declarativedocs 0.1.2

Declarative Document Format (DDF) — YAML to documents compiler

declarativedocs

Python reference compiler for the Declarative Document Format (DDF). Turns YAML into real documents.

Install

pip install declarativedocs

Supported formats

Format	Schema	Status
Presentation (.pptx)	presentation:	✅ Available
Document (.docx)	document:	🔜 Planned
PDF (.pdf)	pdf:	🔜 Planned

The compiler auto-detects the format from the YAML root key.

Usage

CLI

ddf my-deck.yaml                    # → my-deck.pptx (auto-detected)
ddf my-deck.yaml output.pptx        # → explicit output path

Python API

from declarativedocs import compile

compile("my-deck.yaml", "output.pptx")

Element types (Presentation)

Type	Description
text	Text box with plain text, rich text runs, or bullet lists
shape	Rectangle, rounded rectangle, oval, line
image	Image from file path or base64
chart	Bar, line, pie, doughnut, scatter, radar
table	Data table with styled cells
group	Auto-layout container (row or grid) with card styling
icon	Colored circle with glyph character

Features

• Theme variables — Define colors and fonts once, reference with $name
• Slide masters — Reusable templates with {{placeholder}} data binding
• Auto-layout — row and grid with configurable gap
• Card styling — Fill, radius, shadow, border on group items
• Shadow presets — default, soft, hard, glow, up
• Bullet lists — Proper indentation and tight spacing
• Icons in cards — Automatic positioning above text

Quick example

presentation:
  layout: "16x9"
  theme:
    colors:
      primary: "1E2761"
      white: "FFFFFF"
    fonts:
      heading: Georgia

  slides:
    - background: $primary
      elements:
        - type: text
          x: 1  y: 2  w: 8  h: 1.5
          text: "Hello DDF"
          font: $heading
          size: 44
          color: $white
          bold: true
          align: center

ddf hello.yaml  # → hello.pptx

Schema

See the full DDF specification.

Examples

See the examples/ directory for complete YAML files.

Why DDF?

LLMs currently write 800+ lines of imperative library code to produce a 10-slide deck. DDF describes the same deck in ~200 lines of YAML. That's 3-5x fewer tokens, with significantly fewer bugs.

DDF is a declarative spec — the LLM describes what it wants, the compiler figures out how. The YAML is the contract, the compiler is a swappable implementation detail.

License

Apache 2.0