PyDAS provides a means for Python code to utilize several notable Fortran-based differential algebraic system solvers from Python code. The solvers made available -- DASSL, DASPK, and DASKR -- are all publicly-available from Netlib, and are distributed with PyDAS. PyDAS provides a Python extension type for each solver, which in turn provides a Pythonic means of setting the solver options, providing residual and jacobian functions, and running the solver.
The DASSL, DASPK, and DASKR solvers are all substantially more robust than VODE, the solver used within the ODE solver functionality provided by SciPy.
Copyright (c) 2010 by Joshua W. Allen ([email protected]).
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the 'Software'), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
PyDAS has been tested on Python versions 2.5 and 2.6. It may or may not work for other Python versions.
There are several Python dependencies that you must install before installing PyDAS:
- Python (versions 2.5.x and 2.6.x are known to work)
- NumPy (version 1.3.0 or later is recommended)
- Cython (version 0.12.1 or later is recommended)
In addition, you will also need a Fortran compiler and a C compiler that
produce object files that can interoperate. The gfortran and gcc
compiles from the GNU Compiler Collection are known to work. On Windows the
MinGW compiler collection provides these compilers.
The code for the differential algebraic system solvers DASSL, DASPK, and DASKR has been provided with the PyDAS package. The licenses for these solvers is different than that of the PyDAS wrapper code. You are responsible for knowing and abiding by all licenses associated with each solver as well as with PyDAS as a whole.
Note
Currently only the DASSL solver has been wrapped. The installation scripts therefore only build and install the DASSL wrapper by default.
The provided batch scripts will compile all of the solvers and the PyDAS wrapper code. These scripts presume that you have the 32-bit version of the MinGW C and Fortran compilers installed. Once you have run the batch script, you can install PyDAS into your Python packages if you desire by running the following command from the base package directory:
> python setup.py install
A Makefile has been provided that can be used to compile all of the solvers and the PyDAS wrapper code. To use, invoke the following command from the base package directory:
$ make
This command will build PyDAS in-place, rather than installing it to your
Python package directory. If you wish to formall install PyDAS, run the
following command from the base package directory after the make command
(you may need root privileges for this):
$ python setup.py install
You may wish to write a file make.inc that sets certain variables used by the Makefiles (e.g. the Fortran compiler). An example of such a file, make.inc.example, has been provided.
Homebrew (http://brew.sh) is an easy way to get gfortran:
$ brew install gcc
But your system may still not be able to find the correct libgfortran.a library file (see Homebrew/legacy-homebrew#8539 ). Also, there are some problems linking with clang, so you need to make it link with gcc. This one-liner should build and install, assuming you have NumPy, Cython, etc. all set up:
$ LIBRARY_PATH=/usr/local/lib/gcc LDSHARED='gcc -bundle -undefined dynamic_lookup -arch x86_64' make F77=gfortran install
Or perhaps, with a newer version of Homebrew / Python / gfortran / NumPy / Cython, it will be a simple:
$ LIBRARY_PATH=/usr/local/Cellar/gfortran/4.8.0/gfortran/lib make F77=gfortran
It seems to keep on changing. If you have difficulty, check the issue tracker, and if you solve your difficulty, please share your successful approach.