matamata 0.0.5

Model tournament knockout stages in JSON format and render the schedule in SVG or html table format

matamata

Model a tournament knockout stage in a small JSON "language" and render the schedule in SVG or HTML table format.

matamata also lets a host system map documents representing championship knockout stages onto its own business objects (e.g. a Championship or Cup entity) and persist them apart from any presentation concern — updating results means editing a document, never the code.

Example rendered from examples/libertadores-2026.json:

Quickstart

Requires Python ≥ 3.10. No runtime dependencies. Install from PyPI:

pip install matamata

Render a knockout stage document — plain JSON following the format — to an SVG file, or to an HTML table for small screens:

# the installed command
matamata stage.json -o schedule.svg

# or via the module, writing to stdout
python -m matamata stage.json > schedule.svg

# or as an HTML table
matamata stage.json -o schedule.html

Open the result in a browser to view the schedule. From Python:

from matamata import load_stage, render_svg

svg = render_svg(load_stage("stage.json"))

Ready-to-render worked examples live in examples/. For the latest unreleased commit use pip install git+https://github.com/anibalpacheco/matamata.git; to work on matamata itself, clone the repo and pip install -e ..

Examples

Both examples are rendered from the JSON files in examples/.

The Copa Libertadores example above shows two-legged ties — each leg's goals are shown, shootouts appear in parentheses, and the winner of each tie is emphasized. The first quarterfinal is host-resolved: its legs carry only a ref, so its teams and scores come from get_match (see examples/libertadores_host.py) rather than from the document. Played ties take their team names from the legs; the final is a single match — one leg, so just one goal figure per side — with its winnerof links wiring the advancement tree.

Single matches (knockout-8.json) — one goal figure per side, with shootouts in parentheses. Sides that have not been resolved yet fall back to placeholders such as "Winner SF2":

The format

The document is plain JSON, so any system can store and exchange it natively, and the language is designed so a match can evolve from a placeholder (e.g. "winner of QF1") into a reference to a real match entity — each leg can carry a ref to the real game, resolved dynamically by the host — without changing the language. The advancement tree is laid out deterministically: coordinates are computed directly and the SVG is emitted as a string, with no layout engine and no heavy dependencies.

See the format specification and its JSON Schema. Worked examples live in examples/.

Minimal example:

{
  "tournament": "Copa Libertadores",
  "season": "2026",
  "rounds": [
    {
      "name": "Final",
      "matches": [
        {
          "id": "final",
          "legs": [
            { "team1": "Flamengo", "goals1": 1, "team2": "Nacional", "goals2": 2 }
          ],
          "winner": 2
        }
      ]
    }
  ]
}

Documentation

The manual covers rendering from the CLI and from Python (e.g. a Django view), feeding live data from your own database through KnockoutStage, updating the stored document with apply_results — including a before/after walkthrough of one call — and running the test suite.

Status

Working: the language spec, JSON Schema, Python renderer, CLI and tests are in place, and the package is installable from GitHub.