jlfricas

Copy of the official repository of the FriCAS computer algebra system embedding a small interface to Julia (experimental work in progress)

Project maintained by gvanuxem Hosted on GitHub Pages — Theme by mattgraham

jlFriCAS

FriCAS CI on x64 Linux (with Julia support - SBCL based)
FriCAS CI on x64 Linux (with Julia support - Clozure CL based)
FriCAS CI on Windows (with Julia support - SBCL based)
FriCAS CI on macOS (with Julia support - SBCL based)

FriCAS is a general purpose computer algebra system (CAS).

In this experimantal work-in-progress repository, a C wrapper using libjulia is embedded in FriCAS to support some Julia specialized operations (for example, hardware architecture optimized BLAS and LAPACK libraries). The build process supports Clozure CL and SBCL, but only Julia 1.10.0 and Julia 1.11.* are supported with SBCL, see Caveats. It must not be considered production-ready. Support of general Julia object is now also supported. An overview of what is added to FriCAS is available here.

A general source of information can also be found in HTML format here. Take into account that this is absolutely not the official documentation even though it is highly based on the official FriCAS web site which can be built from the FriCAS source code (thanks to Ralf Hemmecke and Kurt Pagani for their amazing work).

Capture1

Graphics: Capture2

Building and Installing

For general installation instructions see INSTALL. For general documentation consult https://fricas.github.io.

To build FriCAS with Julia support, the julia executable needs to be available in your PATH, and a simple ./configure --enable-julia && make && sudo make install should do the trick. We require Julia 1.7 or higher. Please see https://julialang.org/downloads/ for instructions on how to obtain Julia for your system. The required Julia packages are (see the Julia package manager):

And optionnally:

As of now with Clozure CL queues and bordeaux-threads are also required. Use installed quicklisp with queues and bordeaux-threads, and at configure time use the --with-quicklisp option, see the quicklisp documentation for how to load it and install them. Another possibility, easier, is to use roswell with added ccl-bin and queues. GitHub actions for Clozure CL use them to build jlFriCAS.

If you want to visualize your data using Julia, small support is provided using Plots and eventually LaTeXStrings Julia packages.

If you want to use jFriCAS i.e. Jupyter support for FriCAS built with SBCL, make sure hunchentoot is installed. On a Debian like system you can add hunchentoot with sudo apt install cl-hunchentoot and issue, for example, ./configure --enable-gmp --enable-julia --enable-hunchentoot.

If you want to build and install the HTML documentation, you need to install Sphinx. On a Debian like system, to add it, issue in a terminal sudo apt install python3-sphinx. After building jlFriCAS, and before the installation, issue in your terminal make htmldoc.

Description

The basic goal of FriCAS is to create a free advanced world-class CAS. In 2007 FriCAS forked from Axiom. Currently the FriCAS algebra library is one of the largest and most advanced free general purpose computer algebra systems -- this gives a good foundation to build on. Additionally, the FriCAS algebra library is written in a high level strongly typed language (Spad), which allows natural expression of mathematical algorithms. This makes FriCAS easier to understand and extend.

FriCAS uses lightweight development methodology. Compared to Axiom, FriCAS is significantly restructured -- it is more portable and fixed several defects. FriCAS removed rather large unused parts (without removing functionality).

Goals

Current development goals:

Caveat: SBCL

Julia support for FriCAS built with SBCL is/was erratic, depending on the Julia version used and the loaded libraries used by Julia. The 1.10.0 version seems to have solved some issues related to memory management interactions with SBCL, but with Julia 1.10.1 and 1.10.2 some problems occur again. Note that with Julia 1.11.* and later, FriCAS seems to work fine again. More work needs to be done in this regard. So, if you use SBCL to build FriCAS, imperatively use a version of Julia that is known to be compatible. Additionnaly, if you use WS domains/packages through MathLink, SBCL is known to crash if you start HyperDoc after loading WS domains/packages. If you need to use HyperDoc, start it first whith the system command )hd.