LibPARI.jl

A Julia wrapper for the PARI/GP number-theory library.

LibPARI embeds the PARI/GP C library in a Julia process and exposes it two ways: over 1200 functions generated directly from PARI's own machine-readable function database (the LibPARI.PARI submodule), and a hand-written core — a Gen value type wrapping every PARI object, idiomatic numeric operators, conversions to and from Julia numbers, a GP expression evaluator, and safe library lifecycle and error handling.

LibPARI.jl is an independent, community-maintained Julia interface to the PARI/GP computer algebra system. This project is not affiliated with, endorsed by, or maintained by the PARI Group or the PARI/GP development team. PARI/GP is developed by the PARI Group and distributed under the GNU General Public License. This wrapper merely calls into the official PARI library (distributed here via PARI_jll) and does not modify it.

• Bugs or feature requests concerning this Julia wrapper should be reported to this repository's issue tracker.

• Bugs or feature requests concerning PARI/GP itself should be directed to the official PARI/GP bug tracker or the pari-users mailing list.

Any errors in the Julia interface, the high-level API, or its documentation are the responsibility of this project's maintainers, not of the PARI Group.

Substantial portions of this code were generated with AI assistance and reviewed by the maintainer. See [topic ai-generated-code] for details.

Documentation

Full documentation — introduction, getting-started guide, and API reference — is published at https://s-celles.github.io/LibPARI.jl.

Features

• Gen <: Number — PARI objects are first-class Julia numbers, with +, -, *, /, ^, and ==, and mixed Gen/Julia arithmetic.
• Exact arbitrary-size arithmetic — PARI bignum integers, rationals, and reals, with conversions back to BigInt and fixed-width integer types.
• gp_eval — evaluate any GP-language expression, the escape hatch to every PARI capability.
• A safe C boundary — every PARI error becomes a catchable PariError; PARI's stack stays leak-free and each Gen frees its own storage.
• Thread-safe — every libpari call is marshalled onto one dedicated worker task, so LibPARI is correct under multi-threaded Julia.
• Type-stable and precompiled — built to SciML inference standards.

Installation

using Pkg
Pkg.add("LibPARI")

The native PARI library is supplied by PARI_jll, a registered binary build of PARI 2.17.3 for Linux, macOS, and Windows. Installation is fully automatic — no system PARI is required.

LibPARI requires Julia 1.10 (the long-term-support release) or later.

Quick example

using LibPARI

a = LibPARI.Gen(42)
a * a + LibPARI.Gen(1)                  # 1765
LibPARI.Gen(2)^100                      # 1267650600228229401496703205376
BigInt(LibPARI.Gen(2)^64)               # 18446744073709551616

LibPARI.PARI.nextprime(LibPARI.Gen(1000))   # 1009
LibPARI.gp_eval("sum(k = 1, 100, k^2)")     # 338350

See the getting-started guide for a fuller tour.

Development

• Tests: julia --project=. -e 'using Pkg; Pkg.test()'
• Documentation: julia --project=docs docs/make.jl (after Pkg.develop(path = ".") + Pkg.instantiate() in the docs/ environment).
• Formatting: 80-column margin, enforced by JuliaFormatter and pre-commit.
• The project follows Semantic Versioning; notable changes are recorded in CHANGELOG.md. The implementation plan is in ROADMAP.md.