---
title: "Python Runtime Semantics in pyrt"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Python Runtime Semantics in pyrt}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
```

## Overview

`pyrt` provides the Python 3 runtime semantics required by R packages generated
with `pyport`:

- Reference-semantics containers (`py_list()`, `py_dict()`, `py_set()`, `py_tuple()`)
- C3 MRO classes, multiple inheritance, descriptors (`py_property()`, `py_super()`)
- 0-based negative and step slicing (`py_slice()`, `py_getitem()`, `py_setitem()`)
- Arbitrary-precision integers (`gmp::bigz`) with floor division toward $-\infty$
- Bit-exact CPython 3.12 `random` (`py_random`) and `numpy.random.RandomState`
- Standard library (`math`, `itertools`, `functools`, `collections`, `datetime`, `re`, `json`)
  and `numpy` / `scipy` / `pandas` bridges via `py_stdlib()`

```{r}
library(pyrt)

# Reference-semantics list
xs <- py_list(list(10L, 20L, 30L))
alias <- xs
xs$append(40L)
py_len(alias)

# Bit-exact CPython 3.12 random
rng <- py_random$Random(42L)
rng$random()
```