deferred to v0.3.5)

• Bundled fonts (Inter + Noto Sans SC)
• Selector-relative position resolution (smix-adapter-maestro runtime handles it)
• Auto-annotate for --debug-output fail-step screenshots