BenchmarksΒΆ
The repository includes a benchmark comparing hyperscribe with
Jinja, Mako, Cheetah3, Airium, Yattag, dominate, Ludic, Hyperscript, Tagflow,
and xml.etree.ElementTree.
Each library renders the same article list
with conditional badges, optional authors, tag loops, a reusable component,
comments, void elements and optional or boolean attributes.
uv sync --group benchmarks
uv run --group benchmarks pytest benchmarks
uv run --group benchmarks pytest benchmarks -k "Jinja or Hyperscribe" --items 1000
The benchmarks use pytest-benchmark,
so its options for sorting, comparing, and saving results all apply.
The dependencies are in the optional benchmarks dependency group
and require Python 3.14.
Every renderer is first checked to produce the same document.
The report then shows the timing statistics for each library,
followed by a table of the peak traced Python memory and output size of one render.
See benchmarks/README.md in the repository for the methodology.