Skip to content

Development Tips

This page is dedicated to providing further context and information around specific implementation details such as code organization, running migrations and more.

If you're looking for a general getting started guide, please see the Overview and Development Guide.

API Endpoints


API URLs

We define API URLs for specific API versions as constants within fides.common.api.v1.urn_registry (where v1 can be substituted for that particular API version), then import those URLs into their specific API views. Since we are on the first version, there is no clear precedent set for overriding URLs between versions yet. The most likely change is that we'll override the APIRouter class instantiation with a different base path (ie. /api/v2 instead of /api/v1). For example:

1
2
PRIVACY_REQUEST = "/privacy-request"
PRIVACY_REQUEST_DETAIL = "/privacy-request/{privacy_request_id}"

would both resolve as /api/v1/privacy-request and /api/v1/privacy-request/{privacy_request_id} respectively.

Database and Models


The ORM -- SQLAlchemy

SQLAlchemy is an Object Relational Mapper, allowing us to avoid writing direct database queries within our codebase, and access the database via Python code instead. The ORM provides an additional configuration layer allowing user-defined Python classes to be mapped to database tables and other constructs, as well as an object persistence mechanism known as the Session. Some common uses cases are listed below, for a more comprehensive guide see: https://docs.sqlalchemy.org/en/14/tutorial/index.html

Adding models

Database tables are defined with model classes. Model files should live in src/fides/api/models/. Individual model classes must inherit from our custom base class at fides.api.db.base_class.Base to ensure uniformity within the database. Multiple models per file are encouraged so long as they fit the same logical delineation within the project. An example model declaration is added below. For a comprehensive guide see: https://docs.sqlalchemy.org/en/14/orm/mapping_styles.html#declarative-mapping You should also import your model in src/fides/api/db/base.py so it is visible for alembic.

1
2
3
4
5
6
7
class Book(Base):
    __tablename__ = 'book'

    id = Column(Integer, primary_key=True)
    name = Column(String, index=True)
    page_count = Column(Integer, nullable=True)
    author_id = Column(Integer, ForeignKey("author.id"), nullable=False)

When models are added to the project, we must then add them to the database in a recordable and repeatable fashion using migrations.

Using the database via models

Once you've added database tables via project models, you're ready to read, write, and update them via Python code. Some examples of common use cases here are listed below. Official documentation is here: https://docs.sqlalchemy.org/en/14/orm/query.html#sqlalchemy.orm.Query.

  • Import our application's database session: from fides.api.db.session import get_db_session
  • Instantiate the database interaction object:
1
2
SessionLocal = get_db_session(config)
db = SessionLocal()
  • Create a new row in a table:
1
2
3
4
5
6
7
8
9
db_obj = User(
    email="admin@fides.app",
    full_name="Fides Admin",
    is_superuser=True,
    is_active=True,
)
db.add(db_obj)
db.commit()
db.refresh(db_obj)
  • Fetch all objects in a table: users = db.query(User).all()
  • Fetch all objects in a table that meet some criteria: active_users = db.query(User).filter(User.is_active == True)
  • Get a specific row in a table: user = db.query(User).get(User.email == "admin@fides.app")
  • Update a specific row in a table:
1
2
3
4
user.email = "updated@fides.app"
db.add(user)
db.commit()
db.refresh()

Connecting to the database

When you run nox -s dev, the database will spin up in a Docker container with port 5432 exposed on localhost. You can connect to it using the credentials found in .fides.toml, e.g.

  • Hostname: localhost
  • Port: 5432
  • Username: see database.user in .fides.toml
  • Password: see database.password in .fides.toml

Alembic migrations

Some common Alembic commands are listed below. For a comprehensive guide see: https://alembic.sqlalchemy.org/en/latest/tutorial.html.

The commands will need to be run inside a shell on your Docker containers, which can be opened with nox -s dev -- shell.

In the /src/fides/api/alembic directory:

  • Migrate your database to the latest state: alembic upgrade head
  • Merge heads (for when you have conflicting heads from a merge/rebase): alembic merge heads
  • Get revision id of previous migration: alembic current
  • Automatically generate a new migration: alembic revision --autogenerate -m "<a message describing your changes>"
  • Create a new migration file to manually fill out: alembic revision -m "<a message describing your changes>"
  • Migrate your database to a specific state alembic upgrade <revision-id> or alembic downgrade <revision-id>, (or if you want to be smart alembic upgrade <revision-id> || alembic downgrade <revision-id> is handy when you don't know whether the target revision is an upgrade or downgrade)
  • NB. You can find the revision-id inside each migration file in alembic/versions/ on line 3 next to Revision ID: ...

When working on a PR with a migration, ensure that down_revision in the generated migration file correctly references the previous migration before submitting/merging the PR.

Exception Handling

Our preference for exception handling is by overriding the nearest sensible error, for example:

1
2
3
4
5
6
class SomeException(ValueError):
    "a docstring"


def some_method():
    raise SomeException("a message")

General debugging -- pdb

The project uses pdb for debugging as a dev-requirement. You can set breakpoints with pdb in much the same way you'd set them using debugger in JavaScript. Insert import pdb; pdb.set_trace() into the line where you want the breakpoint to set, then run your Python code.

Docker

As a last resort you may need to tear everything down in Docker and rebuild. The following commands will achieve that, but be warned that rebuild times can be long!

1
nox -s clean

Warning

If you find yourself feeling the need to run this command regularly, open an issue or slack a member of the dev team as it is not expected that this will need to be run regularly.

Performance and Benchmarking

The following are a few options we have for monitoring and benchmarking application performance:

  1. docker stats - Running this command will show you the CPU and Memory usage of your containers. This is very handy for quickly checking the memory footprint of an image while running.
  2. drill - This is a CLI tool used for load-testing applications. It requires Rust and OpenSSL. This is used in CI to continually benchmark performance.
  3. nox -s performance_tests - This is a convenient combination of the prior two tools that is used in CI but also possible to use locally.
  4. profile-request - When in dev_mode, Adding profile-request: true to any request header will tell the server to profile the request and send back a text response containing the profile data, instead of returning the typical JSON response. This allows arbitrary profiling of any endpoint or feature.
Back to top