Python - Subprocess

Overview

Questions:
  • How can I run another program?

Objectives:
  • Run a command in a subprocess.

  • Learn about check_call and check_output and when to use each of these.

  • Read it’s output.

Requirements:
Time estimation: 45 minutes
Level: Intermediate Intermediate
Supporting Materials:
Last modification: Apr 25, 2022
License: Tutorial Content is licensed under Creative Commons Attribution 4.0 International License The GTN Framework is licensed under MIT

Best viewed in a Jupyter Notebook

This tutorial is best viewed in a Jupyter notebook! You can load this notebook one of the following ways

Launching the notebook in Jupyter in Galaxy

  1. Instructions to Launch JupyterLab
  2. Open a Terminal in JupyterLab with File -> New -> Terminal
  3. Run wget https://training.galaxyproject.org/archive/2022-08-01/topics/data-science/tutorials/python-subprocess/data-science-python-subprocess.ipynb
  4. Select the notebook that appears in the list of files on the left.

Downloading the notebook

  1. Right click one of these links: Jupyter Notebook (With Solutions), Jupyter Notebook (Without Solutions)
  2. Save Link As..

Sometimes you need to run other tools in Python, like maybe you want to Here we’ll give a quick tutorial on how to read and write files within Python.

Agenda

In this tutorial, we will cover:

  1. Subprocesses
  2. Check Call: Downloading Files
  3. Check Output: Gene Calling with Augustus
  4. Aside: stdin, stderr, stdout
  5. Pipes

Subprocesses

Programs can run other programs, and in Python we do this via the subprocess module. It lets you run any other command on the system, just like you could at the terminal.

The first step is importing the module

import subprocess

You’ll primarily use two functions:

help(subprocess.check_call)

Which executes a command and checks if it was successful (or it raises an exception), and

help(subprocess.check_output)

Check Call: Downloading Files

Which executes a command returns the output of that command. This is really useful if you’re running a subprocess that writes something to stdout, like a report you need to parse. We’ll learn how to use these by running two gene callers, augustus and glimmer. You can install both from Conda if you do not have them already.

conda create -n subprocess augustus glimmer3

Additionally you’ll need two files, you generally should not do this, but you can use a subprocess to download the file! We’ll use subprocess.check_call for this which simply executes the program, and continues on. If there is an error in the execution, it will raise an exception and stop execution.

url = "https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/836/945/GCF_000836945.1_ViralProj14044/GCF_000836945.1_ViralProj14044_genomic.fna.gz"
genome = 'Escherichia virus T4.fna.gz'
subprocess.check_call(['wget', url, '-O', genome])
subprocess.check_call(['gzip', '-d', genome])

tip Tip: What do these commands look like on the CLI?

wget https://ftp.ncbi.nlm.nih.gov/.... -O "Escherichia virus T4.fna.gz"
gzip -d "Escherichia virus T4.fna.gz"

The above segment

  • sets a url variable
  • sets an output filename, Escherichia virus T4.fna.gz
  • runs check_call with a single argument: a list
    • wget a tool we use to download files
    • the URL
    • -O indicating the next argument will be the ‘output name’
    • what we want the output filename to be called
  • runs check_call with a single argument: a list
    • gzip a tool to decompress files --d indicating we want to decompress
    • and the filename.

This list is especially important. When you run commands on the command line, normally you just type in a really bit of text by yourself. It’s one big string, and you’re responsible for making sure quotation marks appear in the right place. For instance, if you have spaces in your filenames, you have to quote the filename. Python requires you specify a list of arguments, and then handles the quoting for you! Which, honestly, is easier and safer.

code-in Terminal

Here we manually quote the argument

glimmer3 "bow genome.txt"

code-out Python

Here python handles that for us

subprocess.check_call(['glimmer3', 'bow genome.txt'])

tip Tip: Exploitation!

This is one of the major reasons we don’t use os.system or older Python interfaces for running commands. If you’re processing files, and a user supplies a file with a space, if your program isn’t expecting that space in that filename, then it could do something dangerous! Like exploit your system!

So, always use subprocess if you run to commands, never any other module, despite what you see on the internet!

There are more functions in the module, but the vast majority of the time, those are sufficient.

Check Output: Gene Calling with Augustus

gff3 = subprocess.check_output([
    'augustus', # Our command
    '--species=E_coli_K12', # the first argument, the species, we're using a phage so we call genes  based on it's host organism
    'Escherichia virus T4.fna', # The path to our genome file, without the .gz because we decompressed it.
    '--gff3=on' # We would like gff3 formatted output (it's easier to parse!)
])

gff3 = gff3.decode('utf-8')
gff3 = gff3.split('\n')

tip Tip: What does this commands look like on the CLI?

augustus --species=E_coli_K12 'Escherichia virus T4.fna' --gff3=on

If you’re using subprocess.check_output() python doesn’t return plain text str to you, instead it returns a bytes object. We can decode that into text with .decode('utf-8'), a phrase you should memorise as going next to check_output(), for 99% of use cases.

Let’s look at the results!

print(gff3[0:20])

It’s a lot of comment lines, starting with #. Let’s remove those

cleaned_gff3 = []
for line in gff3:
    if line.startswith('#'):
        continue
    cleaned_gff3.append(line)

print(cleaned_gff3[0:20])

And now you’ve got a set of gff3 formatted gene calls! You can use all of your loop processing skills to slice and dice this data into something great!

Aside: stdin, stderr, stdout

All unix processes have three default file handles that are available to them:

  • stdin, where data is passed to the program via a pipe. E.g. generate-data | my-program, there the program would read the output of generate-data from the pipe.
  • stdout, the default place where things are written. E.g. if you print() in python, it goes to stdout. People often redirect stdout to a file, like my-program > output.txt to save the output.
  • stderr, generally if your program produces output on stdout, you might still want to log messages (errors, % done, etc.) If you write to stdout, it might get mixed in with the user’s outputs, so we write to stderr, which also gets printed to the screen, and looks identical as any print statement, but it’s coming from a separate pipe.

Pipes

One of the more complicated cases, however, is when you need pipes.

url = "https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/001/721/125/GCF_001721125.1_ASM172112v1/GCF_001721125.1_ASM172112v1_cds_from_genomic.fna.gz"
cds = 'E. Coli CDSs.fna.gz'
subprocess.check_call(['wget', url, '-O', cds])
subprocess.check_call(['gzip', '-d', cds])

With subprocesses, you can control the stdin, and stdout of the process by using file handles.

code-in Terminal

Here we pipe a file to a process named build-icm which takes one argument, the output name. It reads sequences from stdin.

cat seq.fa | build-icm test.icm
# OR
build-icm test.icm < seq.fa

code-out Python

Here we need to do a bit more.

  1. Open a file handle
  2. Pass that file handle to check_call or check_output. This determines where stdin comes from.
    with open('seq.fa', 'r') as handle:
     subprocess.check_call(['build-icm', 'test.icm'], stdin=handle)
    

We’ll do that now:

with open('E. Coli CDSs.fna', 'r') as handle:
    subprocess.check_call(['build-icm', 'test.icm'], stdin=handle)

tip Tip: What does this commands look like on the CLI?

build-icm test.icm < 'E. Coli CDSs.fna'

Here we build a model, based on the sequences of E. Coli K-12, that Glimmer3 can use.

output = subprocess.check_output([
    'glimmer3', # Our program
    'Escherichia virus T4.fna', # The input genome
    'test.icm', # The model we just built
    't4-genes'  # The base name for output files. It'll produce t4-genes.detail and t4-genes.predict.
]).decode('utf-8') # And of course we decode as utf-8

print(output)

tip Tip: What does this commands look like on the CLI?

glimmer3 'Escherichia virus T4.fna' test.icm t4-genes

What happened here? The output of the program was written to stderr, not stdout, so Python may print that out to your screen, but output will be empty. To solve this common problem we can re-run the program and collect both stdout and stderr.

output = subprocess.check_output([
    'glimmer3', # Our program
    'Escherichia virus T4.fna', # The input genome
    'test.icm', # The model we just built
    't4-genes'  # The base name for output files. It'll produce t4-genes.detail and t4-genes.predict.
], stderr=subprocess.STDOUT).decode('utf-8') # And of course we decode as utf-8

print(output)

Here we’ve re-directed the stderr to stdout and mixed both of them together. This isn’t always what we want, but here the program produces no output, and we can do that safely, and now we can parse it or do any other computations we need with it! Our Glimmer3 gene calls are in t4-genes.detail and t4-genes.predict if we want to open and process those as well.

Key points

  • DO NOT USE os.system

  • DO NOT USE shell=True

  • 👍Use subprocess.check_call() if you don’t care about the output, just that it succeeds.

  • 👍Use subprocess.check_output() if you want the output

  • Use .decode('utf-8') to read the output of check_output()

Frequently Asked Questions

Have questions about this tutorial? Check out the FAQ page for the Foundations of Data Science topic to see if your question is listed there. If not, please ask your question on the GTN Gitter Channel or the Galaxy Help Forum

Feedback

Did you use this material as an instructor? Feel free to give us feedback on how it went.
Did you use this material as a learner or student? Click the form below to leave feedback.

Click here to load Google feedback frame

Citing this Tutorial

  1. Helena Rasche, Donny Vrins, Bazante Sanders, 2022 Python - Subprocess (Galaxy Training Materials). https://training.galaxyproject.org/archive/2022-08-01/topics/data-science/tutorials/python-subprocess/tutorial.html Online; accessed TODAY
  2. Batut et al., 2018 Community-Driven Data Analysis Training for Biology Cell Systems 10.1016/j.cels.2018.05.012

details BibTeX

@misc{data-science-python-subprocess,
author = "Helena Rasche and Donny Vrins and Bazante Sanders",
title = "Python - Subprocess (Galaxy Training Materials)",
year = "2022",
month = "04",
day = "25"
url = "\url{https://training.galaxyproject.org/archive/2022-08-01/topics/data-science/tutorials/python-subprocess/tutorial.html}",
note = "[Online; accessed TODAY]"
}
@article{Batut_2018,
    doi = {10.1016/j.cels.2018.05.012},
    url = {https://doi.org/10.1016%2Fj.cels.2018.05.012},
    year = 2018,
    month = {jun},
    publisher = {Elsevier {BV}},
    volume = {6},
    number = {6},
    pages = {752--758.e1},
    author = {B{\'{e}}r{\'{e}}nice Batut and Saskia Hiltemann and Andrea Bagnacani and Dannon Baker and Vivek Bhardwaj and Clemens Blank and Anthony Bretaudeau and Loraine Brillet-Gu{\'{e}}guen and Martin {\v{C}}ech and John Chilton and Dave Clements and Olivia Doppelt-Azeroual and Anika Erxleben and Mallory Ann Freeberg and Simon Gladman and Youri Hoogstrate and Hans-Rudolf Hotz and Torsten Houwaart and Pratik Jagtap and Delphine Larivi{\`{e}}re and Gildas Le Corguill{\'{e}} and Thomas Manke and Fabien Mareuil and Fidel Ram{\'{\i}}rez and Devon Ryan and Florian Christoph Sigloch and Nicola Soranzo and Joachim Wolff and Pavankumar Videm and Markus Wolfien and Aisanjiang Wubuli and Dilmurat Yusuf and James Taylor and Rolf Backofen and Anton Nekrutenko and Björn Grüning},
    title = {Community-Driven Data Analysis Training for Biology},
    journal = {Cell Systems}
}
                

Congratulations on successfully completing this tutorial!