Build a Python app with PyInstaller
This tutorial shows you how to use Jenkins to orchestrate building a simple Python application with PyInstaller.
If you are a Python developer who is new to CI/CD concepts, or you might be familiar with these concepts but don’t know how to implement building your application using Jenkins, then this tutorial is for you.
The simple Python application (which you’ll obtain from a sample repository on GitHub) is a command line tool "add2vals" that outputs the addition of two values. If at least one of the values is a string, "add2vals" treats both values as a string and instead concatenates the values. The "add2" function in the "calc" library (which "add2vals" imports) is accompanied by a set of unit tests. These are tested with pytest to check that this function works as expected and the results are saved to a JUnit XML report.
The delivery of the "add2vals" tool through PyInstaller converts this tool into a standalone executable file for Linux, which you can download through Jenkins and execute at the command line on Linux machines without Python.
Note: Unlike the other tutorials in this documentation, this tutorial requires approximately 500 MB more Docker image data to be downloaded.
Duration: This tutorial takes 20-40 minutes to complete (assuming you’ve already met the include::partial$_prerequisites[prerequisites] below). The exact duration will depend on the speed of your machine and whether or not you’ve already run Jenkins in Docker from another tutorial.
You can stop this tutorial at any point in time and continue from where you left off.
If you’ve already run though another tutorial, you can skip the include::partial$_prerequisites[prerequisites] and Run Jenkins in Docker sections below and proceed on to forking the sample repository. (Just ensure you have Git installed locally.) If you need to restart Jenkins, simply follow the restart instructions in run Jenkins in Docker and then proceed on.
Unresolved include directive in modules/ROOT/pages/build-a-python-app-with-pyinstaller.adoc - include::partial$_prerequisites[] ** Git and optionally GitHub Desktop
Run Jenkins in Docker
In this tutorial, you’ll be running Jenkins as a Docker container from the
jenkins/jenkins
Docker
image.
To run Jenkins in Docker, follow the relevant instructions below for either macOS and Linux or Windows.
You can read more about Docker container and image concepts in the Docker section of the Installing Jenkins page.
Fork and clone the sample repository
Obtain the simple "add" Python application from GitHub, by forking the sample repository of the application’s source code into your own GitHub account and then cloning this fork locally.
-
Ensure you are signed in to your GitHub account. If you don’t yet have a GitHub account, sign up for a free one on the GitHub website.
-
Fork the
simple-python-pyinstaller-app
on GitHub into your local GitHub account. If you need help with this process, refer to the Fork A Repo documentation on the GitHub website for more information. -
Clone your forked
simple-python-pyinstaller-app
repository (on GitHub) locally to your machine. To begin this process, do either of the following (where<your-username>
is the name of your user account on your operating system):-
If you have the GitHub Desktop app installed on your machine:
-
In GitHub, click the green Clone or download button on your forked repository, then Open in Desktop.
-
In GitHub Desktop, before clicking Clone on the Clone a Repository dialog box, ensure Local Path for:
-
macOS is
/Users/<your-username>/Documents/GitHub/simple-python-pyinstaller-app
-
Linux is
/home/<your-username>/GitHub/simple-python-pyinstaller-app
-
Windows is
C:\Users\<your-username>\Documents\GitHub\simple-python-pyinstaller-app
-
-
-
Otherwise:
-
Open up a terminal/command line prompt and
cd
to the appropriate directory on:-
macOS -
/Users/<your-username>/Documents/GitHub/
-
Linux -
/home/<your-username>/GitHub/
-
Windows -
C:\Users\<your-username>\Documents\GitHub\
(although use a Git bash command line window as opposed to the usual Microsoft command prompt)
-
-
Run the following command to continue/complete cloning your forked repo:
git clone https://github.com/YOUR-GITHUB-ACCOUNT-NAME/simple-python-pyinstaller-app
whereYOUR-GITHUB-ACCOUNT-NAME
is the name of your GitHub account.
-
-
Create your Pipeline project in Jenkins
-
Go back to Jenkins, log in again if necessary and click create new jobs under Welcome to Jenkins!
Note: If you don’t see this, click New Item at the top left. -
In the Enter an item name field, specify the name for your new Pipeline project (e.g.
simple-python-pyinstaller-app
). -
Scroll down and click Pipeline, then click OK at the end of the page.
-
( Optional ) On the next page, specify a brief description for your Pipeline in the Description field (e.g.
An entry-level Pipeline demonstrating how to use Jenkins to build a simple Python application with PyInstaller.
) -
Click the Pipeline tab at the top of the page to scroll down to the Pipeline section.
-
From the Definition field, choose the Pipeline script from SCM option. This option instructs Jenkins to obtain your Pipeline from Source Control Management (SCM), which will be your locally cloned Git repository.
-
From the SCM field, choose Git.
-
In the Repository URL field, specify the directory path of your locally cloned repository above, which is from your user account/home directory on your host machine, mapped to the
/home
directory of the Jenkins container - i.e.-
For macOS -
/home/Documents/GitHub/simple-python-pyinstaller-app
-
For Linux -
/home/GitHub/simple-python-pyinstaller-app
-
For Windows -
/home/Documents/GitHub/simple-python-pyinstaller-app
-
-
Click Save to save your new Pipeline project. You’re now ready to begin creating your
Jenkinsfile
, which you’ll be checking into your locally cloned Git repository.
Create your initial Pipeline as a Jenkinsfile
You’re now ready to create your Pipeline that will automate building your Python
application with PyInstaller in Jenkins. Your Pipeline will be created as a
Jenkinsfile
, which will be committed to your locally cloned Git repository
(simple-python-pyinstaller-app
).
This is the foundation of "Pipeline-as-Code", which treats the continuous delivery pipeline a part of the application to be versioned and reviewed like any other code. Read more about Pipeline and what a Jenkinsfile is in the Pipeline and Using a Jenkinsfile sections of the User Handbook.
First, create an initial Pipeline with a "Build" stage that executes the first part of the entire production process for your application. This "Build" stage downloads a Python Docker image and runs it as a Docker container, which in turn compiles your simple Python application into byte code.
-
Using your favorite text editor or IDE, create and save new text file with the name
Jenkinsfile
at the root of your localsimple-python-pyinstaller-app
Git repository. -
Copy the following Declarative Pipeline code and paste it into your empty
Jenkinsfile
:pipeline { agent none (1) stages { stage('Build') { (2) agent { docker { image 'python:2-alpine' (3) } } steps { sh 'python -m py_compile sources/add2vals.py sources/calc.py' (4) stash(name: 'compiled-results', includes: 'sources/*.py*') (5) } } } }
1 The agent
section with thenone
parameter specified at the top of this Pipeline code block means that no global agent will be allocated for the entire Pipeline’s execution and that eachstage
directive must specify its ownagent
section.2 Defines a stage
(directive) calledBuild
that appears on the Jenkins UI.3 This image
parameter (of theagent
section’sdocker
parameter) downloads thepython:2-alpine
Docker image (if it’s not already available on your machine) and runs this image as a separate container. This means that:-
You’ll have separate Jenkins and Python containers running locally in Docker.
-
The Python container becomes the agent that Jenkins uses to run the
Build
stage of your Pipeline project. However, this container is short-lived - its lifespan is only that of the duration of yourBuild
stage’s execution.
4 This sh
step (of thesteps
section) runs the Python command to compile your application and itscalc
library into byte code files (each with.pyc
extension), which are placed into thesources
workspace directory (within the/var/jenkins_home/workspace/simple-python-pyinstaller-app
directory in the Jenkins container).5 This stash
step (of thebasic steps
section) saves the Python source code and compiled byte code files (with.pyc
extension) from thesources
workspace directory for use in later stages. -
-
Save your edited
Jenkinsfile
and commit it to your localsimple-python-pyinstaller-app
Git repository. E.g. Within thesimple-python-pyinstaller-app
directory, run the commands:
git add .
then
git commit -m "Add initial Jenkinsfile"
-
Go back to Jenkins again, log in again if necessary and click Open Blue Ocean on the left to access Jenkins’s Blue Ocean interface.
-
In the This job has not been run message box, click Run, then quickly click the OPEN link which appears briefly at the lower-right to see Jenkins running your Pipeline project. If you weren’t able to click the OPEN link, click the row on the main Blue Ocean interface to access this feature.
Note: You may need to wait a few minutes for this first run to complete. After making a clone of your localsimple-python-pyinstaller-app
Git repository itself, Jenkins:-
Initially queues the project to be run on the agent.
-
Runs the
Build
stage (defined in theJenkinsfile
) on the Python container. During this time, Python uses thepy_compile
module to compile the code of your Python application and itscalc
library into byte code, which are stored in thesources
workspace directory (within the Jenkins home directory).
The Blue Ocean interface turns green if Jenkins compiled your Python application successfully.
-
-
Click the X at the top-right to return to the main Blue Ocean interface.
Add a test stage to your Pipeline
-
Go back to your text editor/IDE and ensure your
Jenkinsfile
is open. -
Copy and paste the following Declarative Pipeline syntax immediately under the
Build
stage of yourJenkinsfile
:stage('Test') { agent { docker { image 'qnib/pytest' } } steps { sh 'py.test --junit-xml test-reports/results.xml sources/test_calc.py' } post { always { junit 'test-reports/results.xml' } } }
so that you end up with:
pipeline { agent none stages { stage('Build') { agent { docker { image 'python:2-alpine' } } steps { sh 'python -m py_compile sources/add2vals.py sources/calc.py' stash(name: 'compiled-results', includes: 'sources/*.py*') } } stage('Test') { (1) agent { docker { image 'qnib/pytest' (2) } } steps { sh 'py.test --junit-xml test-reports/results.xml sources/test_calc.py' (3) } post { always { junit 'test-reports/results.xml' (4) } } } } }
1 Defines a stage
(directive) calledTest
that appears on the Jenkins UI.2 This image
parameter (of theagent
section’sdocker
parameter) downloads theqnib:pytest
Docker image (if it’s not already available on your machine) and runs this image as a separate container. This means that:-
You’ll have separate Jenkins and pytest containers running locally in Docker.
-
The pytest container becomes the agent that Jenkins uses to run the
Test
stage of your Pipeline project. This container’s lifespan lasts the duration of yourTest
stage’s execution.
3 This sh
step (of thesteps
section) executes pytest’spy.test
command onsources/test_calc.py
, which runs a set of unit tests (defined intest_calc.py
) on the "calc" library’sadd2
function (used by your simple Python applicationadd2vals
). The:-
--junit-xml test-reports/results.xml
option makespy.test
generate a JUnit XML report, which is saved totest-reports/results.xml
(within the/var/jenkins_home/workspace/simple-python-pyinstaller-app
directory in the Jenkins container).
4 This junit
step (provided by the JUnit Plugin) archives the JUnit XML report (generated by thepy.test
command above) and exposes the results through the Jenkins interface. In Blue Ocean, the results are accessible through the Tests page of a Pipeline run. Thepost
section’salways
condition that contains thisjunit
step ensures that the step is always executed at the completion of theTest
stage, regardless of the stage’s outcome. -
-
Save your edited
Jenkinsfile
and commit it to your localsimple-python-pyinstaller-app
Git repository. E.g. Within thesimple-python-pyinstaller-app
directory, run the commands:
git stage .
then
git commit -m "Add 'Test' stage"
-
Go back to Jenkins again, log in again if necessary and ensure you’ve accessed Jenkins’s Blue Ocean interface.
-
Click Run at the top left, then quickly click the OPEN link which appears briefly at the lower-right to see Jenkins running your amended Pipeline project. If you weren’t able to click the OPEN link, click the top row on the Blue Ocean interface to access this feature.
Note: It may take a few minutes for theqnib:pytest
Docker image to download (if this hasn’t already been done).
If your amended Pipeline ran successfully, here’s what the Blue Ocean interface should look like. Notice the additional "Test" stage. You can click on the previous "Build" stage circle to access the output from that stage. -
Click the X at the top-right to return to the main Blue Ocean interface.
Add a final deliver stage to your Pipeline
-
Go back to your text editor/IDE and ensure your
Jenkinsfile
is open. -
Copy and paste the following Declarative Pipeline syntax immediately under the
Test
stage of yourJenkinsfile
:stage('Deliver') { agent any environment { VOLUME = '$(pwd)/sources:/src' IMAGE = 'cdrx/pyinstaller-linux:python2' } steps { dir(path: env.BUILD_ID) { unstash(name: 'compiled-results') sh "docker run --rm -v ${VOLUME} ${IMAGE} 'pyinstaller -F add2vals.py'" } } post { success { archiveArtifacts "${env.BUILD_ID}/sources/dist/add2vals" sh "docker run --rm -v ${VOLUME} ${IMAGE} 'rm -rf build dist'" } } }
and add a
skipStagesAfterUnstable
option so that you end up with:pipeline { agent none options { skipStagesAfterUnstable() } stages { stage('Build') { agent { docker { image 'python:2-alpine' } } steps { sh 'python -m py_compile sources/add2vals.py sources/calc.py' stash(name: 'compiled-results', includes: 'sources/*.py*') } } stage('Test') { agent { docker { image 'qnib/pytest' } } steps { sh 'py.test --junit-xml test-reports/results.xml sources/test_calc.py' } post { always { junit 'test-reports/results.xml' } } } stage('Deliver') { (1) agent any environment { (2) VOLUME = '$(pwd)/sources:/src' IMAGE = 'cdrx/pyinstaller-linux:python2' } steps { dir(path: env.BUILD_ID) { (3) unstash(name: 'compiled-results') (4) sh "docker run --rm -v ${VOLUME} ${IMAGE} 'pyinstaller -F add2vals.py'" (5) } } post { success { archiveArtifacts "${env.BUILD_ID}/sources/dist/add2vals" (6) sh "docker run --rm -v ${VOLUME} ${IMAGE} 'rm -rf build dist'" } } } } }
1 Defines a stage
(directive) calledDeliver
that appears on the Jenkins UI.2 This environment
block defines two variables which will be used later in the 'Deliver' stage.3 This dir
step (of thebasic steps
section) creates a new subdirectory named by the build number. The final program will be created in that directory by pyinstaller.BUILD_ID
is one of the pre-defined Jenkins environment variables and is available in all jobs.4 This unstash
step (of thebasic steps
section) restores the Python source code and compiled byte code files (with.pyc
extension) from the previously saved stash image (if it’s not already available on your machine) and runs this image as a separate container. This means that:-
You’ll have separate Jenkins and PyInstaller (for Linux) containers running locally in Docker.
-
The PyInstaller container becomes the agent that Jenkins uses to run the
Deliver
stage of your Pipeline project. This container’s lifespan lasts the duration of yourDeliver
stage’s execution.
5 This sh
step (of thesteps
section) executes thepyinstaller
command (in the PyInstaller container) on your simple Python application. This bundles youradd2vals.py
Python application into a single standalone executable file (via the--onefile
option) and outputs the this file to thedist
workspace directory (within the Jenkins home directory). Although this step consists of a single command, as a general principle, it’s a good idea to keep your Pipeline code (i.e. theJenkinsfile
) as tidy as possible and place more complex build steps (particularly for stages consisting of 2 or more steps) into separate shell script files like thedeliver.sh
file. This ultimately makes maintaining your Pipeline code easier, especially if your Pipeline gains more complexity.6 This archiveArtifacts
step (provided as part of Jenkins core) archives the standalone executable file (generated by thepyinstaller
command above atdist/add2vals
within the Jenkins home’s workspace directory) and exposes this file through the Jenkins interface. In Blue Ocean, archived artifacts like these are accessible through the Artifacts page of a Pipeline run. Thepost
section’ssuccess
condition that contains thisarchiveArtifacts
step ensures that the step is executed at the completion of theDeliver
stage only if this stage completed successfully. -
-
Save your edited
Jenkinsfile
and commit it to your localsimple-python-pyinstaller-app
Git repository. E.g. Within thesimple-python-pyinstaller-app
directory, run the commands:
git stage .
then
git commit -m "Add 'Deliver' stage"
-
Go back to Jenkins again, log in again if necessary and ensure you’ve accessed Jenkins’s Blue Ocean interface.
-
Click Run at the top left, then quickly click the OPEN link which appears briefly at the lower-right to see Jenkins running your amended Pipeline project. If you weren’t able to click the OPEN link, click the top row on the Blue Ocean interface to access this feature.
Note: It may take a few minutes for thecdrx/pyinstaller-linux
Docker image to download (if this hasn’t already been done).
If your amended Pipeline ran successfully, here’s what the Blue Ocean interface should look like. Notice the additional "Deliver" stage. Click on the previous "Test" and "Build" stage circles to access the outputs from those stages.Here’s what the output of the "Deliver" stage should look like, showing you the results of PyInstaller bundling your Python application into a single standalone executable file.
-
Click the X at the top-right to return to the main Blue Ocean interface, which lists your previous Pipeline runs in reverse chronological order.
Follow up (optional)
If you use Linux, you can try running the standalone add2vals
application you
generated with PyInstaller locally on your machine. To do this:
-
From the main Blue Ocean interface, access your last Pipeline run you performed above. To do this, click the top row (representing the most recent Pipeline run) on the main Blue Ocean’s Activity page.
-
On the results page of the Pipeline run, click Artifacts at the top right to access the Artifacts page.
-
In the list of artifacts, click the down-arrow icon at the far right of the dist/add2vals artifact item to download the standalone executable file to your browser’s "Downloads" directory.
-
Back in your operating system’s terminal prompt,
cd
to your browser’s "Downloads" directory. -
Make the
add2vals
file executable - i.e.chmod a+x add2vals
-
Run the command
./add2vals
and follow the instructions provided by your app.
Wrapping up
Well done! You’ve just used Jenkins to build a simple Python application!
The "Build", "Test" and "Deliver" stages you created above are the basis for building more complex Python applications in Jenkins, as well as Python applications that integrate with other technology stacks.
Because Jenkins is extremely extensible, it can be modified and configured to handle practically any aspect of build orchestration and automation.
To learn more about what Jenkins can do, check out:
-
The Tutorials overview page for other introductory tutorials.
-
The User Handbook for more detailed information about using Jenkins, such as Pipelines (in particular Pipeline syntax) and the Blue Ocean interface.
-
The Jenkins blog for the latest events, other tutorials and updates.