Installing Cinfony on Linux

The following instructions take you through the process of installing Cinfony and all of its dependencies. Note that all dependencies are optional so if you don’t want to use a particular toolkit, there is no need to follow the instructions for installing it.

A brief word about installing Python packages

After extracting the .tar.gz file, Python packages are generally installed in one of two ways:

  • globally, with python setup.py install (requires root)
  • locally, with python setup.py install --prefix=$HOME
    • This requires the PYTHONPATH variable to be set to include the directory where the files were installed (in this case, $HOME/lib/python2.7/site-packages or so).

Prerequisites

  • A recent Python, Jython or IronPython in the 2.x series
  • Java - If using Jython, or accessing a Java library (e.g. CDK, or OPSIN)
    • Install JDK 7.0 (1.7.0)
  • Mono - If using IronPython

Cinfony

  • Download and install Cinfony 1.2
    • Test at the Python prompt as follows:
>>> import cinfony
  • If you want to draw 2D diagrams for some of the modules, you need to install the Python Imaging Library (PIL)
    • Use your distribution’s package manager to install PIL. On Debian Etch, for example, the package was split into python-imaging and python-imaging-tk
    • Test at the Python prompt as follows:
>>> import Image
>>> import ImageTk
  • In order for Jython to find Cinfony, add a line to the registry file in the Jython installation directory as follows:
    • python.path = /usr/local/lib/python2.7/site-packages where you should replace the path by the path to the directory where Cinfony was installed
  • In order for IronPython to find Cinfony, add the installation directory to the IRONPYTHONPATH environment variable
  • Test at the Jython or IronPython prompt as follows:
>>> import cinfony

Cinfony contains the following 7 modules (prerequisites listed in parentheses):

  1. pybel - OpenBabel wrapper (Python or Java+Jython or IronPython)
  2. cdk - CDK wrapper (Java and Python+JPype or Jython)
  3. rdk - RDKit wrapper (Python)
  4. opsin - OPSIN wrapper (Java and Python+JPype or Jython)
  5. indy - Indigo wrapper (Python or Java+Jython or IronPython)
  6. jchem - JChem wrapper (Java and Python+JPype or Jython)
  7. webel - Uses web services

To install the dependencies for each wrapper, follow the appropriate instructions below. Note: there is no need to install the dependencies for a wrapper you are not going to use.

Open Babel

Open Babel is a C++ library and can be accessed from all of CPython, Jython and IronPython.

  • Follow the instructions to compile and install Open Babel

    • If you want to be able to create a 2D diagram of a molecule, you need to make sure that you have compiled Open Babel with support for Png output. This requires the Cairo library (and development headers) which you should install using your package manager, prior to compiling Open Babel.

  • Follow the instructions to compile and install the Open Babel Python bindings, Java bindings or CSharp bindings

In the following section $OB_INSTALLDIR refers to the directory where the Open Babel library was installed (typically /usr/local/lib).

  • For use from CPython, add the directory where the Python bindings were installed to the PYTHONPATH environment variable (if necessary)

  • For use from Jython:

    • Add the directory containing libopenbabel_java.so (typically $OB_INSTALLDIR) to the LD_LIBRARY_PATH environment variable
    • Add the directory containing openbabel.jar (typically $OB_INSTALLDIR) to the CLASSPATH environment variable

  • From use from IronPython:

    • Add the directory containing libopenbabel_csharp.so (typically $OB_INSTALLDIR) to the LD_LIBRARY_PATH environment variable
    • Add the directory containing OBDotNet.dll (typically $OB_INSTALLDIR) to the OBDOTNET environment variable

Test

Test at the Python, Jython or IronPython prompt as follows:

>>> from cinfony import obabel
>>> mol = obabel.readstring("smi", "CCC")
>>> print mol.molwt
44.09562
>>> mol.draw()

RDKit

The RDKit is a C++ library currently accessible only from CPython.

  • Download and extract RDKit 2012.09

  • Follow the installation instructions

  • Add the RDKit_2011_09_1 directory to the PYTHONPATH environment variable

  • Test at the Python prompt as follows:

    >>> from cinfony import rdk
>>> mol = rdk.readstring("smi", "CCC")
>>> print mol.molwt
44.097
>>> mol.draw()

CDK

The CDK is a Java library and can be accessed from both CPython and Jython.

  • Download CDK 1.4.15
  • Add the cdk-1.4.15.jar file to your CLASSPATH environment variable (remember to include the full path to the jar file)

No additional configuration is required to use the CDK from Jython, but to use it from CPython:

  • Use your package manager to install JPype, or else download and install it
    • You need to set JAVA_HOME to the Java installation directory before running setup.py.
  • Set JPYPE_JVM environment variable to point to the libjvm.so file in your Java installation directory
    • On my system, it’s /home/user/Tools/jdk1.7.0_4/jre/lib/i386/client/libjvm.so

Test

Test at the Python or Jython prompt as follows:

>>> from cinfony import cdk
>>> mol = cdk.readstring("smi", "CCC")
>>> print mol.molwt
44.0956192017
>>> mol.draw()

Indigo

Indigo is a C++/C library and can be accessed from CPython and Jython (GGASoftware did not provide the .NET bindings for Linux at the time).

  • Download the Python and/or Java API for Linux and unzip them
  • To access Indigo from CPython, add the extracted Python directory to PYTHONPATH
  • To access Indigo from Jython, add all of the jar files in the extracted Java directory to CLASSPATH
  • Test at the Python or Jython prompt as follows:
>>> from cinfony import indy
>>> mol = indy.readstring("smi", "CCC")
>>> print mol.molwt
44.0956192017
>>> mol.draw()

JChem

JChem is a Java toolkit and can be accessed from both CPython (using JPype) and from Jython. This is a commercial toolkit and requires a license.

  • Once installed, add jchem.jar to the CLASSPATH.
  • For use from CPython, see instructions for installing JPype in the CDK section above.
  • Test at the Python or Jython prompt as follows:
>>> from cinfony import jchem
>>> mol = jchem.readstring("smi", "CCC")
>>> print mol.molwt
44.0956001282
>>> mol.draw()

OPSIN

OPSIN can be accessed from both CPython (using JPype) and from Jython.

  • Please follow the same instructions as for the CDK (above), but using the OPSIN 1.3 jar file.
  • Test at the Python or Jython prompt as follows:
>>> from cinfony import opsin
>>> mol = opsin.readstring("iupac", "2-chloro-propane")
>>> print mol.write("smi")
ClC(C)C

Webel

Webel is a Cinfony module that uses web services. It is pure-Python and does not require any additional configuration. It will work equally well from all versions of Python.

Common problems

  1. If when using one of the Java bindings, you get a complaint about not being able to find libawt.so, add to the LD_LIBRARY_PATH environment variable the directory in your Java installation containing libawt.so.

    • On my system, it’s /home/user/Tools/jdk1.7.0_4/jre/lib/i386