Installing Python

There are multiple options to install Python. If you are working with Linux, hopefully this document can help you. If you are working with another OS or the instructions below do not help you, you can visit the official Python site, where there is a link to Download Python. Then install Python using the downloaded file.

In case you need/want to learn more about the Python installation process in general, you might be interested in reading this Python installation guide.

In Linux, if you have administration permissions, the easiest way to install Python is through your system’s package manager. How this can be done is obviously outside the scope of this documentation.

If you don’t have priviledges to install software on your target system using the package manager (that is typically the case if you plan to run PacBio Data Processing on a cluster), you still can easily use Spack to install Python. Alternatively, you can also install Python directly from sources. In the sections below these two options are described.

Default: installing Python with Spack

With Spack (see Using Spack for details) the installation of Python is straigthforward. Once Spack is on your system, which doesn’t require particular priviledges, run:

$ spack install python@3.9.13

to install CPython-3.9.13 (see CPython).

On success, Spack will inform you about the module that has been created and which must be loaded if you want to use it. In my case the name of the module is python-3.9.13-gcc-9.5.0-oesebwh, hence, to use it:

$ module load python-3.9.13-gcc-9.5.0-oesebwh

YMMV. If you forget the name of the module, you can use the <TAB> to autocomplete the line, or list all modules available in your system with:

$ module avail

Alternative: installing Python from sources

Warning

Use this option as a last resort, since you need to be able to install some libraries in order to produce a usable Python installation from the sources. If you cannot install Python with your package manager, you might expect problems getting Python up and running from sources…

If you download the sources (typically a file ending in .tgz, .tar.xz or similar) the procedure is relatively simple:

1. untar the file. For instance:

   $ tar xf Python-3.9.13.tgz
   

2. Enter in the created directory with the sources:

   $ cd Python-3.9.13
   

3. Open the README.rst file and follow the instructions in its Build Instructions section. They schematically amount to 3 steps:

   $ ./configure
   $ make
   $ make install
   

   In the configure step you can customize various features of the compilation and installation process. For instance, two common options are --prefix (to install Python in a custom location and not in the default system-wide location) and --enable-optimizations. The following command (to be executed before make) would use both options:

   $ ./configure --prefix=/home/dave --enable-optimizations
   

   The given prefix (/home/dave) implies that the make install step will create suitable directories under the given location to have the necessary structure to use Python. It is up to the user to leverage that installation. That could be as simple as calling python with the full (or relative) path to its interpreter, e.g. in the case of using --prefix=/home/dave:

   $ /home/dave/bin/python3
   

   If you want to use the compiled interpreter as your default Python interpreter, modify the PATH shell variable to make the terminal aware of your recently installed Python:

   $ export PATH="/home/dave/bin:${PATH}"
   

   After that, the current shell and any subprocesses created from it, will know that by typing:

   $ python
   # or
   $ python3
   

   (or any other executable installed, check the contents of /home/dave/bin) you mean actually /home/dave/bin/python or /home/dave/bin/python3, etc.

   If you want this change to be permanent, i.e. if you don’t want to type the complete path to your newly installed Python interpreter but you don’t want to modify the PATH variable every time you log in, and you are using Bash, add the export line above near the end of your .bashrc file (/home/dave/.bashrc in our example); if .bashrc is missing in your home directory, feel free to create one.

   The CPython README.rst file is worth having a look for more insights about the options available for the manual installation of CPython. It contains very useful hints and a lot more information.