Skip to content

Developing Fides

The primary developer interface for Fides is via a tool called Nox. Nox is a pure-Python replacement for tools like make. You can read more about Nox in the official documentation here.

⚠️ Warning
If you install nox using pipx, you may run into issues with it not detecting your Python version correctly. Install it using pip instead.

Additionally, much of what nox helps abstract is related to Docker. This usually makes it possible to troubleshoot potential nox issues by using various Docker commands directly.

If you haven't already, read the Requirements Installation Guide to get up and running with the required tools and optimal configurations.

Terminology

For the purposes of this documentation, there is some Nox-specific terminology that is used and is helpful to understand before reading the rest of the guide.

  • session - This is the equivalent of a single command, or a target in Make. They can be chained together arbitrarily, and the order of execution will be preserved. For example nox -s isort black
  • posarg - This refers to a Positional Argument. These can be provided to any Nox command by putting -- at the end of a command followed by the argument that you would like to provide. For instance nox -s dev -- shell
  • param - This is a specified value that is provided directly to a session. For instance nox -s "isort(<param>)". Note that due to terminal limitations, the session + param must be wrapped in quotes for proper escaping. Additionally, if a session has multiple params but none are specific, all permutations of that session will be run.

Sessions to Know

While there are many sessions available in our nox setup, these are the ones that you are most likely to use regularly. This is intended as a brief introduction, as many of these commands will be explained further in specific sections below.

Command Description
nox This is the default session that runs when no other session is provided. Automatically opens this page.
nox -l Provides a list of all available nox sessions.
nox -s usage -- <command> Shows the full docstring for a single session. The name of the session is provided as a positional argument. Use this whenever you want to know more about a specific session's usage.
nox -s dev Spins up the Fides webserver with all required infrastructure.
nox -s shell Opens an interactive shell in a running Fides webserver container.
nox -s build Builds and tags docker containers.
nox -s clean Completely wipes out all Docker volumes, images, and caches across the entire machine. Note: This should only be used as a last resort! If you find yourself running this command often, there is probably an issue that needs to be addressed.
nox -s teardown Spins down all of the Docker containers for the Fides project. Including the -- volumes posarg will also delete the application database data.

Getting Started

This section will run through the basics of getting everything up and running so that you can start making changes.

Docker

Given that the majority of commands rely on Docker images, the first step is going to be getting some Docker images built. This is handled by the build session. As an excercise, let's check the docs for the build session via the usage session:

1
nox -s usage -- build

The simplest way to build everything is to run nox -s build. This will build all available Docker images used for development and testing. Additionally, to build only a specific image(s), run nox -s usage -- build to see what params have been documented.

Running the Webserver

Now that all of the potential images we'll need are built, it's time to spin up the webserver:

nox -s dev

Depending on your computer, this will take anywhere from ~30 seconds to ~2 minutes. The webserver will log all of its activity to that window which is useful for debugging and checking potential server-side errors.

Warning

When running nox -s dev, if you get an importlib.metadata.PackageNotFoundError: fides, it means that fides was not properly installed. Run nox -s dev -- shell, and then run pip install -e .

You can verify Fides is installed with pip list | grep fides

Opening a Shell

Now that the webserver is up and running, a convenient way to interact with it as well as the Fides CLI is to open a shell directly into the running webserver container. Open a new terminal and run the following:

nox -s shell

With the shell open, we'll need to authenticate before we can start using the CLI. Run the following to authenticate the shell and get ready to run the CLI:

fides user login

With that done, we can do things like evaluations:

fides evaluate

Listing all resources of a type:

fides ls system

and more. Run fides -h to see the full list of commands offered by the CLI.

Checking Your Changes

Once you've made some changes, it's time to verify that they are working as expected.

If you've made changes to the CLI, you can verify them directly using the shell method described here.

If you've made an API change, there are a few different ways to verify your changes. The first and most simple is to go to localhost:8080/docs and verify that the changes you made show up as expected in the documentation. This is most useful when adding a new endpoint or changing a schema.

To dig in even deeper, you can use the Authorize button at the top of the API docs page to login (credentials are available in the .fides/fides.toml config file). This will allow you to send requests, view responses, etc.

Running Tests

For a full walkthrough on testing, head over to the Testing page.

Static Checks

For the sake of simplicity, all of the typical pre-commit static checks are packaged into a single session:

nox -s static_checks

This installs and runs the various checks in their own virtual environment for maximum consistency. For more information about code style in Fides, check Code Style.

Debugging Fides in VSCode

This is a quick guide to show how VSCode can be used to debug Fides running locally in Docker. The general approach is to allow the local fides Docker Compose service to allow remote debugging connections, and to start a remote debugger from a host VSCode workspace.

Setup

Run Fides with Remote Debugging Enabled

In order to accept incoming remote debugging connections, the fides Docker Compose service must be run with slight alterations. To enable this functionality, simply add the remote_debug flag to a nox command. For example:

nox -s dev -- remote_debug

or

nox -s dev -- remote_debug postgres timescale

With those commands, the fides Docker Compose service that's running the Fides server locally is able to accept incoming remote debugging connections.

Note that, at this point, the remote_debug flag is not enabled for other nox sessions, e.g. fides_env, pytest_ops, etc.

Attach a Remote Debugger to the Fides Server

Now that the running Fides server can accept incoming remote debugging connections, you can attach a remote debugger from a local VSCode workspace to actively debug the server application. A launch configuration is included in the fides repo to facilitate this step.

  • Open up the fides repo in a VSCode workspace
  • Go to the Run and Debug view
  • From the debugger dropdown list, select the Python debugger: Remote Attach Fides configuration
  • Click the Start Debugging play button
  • The remote debugger should now be attached to the Fides server!
  • To confirm the debugger is attached, at least one RUNNING line item should appear in the CALL STACK window

Debug!

At this point, VSCode is ready to debug the running Fides server. Try setting breakpoints and hitting them by, e.g., making certain HTTP requests against the Fides server. This guide provides more information on how to use the VSCode Python debugger.

Some relevant VSCode documentation for reference:

Debugging Fides in IntelliJ IDEA Ultimate

This guide will show how to use the IntelliJ debugger with Fides running in Docker. The setup for PyCharm Professional should be very similar.

Prerequisites

Setup

Connect to Docker daemon

This step will allow the IDE to connect to Docker Desktop.

Go to: Settings/Preferences -> Docker -> +

  • Select Docker for "your operating system"

See the screenshot below:

Screenshot of IDE Docker setup

Configure Python Remote Interpreter

Define a Docker-based remote interpreter.

Go to: File -> Project Structure... -> Platform Settings -> SDKs -> +

  • Set Server to Docker
  • Set Configuration files to .docker-compose.yml
  • Set Python interpreter path to python

After clicking OK the Remote Python Docker Compose should be listed as an SDK.

See screenshots below:

Screenshot of Add Python Interpreter

Screenshot of Project Structure SDKs

Run/Debug Configuration

Set up a Run/Debug Configuration so that breakpoints can be hit in the f sourcecode.

Go to: Run/Debug Configurations -> + -> Python

  • To debug Fides, debug the <path on your machine>/src/fides/main.py script
  • Make sure to select Use specified interpreter set the Remote Python Docker Compose (created in the previous section)
  • Add FIDES__CONFIG_PATH=/fides to Environment variables

See screenshot below:

Screenshot of Run/Debug Configuration for main.py

Hit a Breakpoint

Now the IDE is ready to debug the source code. Click the debug button for main (setup in the previous section).

Try firing a http request to Fides from Postman or Curl and hit a break point.

There is a postman collection in this repo: docs/fides/docs/development/postman/Fides.postman_collection.json

Screenshot of hit breakpoint below:

Screenshot of Debugging from IntelliJ

The information is this guide is largely based on these docs

Back to top