pytest-r-snapshot

A pytest plugin for snapshot testing against reference outputs produced by R code. Particularly useful for test-driven development (TDD) when porting R packages to Python.

It is designed for a portable workflow:

• Record locally (requires R): run labelled R chunks embedded in your Python tests and write snapshot files.
• Replay everywhere (default; no R required): read committed snapshot files and compare them to Python outputs.

The R chunks live close to the test assertion (copy/paste-able from R scripts or R Markdown), but snapshots are stored as deterministic UTF-8 text files.

Installation

You can install pytest-r-snapshot from PyPI:

pip install pytest-r-snapshot

If your project is managed with uv, add it as a dev dependency:

uv add --dev pytest-r-snapshot

Or install the development version from GitHub:

git clone https://github.com/nanxstats/pytest-r-snapshot.git
cd pytest-r-snapshot
python3 -m pip install -e .

Quick start

Embed a labelled R fenced chunk (commented or in a docstring), then compare your Python output to the recorded snapshot:

def test_summary_matches_r(r_snapshot):
    # ```{r, summary}
    # x <- c(1, 2, 3)
    # summary(x)
    # ```

    actual = my_python_summary(...)
    r_snapshot.assert_match_text(actual, name="summary")

Generate (or update) snapshots locally:

pytest --r-snapshot=record

Commit the generated snapshot files, and run CI with the default replay mode (no R required).

What you get

• A pytest fixture r_snapshot with methods:
  • assert_match_text(actual, name=..., ext=".txt", normalize=...)
  • read_text(name=..., ext=".txt")
  • record_text(name=..., ext=".txt")
  • path_for(name=..., ext=".txt")
• Snapshot modes: replay (default), record, auto.
• Chunk extraction for R Markdown-style fenced chunks inside Python source:
  • Commented fences (chunk lines prefixed with #).
  • Raw chunks in docstrings / multi-line strings.
• Configurable R execution context: Rscript path, environment, working directory, timeout, encoding.