Development Guide

Anitya welcomes contributions! Our issue tracker is located on GitHub.

Contribution Guidelines

When you make a pull request, someone from the fedora-infra organization will review your code. Please make sure you follow the guidelines below:

Python Support

Anitya supports Python 3.8 or greater so please ensure the code you submit works with these versions. The test suite will run against all supported Python versions to make this easier.

Code Style

We follow the PEP8 style guide for Python. The test suite includes a test that enforces the required style, so all you need to do is run the tests to ensure your code follows the style. If the unit test passes, you are good to go!

To automatically format the code run the following in project root. The .tox folder will be created when tox will be run.

.tox/format/bin/black .

Unit Tests

The test suites can be run using tox by simply running tox from the repository root. These tests include unit tests, a linter to ensure Python code style is correct, checks for possible security issues, and checks the documentation for Sphinx warnings or errors.

All tests must pass. All new code should have 100% test coverage. Any bugfix should be accompanied by one or more unit tests to demonstrate the fix. If you are unsure how to write unit tests for your code, we will be happy to help you during the code review process.

CI (Continuous Integration)

Anitya has a CI set up to run on each PR. As a CI of choice Anitya is using Fedora zuul and the configuration could be found in .zuul.yaml in Anitya root directory.

The CI runs unit tests for all supported python versions, code style test, coverage test, flake8 test (linter), documentation test build and bandit (to check for any security issue).

The successful run of CI is a requirement for merge of the PR.

Documentation

Anitya uses sphinx to create its documentation. New packages, modules, classes, methods, functions, and attributes all should be documented using “Google style” docstrings. For historical reasons you may encounter plain reStructuredText-style docstrings. Please consider converting them and opening a pull request!

Python API documentation is automatically generated from the code using Sphinx’s autodoc extension. HTTP REST API documentation is automatically generated from the code using the httpdomain extension.

Release notes

To add entries to the release notes, create a file in the news directory with the source.type name format, where type is one of:

  • feature: for new features

  • bug: for bug fixes

  • api: for API changes

  • dev: for development-related changes

  • author: for contributor names

  • other: for other changes

And where the source part of the filename is:

  • 42 when the change is described in issue 42

  • PR42 when the change has been implemented in pull request 42, and there is no associated issue

  • username for contributors (author extention). It should be the username part of their commit’s email address.

For example:

If this PR is solving bug 714 the file inside news should be called 714.bug and the content of the file would be:

Javascript error on add project page

Matching the issue title.

The text inside the file will be used as entry text. A preview of the release notes can be generated with towncrier --draft.

Development Environment

There are two options for setting up a development environment. If you’re not sure which one to choose, pick the Vagrant method.

Vagrant

The Vagrant development environment is set up using Ansible.

To get started, install Vagrant and Ansible. On Fedora:

$ sudo dnf install vagrant libvirt vagrant-libvirt vagrant-sshfs ansible

Next, clone the repository and start the Vagrant machine:

$ git clone https://github.com/fedora-infra/anitya.git
$ cd anitya
$ vagrant up
$ vagrant ssh

When you log in you’ll be presented with a message of the day with more details about the environment.

To start the Anitya instance in vagrant you can run:

$ systemctl --user start anitya

You may then access Anitya on your host at:

http://127.0.0.1:5000

By default, Anitya imports the production database so you’ve got something to start with. If instead you prefer an empty database, add the following to the Ansible provisioner inside your Vagrantfile:

ansible.extra_vars = { import_production_database: false }

The application’s configuration file is /home/vagrant/anitya.toml. You can also look at the sample configuration

Note

Please don’t commit any local changes to Vagrantfile. We are managing it upstream.

Vagrant is using PostgreSQL database. To work with it use psql command:

$ sudo -u postgres psql
postgres=#\connect anitya

After this you can use standard SQL queries or another psql commands:

# Show description of tables
\dt
# Show table description
\d users

For additional psql commands see man psql.

To run libraries.io service simply run:

$ librariesio_consumer.py

To run check service simply run:

$ check_service.py

Docker / Podman

This way, you will be able to control each service (anitya-web, anitya-librariesio-consumer. RabbitMQ, etc) separately. You can turn off RabbitMQ or PostgreSQL or both, then connect to external services or use them with the application.

Requirements:

  • Docker / Podman

  • Docker Compose / Podman Compose

Next, clone the repository and start containers:

$ git clone https://github.com/fedora-infra/anitya
$ cd anitya
$ make up
Container Service Informations:

Name 1

Url

Credentials

RabbitMQ

localhost:5672

anitya:anitya

RabbitMQ Management UI

http://localhost:15672

anitya:anitya

PostgreSQL

localhost:5432

postgres:anypasswordworkslocally

Makefile scripts that provide easier container management:

  • make up Starts all the container services

  • make restart Restarts all the container services that are either stopped or running

  • make halt Stops and removes the containers

  • make bash-web Connects to anitya-web container

  • make bash-consumer Connects to anitya-librariesio-consumer container

  • make init-db Creates database

  • make dump-restore Import production database

  • make logs Shows all logs of all containers

  • make clean Removes all images used by Anitya compose

Project files are bound to each other with host and container. Whenever you change any project file from the host or the container, the same change will happen on the opposite side as well.

Start the development web server included with Flask with:

$ make bash-web
$ FLASK_APP=anitya.wsgi flask run

Start the libraries.io service with:

$ make bash-consumer
$ librariesio_consumer.py

Start the check service with:

$ make bash-consumer or make-bash-web
$ check_service.py

Python virtualenv

Anitya can also be run in a Python virtualenv. For Fedora:

$ git clone https://github.com/fedora-infra/anitya.git
$ cd anitya
$ sudo dnf install python3-virtualenvwrapper
$ source /usr/bin/virtualenvwrapper.sh
$ mkvirtualenv anitya
$ workon anitya

Issuing that last command should change your prompt to indicate that you are operating in an active virtualenv.

Next, install Anitya:

(anitya)$ pip install -r test_requirements.txt
(anitya)$ pip install -e .

Create the database, by default it will be a sqlite database located at /var/tmp/anitya-dev.sqlite:

(anitya) $ python createdb.py

You can start the development web server included with Flask with:

(anitya)$ FLASK_APP=anitya.wsgi flask run

If you want to change the application’s configuration, create a valid configuration file and start the application with the ANITYA_WEB_CONFIG environment variable set to the configuration file’s path. You can look at the sample configuration for guidance.

Release Guide

Testing before release

To test the new version before release just update the staging branch to current master:

git checkout staging
git rebase master
git push origin/staging

This will automatically start the deployment in staging instance. You can then test the new changes there.

If you need to do any changes in configuration of staging instance, just update the release-monitoring role in Fedora infra ansible repository.

If the changes are merged, you can run the playbook by following configuration guide for Anitya in Fedora infra documentation.

Note

Have in mind that everything needs to be only done for staging. In configuration use jinja statements and when deploying don’t forget to use -l staging switch.

Anitya

To do the release you need following python packages installed:

wheel
twine
towncrier

If you are a maintainer and wish to make a release, follow these steps:

  1. Change the version in anitya.__init__.__version__. This is used to set the version in the documentation project and the setup.py file.

  2. Add any missing news fragments to the news folder.

  3. Get authors of commits by python get-authors.py.

Note

This script must be executed in news folder, because it creates files in current working directory.

  1. Generate the changelog by running towncrier.

Note

If you added any news fragment in the previous step, you might see towncrier complaining about removing them, because they are not committed in git. Just ignore this and remove all of them manually; release notes will be generated anyway.

  1. Remove every remaining news fragment from news folder.

  2. Generate new DB schema image by running ./generate_db_schema in docs folder.

  3. Commit your changes with message Anitya <version>.

  4. Tag a release with git tag -s <version>.

  5. Don’t forget to git push --tags.

  6. Sometimes you need to also do git push.

  7. Build the Python packages with python setup.py sdist bdist_wheel.

  8. Upload the packages with twine upload dist/<dists>.