OctSymPy

An implementation of a symbolic toolbox using SymPy.

[https://github.com/cbm755/octsympy]

Goals

Feature parity with the other symbolic toolboxes.

Status

"Beta" quality at best! Contributions welcome.

How to Install

1. The only dependencies are Python and SymPy. Consult the SymPy website (e.g., yum install sympy on Fedora).

2. Download the latest release, e.g., octsympy-0.1.2.tar.gz.

3. Run Octave and change to the folder containing the downloaded file.

4. At Octave prompt, type pkg install octsympy-0.1.2.tar.gz.

5. At Octave prompt, type pkg load octsympy.

6. At Octave prompt, type syms x, then f = (sin(x/2))^3, diff(f,x), etc.

How to Install on Windows

1. Get Octave for Windows.

2. Try the octsympy-windows-0.1.2.zip package. Follow the last three steps above using this file instead.

The octsympy-windows package should have no dependencies other than Octave itself (it includes SymPy and a Python interpreter.)

Alternatively, you can install Python and SymPy yourself and use the standard octsympy-0.1.2.zip package.

If you encounter any difficulties (even minor ones) please read and if possible help us improve the wiki page on Windows Installation.

How to Install on Matlab

Although OctSymPy is designed for GNU Octave, it will work with Matlab. Currently only the slower system()-based communication is available.

1. Download the latest release, e.g., octsympy-matlab-0.1.2.tar.gz.

2. Unzip is somewhere and add it to your Matlab Path.

The .m files for Matlab have been reformatted for Matlab comment conventions, but are otherwise the same as the Octave source.

How to Help

We have a list of things to work on tagged help wanted. Some of these should be quite easy to fix and would be a great way to get involved. Come join us!

How to hack on the code:

1. Clone the repo with git (preferred, but you can use the "Download ZIP" instead of you want).

2. Go to octsympy/src/ and type "make". (You only need to do this again if you change the inst/private/python_header.py or various autogeneration scripts in src/.)

3. Run Octave (or Matlab) in octsympy/inst/.

Implementation

Python code is generated to do the actual work. Each sym objects keep a text field for display purposes and a string (a SymPy srepr). The objects are communicated between Python and Octave by passing the srepr's back-and-forth. Currently pure m-file (and Python) implementation, no code to be compiled.

Communication

We have two IPC mechanisms between Octave and Python. One option is calling "system()". The other uses "popen2()". Others could be implemented.

Related Projects

• There was a previous "symbolic" package in Octave Forge based on GiNaC. Its history has now been merged into octsympy.

• "SymPy CAS" by Jonathan Lister. Calls SymPy commands using system().