This is a reusable Python Flask template. The project is based on Flask in combination with SQLAlchemy ORM.

Complete list of features the template provides:

• Onion architecture
• Maintenance window support
• SQLAlchemy ORM
• Alembic Database migrations
• Local postgres database docker support
• Tests and test containers integration
• Service prefix
• Dependency injection
• Service-repository design pattern

To start a new project, run the following command:

cookiecutter -c v1 https://github.com/microsoft/cookiecutter-python-flask-clean-architecture

This will prompt you for some information about your project. The information you provide will be used to populate the files in the new project directory.

To run the application locally, you need to have a Postgres database running. You can use the run_postgres.sh script in the scripts directory to run a Postgres container.

./scripts/run_postgres.sh

You can then run the application with flask:

flask --app src/app run 

or with gunicorn:

gunicorn wsgi:app -b  0.0.0.0:7000 --workers=1 --preload

The application follows the Onion Architecture pattern. An article is written about our experience integrating an onion architecture with flask in combination with SQL Alchemy ORM that can be found here.

This architecture is a design pattern that organizes the codebase of a software application into multiple layers, where the innermost layer is the domain layer and the outermost layer is the application layer. Each layer depends only on the layers inside of it and not on the layers outside of it, creating a separation of concerns, allowing for a more maintainable and scalable codebase.

For this template we suggest using a service-repository design pattern. This template also provides a set of abc meta classes that you can use to create your repositories and services. For example implementations you can have a look at Service-repository design pattern.

This template provides you with a maintenance window mode. To learn more about maintenance windows in your service you can read this article here

During maintenance mode, clients will receive an http 503 status code.

You can activate maintenance mode in the following ways:

curl -X PATCH http://localhost:7000/<service-prefix>/v1/service-context -d '{"maintenance": true}' -H 'Content-Type: application/json'

or via the command line:

flask activate_maintenance_mode

You can deactivate maintenance mode in the following ways:

curl -X PATCH http://localhost:7000/<service-prefix>/v1/service-context -d '{"maintenance": false}' -H 'Content-Type: application/json'

or via the command line:

flask deactivate_maintenance_mode

The template uses SQLAlchemy ORM for its database connection and database models integration. Its is currently setup with postgres, however you can change it to any other database that is supported by SQLAlchemy. For other databases have a look at the official Flask SQLAlchemy documentation that can be found here

This template provides you with a model base class that you can use to create your models.

from src.infrastructure.models.model_extension import ModelExtension

class User(ModelExtension):
    __tablename__ = 'users'
    id = Column(Integer, primary_key=True)
    attribute_a = Column(String(50), nullable=False)
    attribute_b = Column(String(50), nullable=False)

    def __repr__(self):
        return self.repr(id=self.id, attribute_a=self.attribute_a, attribute_b=self.attribute_b)

Note: The application uses a postgres database. Make sure you have a postgres database running before running the following commands. For local development, you can use the run_postgres.sh script to run a postgres container locally.

1. Make sure you have the diesel cli installed. You can install it with the following command:

   sh ./scripts/run_postgres.sh

2. Create a database migration

   flask db migrate -m <migration-message>

3. Apply the database migration:

   flask db upgrade

You can run a local postgres docker database by using the following script:

 sh ./scripts/run_postgres.sh

This will run a postgres docker container on port 5432. Also it will create a .env file in the root directory of the project. This file contains the database connection string. The service will read this connection string from the .env file and use it to connect to the database.

All tests are can be found under the tests folder. When using the template you can place all you tests in this folder.

The service uses python unittest in combination with [flask testing]

To run the tests, you can use the following command:

python -m unittest discover -s tests

You can use the test containers library to run your tests against a postgres database. You do this by setup_database() in your test class. This will create a postgres container and run your tests against it. After the tests are done, the container will be destroyed.

If you want to run your tests against a different database, you can change the setyp_database method in the test class to use a different database container.

class Test(AppTestBase):

    def setUp(self) -> None:
        super(Test, self).setUp()
        self.setup_database()

The application can use a service prefix for the endpoints. The service prefix is defined in the config.py file. Given a service prefix of 'example', endpoints will be prefixed with '/example/v1/service-context'.

# config.py
SERVICE_PREFIX = os.environ.get(SERVICE_PREFIX, '')

# middleware
class PrefixMiddleware(object):
    ROUTE_NOT_FOUND_MESSAGE = "This url does not belong to the app."

    def __init__(self, app, prefix=''):
        self.app = app
        self.wsgi_app = app.wsgi_app
        self.prefix = prefix

    def __call__(self, environ, start_response):

        if environ['PATH_INFO'].startswith(self.prefix):
            environ['PATH_INFO'] = environ['PATH_INFO'][len(self.prefix):]
            environ['SCRIPT_NAME'] = self.prefix
            return self.wsgi_app(environ, start_response)
        else:
            start_response('404', [('Content-Type', 'text/plain')])
            return [self.ROUTE_NOT_FOUND_MESSAGE.encode()]

During testing the service prefix is not applied. This allows you to test the endpoints without having to add the service prefix to the endpoint.

If the service prefix is not set, the service will not use a service prefix.

This template uses the dependency_injector library for dependency injection. The template provides you with a container class that you can use to register your dependencies. The container class is located in the src/dependency_container.py file.

You can add your dependencies to the container and use them in your routers, services and repositories.

This template provides you with a repository-service pattern. There are two base classes that you can use to create your repositories and services. These base classes are located in the src/services/repository_service.py and src/infrastructure/repositories/repository.py.

A repository example can be seen below, this repository is used to query the MyModel model. For custom query params support override the _apply_query_params method.

from infrastructure.repositories import Repository
from infrastructure.models import MyModel

class MyExampleRepository(Repository):
    base_class = MyModel
    DEFAULT_NOT_FOUND_MESSAGE = "MyModel was not found"

    def _apply_query_params(self, query, query_params):
        name_query_param = self.get_query_param("name", query_params)
        
        if name_query_param:
           query = query.filter_by(name=name_query_param)
            
        return query

A Service example can be seen below, this service expect a repository to be injected in its contuctor.

from services.repository_service import RepositoryService

class MyExampleService(RepositoryService):
    # The RepositoryService gives you access to crud repository operations by the inheritance 
    # RepositoryService.
    pass

# Can be instantiate by injecting the repository
my_example_service = MyExampleService(my_example_repository)