Difference between revisions of "Applications/Python"

From HPC
Jump to: navigation , search
m (Pysdlb moved page Python to Applications/Python without leaving a redirect)
m (Python Virtual Environment)
 
(48 intermediate revisions by the same user not shown)
Line 1: Line 1:
 
__TOC__
 
__TOC__
 +
 
=== Application Details ===
 
=== Application Details ===
 
* Description: Python is a high-level interpreted programming language for general-purpose programming, supported by a large number of libraries for many tasks
 
* Description: Python is a high-level interpreted programming language for general-purpose programming, supported by a large number of libraries for many tasks
 
* Versions: Python 2.7.11 and 3.5.1
 
* Versions: Python 2.7.11 and 3.5.1
* Module names: python/2.7.11 python/3.5.1  
+
* Module names: python/anaconda/4.0/2.7, python/anaconda/4.0/3.5, python/anaconda/4.1.1/2.7 and python/anaconda/4.3.31/3.6-VE
 +
* Additional module: python/anaconda/4.6/miniconda/3.7 and python/anaconda/202111/3.9 (used for virtual environments)
 
* License: Free to use - [https://en.wikipedia.org/wiki/Python_Software_Foundation_License Python Software Foundation License]
 
* License: Free to use - [https://en.wikipedia.org/wiki/Python_Software_Foundation_License Python Software Foundation License]
  
 +
== Introduction ==
 +
 +
* Python is provided by the [[applications/Anaconda|Anaconda package]] too.
 +
* Anaconda is the leading open data science platform powered by Python.
 +
 +
== Virtual Environments ==
 +
 +
A Python virtual environment allows users to create a custom environment(s) in which they can have the packages, and versions of those packages that are required without having elevated user privileges.
 +
 +
We recommend the Virtualenv installation when you use a specialised package that would not be used by the wider HPC community. Virtualenv is a virtual Python environment isolated from other Python development, incapable of interfering with or being affected by other Python programs on the same HPC. During the Virtualenv installation process, you will install can install not only the additional package but all the dependencies that go with it. (This is actually pretty easy.)  All in all, Virtualenv provides a safe and reliable mechanism for installing and running additional packages.
 +
 +
There are many benefits to using a virtual environment on a system. Here is a short non-exhaustive list of a few of the key benefits:
 +
 +
* Control over packages and package versions.
 +
* An arbitrary number of virtual environments can be created for different tasks.
 +
* Reproducible - A user can replicate a python environment on any system, so as to be able to run or re-run a particular task or job under the same conditions.
 +
 +
 +
===Creation of Virtual Environment Using Anaconda on Viper===
 +
 +
* See [[Applications/Miniconda]] page for creating your own virtual environment. This is the preferred method for specialised python modules required such as biopython or TensorFlow for example.
  
== Usage Examples ==
 
  
Python is provided by the [[applications/Anaconda|Anaconda package]] too
+
==Usage Examples==
  
=== Interactive ===
+
=== Interactive Session ===
  
 
Interactive with command line:
 
Interactive with command line:
Line 25: Line 47:
 
</pre>
 
</pre>
  
Interactive with command line with IPython:
+
ipython with the command line  
 +
 
 
<pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
 
<pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
 
[username@c170 ~]$ ipython
 
[username@c170 ~]$ ipython
Line 41: Line 64:
  
 
=== Batch Submission ===
 
=== Batch Submission ===
 +
 
<pre style="background-color: #C8C8C8; color: black; font-family: monospace, sans-serif;">
 
<pre style="background-color: #C8C8C8; color: black; font-family: monospace, sans-serif;">
 
#!/bin/bash
 
#!/bin/bash
Line 50: Line 74:
 
#SBATCH -p compute                  # Slurm partition, where you want the job to be queued  
 
#SBATCH -p compute                  # Slurm partition, where you want the job to be queued  
  
 
 
module purge
 
module purge
module add python/3.5.1
+
module add python/anaconda/4.0/3.5
 
   
 
   
 
python PythonTest.py
 
python PythonTest.py
Line 60: Line 83:
 
<pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
 
<pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
 
[username@login01 ~]$ sbatch Pythontest.job
 
[username@login01 ~]$ sbatch Pythontest.job
Submitted batch job 289522
+
Submitted batch job 2895122
 
</pre>
 
</pre>
  
== Virtual Environments ==
+
===Popular Python Packages===
A Python Virtual Environment allows users to create a custom environment/s in which they can have the packages, and versions of said packages that they require without having to have elevated user privileges on the system.
 
 
 
There are many benefits to using a virtual environment on a system. Here is a short non-exhaustive list of a few of the key benefits:
 
 
 
* Control over packages and package versions
 
* An arbitrary number of virtual environments can be created for different tasks
 
* Reproducibility - A user can replicate a python environment on any system, so as to be able to run or re-run a particular task or job under the same conditions
 
  
===Creation of Virtual Environment Using Python 2.7.11 or 3.5.1 on Viper===
+
* TensorFlow
 +
* Scikit-Learn
 +
* Numpy
 +
* PyTorch
 +
* SciPy
 +
* Matplotlib
 +
* Pandas
  
 +
===Python Virtual Environment===
  
To create a virtual environment you would issue the following command:
+
This part refers to non-conda virtual environments.
  
'''IMPORTANT NOTE: By default the virtual environment does not use the python system packages. However because of the way viper is configured it will see python system packages because of the PYTHONPATH environment variable. So it is advised if you would like a clean environment (meaning no system packages being included) to set this environment variable to empty as follows: ''export PYTHONPATH=''
+
The creation of a virtual environment is done by executing the command venv:
  
<pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
+
<pre>
[user@c001 ~ ]$ virtualenv foo
+
python -m venv machinelearn
</pre>
 
This creates a directory containing your virtual environment named '''foo''' . To activate this virtual environment you would issue the following command:
 
<pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
 
[user@c001 ~ ]$ source foo/bin/activate
 
 
</pre>
 
</pre>
If the environment starts correctly you should see the name of your environment in round brackets in front of the command prompt. For this particular example it should appear as follows:
 
  <pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
 
(foo)  [user@c001 ~ ]$
 
</pre>
 
To exit the virtual environment use the key combination '''ctrl + d'''
 
  
===Creation of Virtual Environment Using Anaconda on Viper===
+
Running this command creates the target directory (creating any parent directories that don’t exist already) and places a pyvenv.cfg file in it with a home key pointing to the Python installation from which the command was run (a common name for the target directory is .venv).
  
To create a virtual environment using anaconda 4.1.1 with python version 2.7 on Viper,  you would use the conda create command as follows:
+
To activate
  
'''IMPORTANT NOTE: By default the virtual environment does not use the python system packages. However because of the way viper is configured it will see python system packages because of the PYTHONPATH environment variable. So it is advised if you would like a clean environment (meaning no system packages being included) to set this environment variable to empty as follows: ''export PYTHONPATH=''
+
<pre>
 
+
source machinelearn/bin/activate
<pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
 
[user@c001 ~ ]$ conda create –n newenv pip numpy
 
</pre>
 
The above command creates a new virtual environment called newenv and installs pip and numpy within this environment.
 
 
 
To activate this virtual environment you would issue the following command:
 
<pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
 
[user@c001 ~ ]$ source activate newenv
 
</pre>
 
On successful activation of this virtual environment you should the name of your environment in front of your login prompt like so:
 
<pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
 
(newenv)  [user@c001 ~ ]$
 
</pre>
 
To exit the virtual environment use the key combination '''ctrl + d'''
 
 
 
===Creation of a Virtual Environment in Anaconda Using a YAML File===
 
 
 
To create a virtual environment from a YAML file you would issue the following command:
 
 
 
<pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
 
[user@c001 ~ ]$ conda env create -f myenv.yml
 
</pre>
 
 
 
This above command is creating a virtual environment from the YAML called myenv.yml. Below is a copy of the mark-up contained within the file called "myenv.yml".
 
 
 
<pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
 
name: ytenv
 
channels:
 
- defaults
 
dependencies:
 
- ca-certificates=2017.08.26=h1d4fec5_0
 
- certifi=2018.1.18=py27_0
 
- intel-openmp=2018.0.0=hc7b2577_8
 
- libedit=3.1=heed3624_0
 
- libffi=3.2.1=hd88cf55_4
 
- libgcc-ng=7.2.0=h7cc24e2_2
 
- libgfortran-ng=7.2.0=h9f7466a_2
 
- libstdcxx-ng=7.2.0=h7a57d05_2
 
- mkl=2018.0.1=h19d6760_4
 
- ncurses=6.0=h9df7e31_2
 
- numpy=1.14.0=py27h3dfced4_1
 
- openssl=1.0.2n=hb7f436b_0
 
- pip=9.0.1=py27ha730c48_4
 
- python=2.7.14=h1571d57_29
 
- readline=7.0=ha6073c6_4
 
- setuptools=38.4.0=py27_0
 
- sqlite=3.22.0=h1bed415_0
 
- tk=8.6.7=hc745277_3
 
- wheel=0.30.0=py27h2bc6bb2_1
 
- zlib=1.2.11=ha838bed_2
 
- pip:
 
  - backports.functools-lru-cache==1.5
 
  - backports.shutil-get-terminal-size==1.0.0
 
  - cycler==0.10.0
 
  - decorator==4.2.1
 
  - enum34==1.1.6
 
  - h5py==2.7.1
 
  - ipython==5.5.0
 
  - ipython-genutils==0.2.0
 
  - matplotlib==2.1.2
 
  - mpmath==1.0.0
 
  - pathlib2==2.3.0
 
  - pexpect==4.4.0
 
  - pickleshare==0.7.4
 
  - prompt-toolkit==1.0.15
 
  - ptyprocess==0.5.2
 
  - pygments==2.2.0
 
  - pyparsing==2.2.0
 
  - python-dateutil==2.6.1
 
  - pytz==2018.3
 
  - scandir==1.7
 
  - simplegeneric==0.8.1
 
  - six==1.11.0
 
  - subprocess32==3.2.7
 
  - sympy==1.1.1
 
  - traitlets==4.3.2
 
  - wcwidth==0.1.7
 
  - yt==3.4.1
 
</pre>
 
 
 
===Exporting a Virtual Environment in Anaconda to a YAML File===
 
 
 
To export a virtual environment to a YAML file so that you or a third party can replicate your environment using Anaconda can be done using the following steps:
 
 
 
'''1.)'''  Activate the Virtual environment you wish to export
 
<pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
 
[user@c001 ~ ]$ source activate newenv
 
</pre>
 
'''2.)'''Export your active virtual environment using the following command
 
<pre style="background-color: #000000; color: white; border: 2px solid black; font-family: monospace, sans-serif;">
 
[user@c001 ~ ]$ conda env export > env.yml
 
 
</pre>
 
</pre>
  
==Further Information==
+
==Next Steps==
  
* [https://www.hpc.hull.ac.uk/forum/viewforum.php?f=15 UoH Python forum]
+
* [[Applications/Miniconda|Miniconda Python virtual environments]]
* [[Programming/Python|Wiki: Python Programming]]
+
* [[Applications/Anaconda|Anaconda Python]]  
* [[Main Page#General_Support|Wiki: General Support]]
+
* [[Programming/Python|Python Programming]]
* [[Main Page|Wiki: Home]]
 
  
{|
+
{{Modulepagenav}}
|style="width:5%; border-width: 0" | [[File:icon_home.png]]
 
|style="width:95%; border-width: 0" |
 
* [[Main_Page|Home]]
 
* [[Applications|Application support]]
 
* [[General|General]]
 
* [[Training|Training]]
 
* [[Programming|Programming support]]
 
|-
 
|}
 

Latest revision as of 12:30, 17 February 2023

Application Details

  • Description: Python is a high-level interpreted programming language for general-purpose programming, supported by a large number of libraries for many tasks
  • Versions: Python 2.7.11 and 3.5.1
  • Module names: python/anaconda/4.0/2.7, python/anaconda/4.0/3.5, python/anaconda/4.1.1/2.7 and python/anaconda/4.3.31/3.6-VE
  • Additional module: python/anaconda/4.6/miniconda/3.7 and python/anaconda/202111/3.9 (used for virtual environments)
  • License: Free to use - Python Software Foundation License

Introduction

  • Python is provided by the Anaconda package too.
  • Anaconda is the leading open data science platform powered by Python.

Virtual Environments

A Python virtual environment allows users to create a custom environment(s) in which they can have the packages, and versions of those packages that are required without having elevated user privileges.

We recommend the Virtualenv installation when you use a specialised package that would not be used by the wider HPC community. Virtualenv is a virtual Python environment isolated from other Python development, incapable of interfering with or being affected by other Python programs on the same HPC. During the Virtualenv installation process, you will install can install not only the additional package but all the dependencies that go with it. (This is actually pretty easy.) All in all, Virtualenv provides a safe and reliable mechanism for installing and running additional packages.

There are many benefits to using a virtual environment on a system. Here is a short non-exhaustive list of a few of the key benefits:

  • Control over packages and package versions.
  • An arbitrary number of virtual environments can be created for different tasks.
  • Reproducible - A user can replicate a python environment on any system, so as to be able to run or re-run a particular task or job under the same conditions.


Creation of Virtual Environment Using Anaconda on Viper

  • See Applications/Miniconda page for creating your own virtual environment. This is the preferred method for specialised python modules required such as biopython or TensorFlow for example.


Usage Examples

Interactive Session

Interactive with command line:

[username@c170 ~]$ module add python/anaconda/4.0/2.7
[username@c170 ~]$ python
Python 2.7.5 (default, Nov 20 2015, 02:00:19)
[GCC 4.8.5 20150623 (Red Hat 4.8.5-4)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>>

ipython with the command line

[username@c170 ~]$ ipython
Python 2.7.13 |Anaconda custom (64-bit)| (default, Dec 20 2016, 23:09:15)
Type "copyright", "credits" or "license" for more information.

IPython 5.1.0 -- An enhanced Interactive Python.
?         -> Introduction and overview of IPython's features.
%quickref -> Quick reference.
help      -> Python's own help system.
object?   -> Details about 'object', use 'object??' for extra details.

In [1]:

Batch Submission

#!/bin/bash
#SBATCH -J PythonTest                # Job name, you can change it to whatever you want
#SBATCH -N 1                                  # Number of nodes 
#SBATCH --ntasks-per-node 1    # Number of cores per node
#SBATCH -o %N.%j.out                # Standard output will be written here
#SBATCH -e %N.%j.err                 # Standard error will be written here
#SBATCH -p compute                  # Slurm partition, where you want the job to be queued 

module purge
module add python/anaconda/4.0/3.5
 
python PythonTest.py

This is then submitted as follows:

[username@login01 ~]$ sbatch Pythontest.job
Submitted batch job 2895122

Popular Python Packages

  • TensorFlow
  • Scikit-Learn
  • Numpy
  • PyTorch
  • SciPy
  • Matplotlib
  • Pandas

Python Virtual Environment

This part refers to non-conda virtual environments.

The creation of a virtual environment is done by executing the command venv:

python -m venv machinelearn

Running this command creates the target directory (creating any parent directories that don’t exist already) and places a pyvenv.cfg file in it with a home key pointing to the Python installation from which the command was run (a common name for the target directory is .venv).

To activate

source machinelearn/bin/activate

Next Steps





Modules | Main Page | Further Topics