CBMPy: Installation Guide

Introduction

Support

PySCeS CBMPy is Open Source software released under the GNU GPL 3 licence (included with the source code) and is in constant development. All the latest downloads, documentation and development information is available on SourceForge: http://cbmpy.sourceforge.net

CBMPy is developed in the Department of Systems Bioinformatiocs at the Vrije Universiteit Amsterdam, as part of the BeBasic Metatoolkit project, by Brett Olivier (bgoli@users.sourceforge.net)

Python standard library modules

CBMPy is developed and tested on Python 2.7.x. but should also work on Python 3. The following Python Standard Library modules are used in CBMPy and should be available as part of any CPython distribution and not require additional installation:

'cPickle', 'cStringIO', 'cgi', 'copy', 'gc', 'itertools', 'locale', 'math',
'multiprocessing', 'os', 'pprint', 'random', 're', 'shutil', 'subprocess',
'time', 'urllib2', 'webbrowser', 'xml'

Required libraries (Python bindings)

Besides those mentioned above, the following packages are required for CBMPy’s core functionality. Note that it is possible to install CBMPy using only numpy but that only very limited subset of functionality is then available. CBMPy is primarily developed on Microsoft Windows and Ubuntu Linux and where possible the package name is provided such that can be used with the software center or package manager sudo apt-get install <package> (please see the man pages for sudo and apt-get if you don’t know what this). A comprehensive list of modules are listed at the end of this document. In the case of external C/C++ libraries the Python bindings should be installed (e.g. libSBML). Many of these are available in batteries included Python distributions and via the pip installer.

Overview

CBMPy has been designed to be used as a framework and can be used in different contexts. Here are the minimal requirements for each role. The fastest way to install the latest official version of CBMPy is using PyPI: https://pypi.python.org/pypi/cbmpy while development source, binaries, tools and utilities are available from SourceForge: https://sourceforge.net/projects/cbmpy/files

PyPI

Try one of:

pip install cbmpy
easy_install cbmpy

Minimal

Read, write and convert model files as well as view, create and edit model components and annotation.

  • numpy
  • libSBML (Python bindings)
  • biopython

Normal

In addition to the minimal requirements a linear optimizer is required

  • Optimization libraries (one or more of):
  • matplotlib
  • sympy
  • xlwt
  • xlrd

Full

This includes graphical interface support

  • pyqt4
  • wxPython
  • suds
  • scipy
  • h5py
  • networkx

User tools

These are highly recommended user tools but are not required for using CBMPy.

  • iPython
  • iPython-notebook
  • SCiTE

Installing on Ubuntu Linux

Python 2.7 (full install)

First we create a scientific Python workbench using the Ubuntu package manager:

sudo apt-get install python-dev python-numpy python-scipy python-matplotlib  python-pip
sudo apt-get install python-sympy python-suds python-xlrd python-xlwt python-h5py
sudo apt-get install python-biopython python-wxgtk2.8 python-qt4
sudo apt-get install ipython ipython-notebook

libSBML

Installing libSBML is now easy using PiP and PyPI, first we need some dependencies:

sudo apt-get install libxml2 libxml2-dev
sudo apt-get install zlib1g zlib1g-dev
sudo apt-get install bzip2 libbz2-dev

Then we can install the latest and greatest with:

sudo pip install python-libsbml

glpk/python-glpk

Currently, CBMPy still requires GLPK version 4.47 to work with glpk-0.3 so it is important not to install the latest using the ubuntu package manager but rather use these instructions. First download the GLPK source and bindings:

https://sourceforge.net/projects/cbmpy/files/tools/glpk/glpk-4.47.tar.gz
https://sourceforge.net/projects/cbmpy/files/tools/glpk/glpk-0.3.tar.bz2

and then install the the following dependency:

sudo apt-get install libgmp-dev

Unpack the glpk-4.47 source and make it your current directory:

cd GLPK source (e.g. glpk-4.47)

Build and install GLPK:

./configure --with-gmp
make
make check
sudo make install
sudo ldconfig

Change to the python-glpk source (glpk-0.3) and make it the current directory:

cd to python-glpk source (glpk-0.3):

Then build and install it with the following commands:

make
sudo make install

Finally, install CBMPy either using PyPI:

pip install cbmpy
     or
easy_install cbmpy

Or download a source file from either:

https://pypi.python.org/pypi/cbmpy
https://sourceforge.net/projects/cbmpy/files

and install manually:

python setup.py build sdist
sudo python setup.py install

Installing on Microsoft Windows

For the modeller that does not want to customize his installation and install all of the individual packages by hand there are some batteries included Python distributions which have many (if not all) of the required packages required.

CBMPy is developed using a 64 bit version of Anaconda Python 2.7 (http://continuum.io) has also been installed on 32 bit Python(x,y) and Enthought Python Distribution (EPD) http://www.enthought.com

Anaconda

Most of the required packages can be installed using the conda package manager for example:

conda install sympy pyqt4

or using pip (libSBML) or setuptools

Python(x,y)

Select the following packages from the Python branch of the Python(x,y) installation directory:

  • WxPython
  • Sympy
  • NetworkX
  • xlrd
  • xlwt
  • h5py
  • wxPython
  • PyQT4

Installing CBMPy

There are two ways to install CBMPy either download the latest release as source bundle or binary from http://cbmpy.sourceforge.net and unzip or execute from a a temporary directory (recommended). Or, if you want the latest (greatest and potentially broken) version grab the latest revision from the the CBMPy Subversion repository:

svn co http://sourceforge.net/p/cbmpy/code/HEAD/tree/trunk/cbmpy cbmpy

In both cases you should should now have a directory that contains a file setup.py which can install by simply typing the following into a Windows shell (command line):

python setup.py build
python setup.py install

Installing libSBML with Python bindings

It is highly recommended to install libSBML which CBMPy uses to provide support for the Systems Biology Markup Language (SBML). First go to the libSBML download page http://sbml.org/Software/libSBML page follow the Download libSBML –> Stable –> Windows –> 32bit path and download libSBML (e.g. libSBML-5.10.0-win-x86.exe). The latest stable version can be found at http://sbml.org/Software/libSBML

Run the installer and make sure you select the Python Bindings during installation or install the appropriate Python bindings that match your Python(x,y) version directly e.g. (libSBML-5.10.0-win-py2.7-x86.exe)

Optmization (1): IBM cplex optimization studio (Academic)

If you have access to the the IBM CPLEX solver. It is a a good idea to use the latest available version. Again choose the appropriate 32 or 64 bit version and an installation path that suites your setup.

  • Run cplex_studio126.win-x86-32.exe
  • Select English language and accept licence
  • Set “Program” install directory to C:\ILOG\CPLEX_Studio126
  • Allow default associations to be set and PATH update

Once installation is complete we need to install the Python bindings

  • Open a terminal
  • Execute cd c:\\ILOG\\CPLEX_Studio126\\cplex\\python\\x86_win32
  • Execute python setup.py install

Optmization (2): GLPK

CBMPy 0.7.4 includes support for the free, Open Source GLPK solver. This allows access to CBMPy’s LP functionality (MILP’s requires CPLEX). A port of PyGLPK 0.3 is maintained by the OpenCOBRA project which is mirrored here:

Select the binary or source distribution you require and either execute the binary:

  • Execute glpk-0.3.win32-py2.7.exe

Testing your new installation

If everything has gone according to plan you can test your installation:

  • Open a terminal
  • Execute ipython
  • In ipython shell, execute import numpy, h5py, xlrd, xlwt

No import errors should occur.

  • Execute import libsbml
  • Execute libsbml.LIBSBML_VERSION_STRING

A successful test should return (for example):

In : libsbml.LIBSBML_VERSION_STRING
Out: '51000'
  • Execute import cbmpy as cbm

This should return:

In [1]: import cbmpy as cbm

*******************************************************************
* Welcome to CBMPy (0.7.4) - PySCeS Constraint Based Modelling    *
*                http://cbmpy.sourceforge.net                     *
* Copyright(C) Brett G. Olivier 2010 - 2015                       *
* Dept. of Systems Bioinformatics                                 *
* Vrije Universiteit Amsterdam, Amsterdam, The Netherlands        *
* CBMPy is distributed under the GNU GPL v 3.0 licence, see       *
* LICENCE (supplied with this release) for details                *
*******************************************************************

Exit ipython with CTRL-D

If you installed CPLEX then try:

  • Open a terminal
  • Execute ipython
  • Execute import cplex
  • Execute lp = cplex.Cplex()
  • Execute lp.solve()

A succesful test should return:

In : lp.solve()
Tried aggregator 1 time.
No LP presolve or aggregator reductions.
Presolve time =    0.00 sec.

Exit ipython with CTRL-D

If you installed GLPK then try:

  • Open a terminal
  • Execute ipython
  • Execute import glpk
  • Execute lp = glpk.LPX()

A succesful test should return:

In : glpk.LPX()
<glpk.LPX 0-by-0 at 0x036C24C8>

Exit ipython with CTRL-D

Install CBMPy (http://cbmpy.sourceforge.net)

Download the latest version of CBMPy

  • Run cbmpy-0.7.x.win32.exe (or newer for 32 bit Windows)
  • Run cbmpy-0.7.x.amd64.exe (or newer for 64 bit Windows)

Test installation:

  • Open a terminal
  • Execute ipython
  • Execute import cbmpy as cbm

This should return:

In [1]: import cbmpy as cbm

*****
Using GLPK
*****

WX GUI tools available.
Qt4 GUI tools available

CBMPy environment
******************
Revision: r346


*******************************************************************
* Welcome to CBMPy (0.7.4) - PySCeS Constraint Based Modelling    *
*                http://cbmpy.sourceforge.net                     *
* Copyright(C) Brett G. Olivier 2010 - 2015                       *
* Dept. of Systems Bioinformatics                                 *
* Vrije Universiteit Amsterdam, Amsterdam, The Netherlands        *
* CBMPy is distributed under the GNU GPL v 3.0 licence, see       *
* LICENCE (supplied with this release) for details                *
*******************************************************************

Exit ipython with CTRL-D

Linux: Ubuntu

On Linux many of the base dependencies are available as packages or from the Python Cheeseshop (http://pypi.python.org/pypi). For libSBML, CPLEX and/or GLPK please see the Generic installation on Microsoft Windows (XP, 7, 2008) for more details. For example using Ubuntu the base dependencies can be easily installed (depending on what functionality is required). If you don’t know what these packages are please look them up before installing.

Required:

sudo apt-get install python-dev python-numpy

- libSBML for SBML support.

Please see http://sbml.org/Software/libSBML or try the following. Depending on your configurationyou need to install libxml2, bzip2 and their associated “dev” packages:

apt-get install libxml2 libxml2-dev
apt-get install zlib1g zlib1g-dev
apt-get install bzip2 libbz2-dev

easy_install pip

# for standard libSBML
pip install python-libsbml

# for "experimental" libSBML (for FBC V2 and Groups support)
pip install python-libsbml-experimental

Please note that due to changes in the GLPK API the current version of PyGLPK (0.3) only supports GLPK up until version 4.47. If your system has a newer version of GLPK then the current workaround is to uninstall the newer version and compile 4.47 from source (also available from the above directory). Dependencies are standard Linux build tools and GMP etc:

tar xzf glpk-4.47.tar.gz
cd glpk-4.47
./configure --with-gmp
make
make check
sudo make install

Graphical interfaces (highly recommended):

sudo apt-get install python-wxgtk2.8 python-qt4 python-matplotlib

Extended IO (highly recommended):

sudo apt-get install python-xlrd python-xlwt python-sympy

Web services and database:

sudo apt-get install python-suds python-pysqlite2

Advanced functionality:

sudo apt-get install python-scipy python-h5py python-networkx

User tools (highly recommended):

sudo apt-get install ipython ipython-notebook scite

Linux: Ubuntu 14.04

Python2

First we create a scientific Python workbench:

sudo apt-get install python-dev python-numpy python-scipy
sudo apt-get install python-matplotlib  python-pip
sudo apt-get install python-sympy python-suds python-xlrd
sudo apt-get install python-xlwt python-h5py
sudo apt-get install python-wxgtk2.8 python-qt4
sudo apt-get install ipython ipython-notebook

libSBML

Installing libSBML is now easy using Pip:

sudo apt-get install libxml2 libxml2-dev
sudo apt-get install zlib1g zlib1g-dev
sudo apt-get install bzip2 libbz2-dev

sudo pip install python-libsbml

glpk/python-glpk

GLPK needs to be version 4.47 to work with glpk-0.3:

sudo apt-get install libgmp-dev

cd GLPK source (e.g. glpk-4.47):

./configure --with-gmp
make
make check
sudo make install
sudo ldconfig

cd to python-glpk source (glpk-0.3):

make
sudo make install

CBMPy

Finally, install CBMPy:

python setup.py build sdist
sudo python setup.py install

Installing PyscesMarinerCBM

This will install PySCeS Mariner that adds SOAP web-services capability to CBMPy. First unpack pyscesmariner-0.7.7.zip and install the cherrypy webserver:

sudo apt-get install python-cherrypy

Install soaplib

cd <pysces_cbm_mariner>/misc:

tar -xf soaplib-0.8.1.tar.gz
cd soaplib-0.8.1
python setup.py build sdist
sudo python setup.py install

Install Mariner

cd <pysces_cbm_mariner> and set mariner configuration (not needed for Ubuntu, Windows or if the server does not read SBML):

sudo nano /usr/local/lib/python2.7/dist-packages/pyscesmariner/MarinerConfig.py
PATH_LIBSBMLTHREAD = '/usr/local/lib/python2.7/dist-packages/pyscesmariner/libSBMLthread.pyc'
PATH_LIBSBML_CONVERTTHREAD = '/usr/local/lib/python2.7/dist-packages/pyscesmariner/libSBMLConvertThread.py'

cd to <pysces_cbm_mariner>:

python setup.py build sdist
sudo python setup.py install

Test installation

Open a new terminal window:

# cd <pysces_cbm_mariner>/demo
python cbm_server_demo.py

Open another terminal and run the client demo:

python cbm_client_demo.py

Kill the server by closing the terminal window.

Python3

Not all dependencies are available for Python3:

sudo apt-get install python3-dev python3-numpy python3-scipy
sudo apt-get install python3-matplotlib  python3-pip
sudo apt-get install python3-xlrd python3-h5py

# need to find out what is going on with Python3 and xlwt suds
# easy_install3 sympy ???
# wxPython and PyQt4 not in Ubuntu P3 builds yet

sudo apt-get install ipython3 ipython3-notebook

sudo apt-get install libxml2 libxml2-dev
sudo apt-get install zlib1g zlib1g-dev
sudo apt-get install bzip2 libbz2-dev

sudo pip3 install python-libsbml

sudo apt-get install python-qt4 python-qt4-dev python-sip
sudo apt-get install python-sip-dev build-essential

Apple Macintosh: OS X

Installation is similar to Linux except packages are installed using distutils and pip. The first step is to install the Mac development tools xcode

Install Python packages:

sudo easy_install numpy ipython scipy matplotlib
sudo easy_install xlrd xlwt sympy suds pyparsing pip

Use pip to install advanced Ipython and libsbml:

sudo pip install ipython[notebook]
ARCHFLAGS=-Wno-error=unused-command-line-argument-hard-error-in-future  pip install python-libsbml

For solvers, either install your own copy of CPLEX or build PyGLPK which requires building both the GMP and GLPK libraries.

GMP (https://gmplib.org/):

download gmp
./configure --prefix=/usr/local
make
make check
sudo make install

GLPK (http://sourceforge.net/projects/cbmpy/files/tools/glpk):

download glpk-4.47.tar.gz
./configure --prefix=/usr/local --with-gmp
make
sudo make install

PyGLPK (http://sourceforge.net/projects/cbmpy/files/tools/glpk):

download python-glpk-0.3
python setup.py build
sudo python setup.py install

Installing PySCeS-CBM Mariner (Microsoft Windows and Linux)

The PySCeS Mariner module exposes the CBMPy functionality as SOAP web services (e.g. as a backend to FAME (http://F-A-M-E.org)). It is available for download from SourceForge:

Dependencies: CherryPy, libXML and SOAPlib

PySCeS-CBM Mariner requires (pure python) soaplib 0.8.1 (supplied with it) or downloadable from:

https//sourceforge.net/projects/cbmpy/files/tools/soaplib/

Soaplib itself has two dependencies which should be installed first:

  • LXML (http://lxml.de)
    • Windows: install with easy_install lxml
    • Linux (Ubuntu) use sudo apt-get install python-lxml
  • CherryPy (http://www.cherrypy.org)
    • Windows: install with easy_install cherrypy
    • Linux (Ubuntu) use sudo apt-get install python-cherrypy
  • SOAPLIB 0.8.1:
    • Windows: Execute soaplib-0.8.1.win32.exe
    • Linux: Unpack the zip archive and run sudo python setup.py install

Test installation:

  • Open a terminal
  • Execute “ipython”
  • Execute “import cherrypy, lxml, soaplib” no errors or warnings should be generated
  • Exit ipython with CTRL-D
  • change directory to supplied soaplib tests e.g. “cd e:\cbmpy\tests\soaplib”
  • Execute “python binary_test.py”
  • Execute “python primitive_test.py”

All tests should pass.

PySCeS-CBM Mariner (http://cbmpy.sourceforge.net)

Download and install the latest version (0.7.4 or newer is required for CBMPy 0.7+):

  • Windows: Execute pyscesmariner-0.7.7.zip
  • Linux: unpack the archive and run sudo python setup.py install

To test installation, on Linux execute the commands in run_server.bat from the terminal directly.

  • Open two terminals and in both
  • Change directory to supplied PySCeS-CBM Mariner tests e.g. cd e:\\cbmpy\\tests\\pyscesmariner
  • In terminal one Execute run_server.bat

Which should now display:

E:\\cbmpy\\tests\\pyscesmariner>python cbm_server_demo.py
Mariner using E:\\cbmpy\\tests\\pyscesmariner as a working directory
Mariner server name: 10.0.2.15
Mariner using port: 31313

Welcome to the PySCeS Constraint Based Modelling Toolkit (0.7.4)

<snipped>

Multiple Environment Module (0.6.2 [r1147])

PySCeSCBM/Mariner initialising ... this console is now blocked

In terminal two:

  • Execute python cbm_client_demo.py

This should end without errors and display done. Congratulations you have successfully installed CBMPy and PySCeS-CBM Mariner!