ECG (electrocardiograms) service

This repository contains the code of a microservice that allows users to register or store ECGs for processing, and then query various insights about them, such as the number of zero crossings of the signal.

Description

The project is built using Python + FastAPI stack, following Hexagonal (also known as Ports and Adapters) architectural style and a simplistic approach of the CQRS pattern.

In the following image (source), we can see a diagram representing the different parts that may for a system following this architecture design:

The directory structure for each feature (ECGs, Insights and Users) is as follows:

• src → main package with all the source code.
  • [feature] → package with all the code related to the feature.
    • [adapter] → implementations of the interfaces (ports) that communicate with the application core. There are two types of adapters:
      • in_adapters → they are implementations of external systems that communicate inwards with the application, such as HTTP, etc.
      • out_adapters → they are implementations of external system that communicate outwards with the application, such as databases, etc.
    • [application] → interfaces and implementation of use cases.
      • [ports] → interfaces and use cases that adapters and services implement for communication with the application core. There are two types as well, in_ports and out_ports, following the same logic as with adapters.
      • [service] → implementation of use cases .
    • [domain] → main domain model and business logic of the application.

Note: To simplify development, only in memory persistence adapters have been implemented. For adding any other solution, it's as simple as implementing a new adapter for it (PostgreSQL, Mongo, Redis, etc.) and changing the dependency in the service that requires it.

Endpoints

There are three main endpoints that the app implements + a health endpoint:

• GET /health/ping → to check if the app is alive.
• POST /users → for registering new users. Example:

  curl --location 'http://0.0.0.0:8000/users' \
  --header 'x-admin-token: secret' \
  --header 'Content-Type: application/json' \
  --data '{
    "username": "user"
  }'
  

• POST /ecgs: → to register a new ECG. Example:

  curl --location 'http://0.0.0.0:8000/ecgs' \
  --header 'x-user-token: <GENERATED_USER_TOKEN>' \
  --header 'Content-Type: application/json' \
  --data '{
      "leads": [
          {
              "name": "I",
              "number_of_samples": 8,
              "signal": [1, 0, -1, 0, 1, 0, 0, 1]
          }
      ]
  }'
  

• GET /ecgs/<ecg-id>/insights → to retrieve insights for an existing ECG. Example:

  curl --location 'http://0.0.0.0:8000/ecgs/<ECG_ID>/insights' \
  --header 'x-user-token: <GENERATED_USER_TOKEN>'
  

The details of the endpoint's contract can also be found at the autogenerated API docs, once the app is deployed.

Running

With Docker

docker build --target deploy -t idoven-backend-challenge .

docker run -p 8000:8000 --rm --name idoven-backend-challenge-local-deploy idoven-backend-challenge

Locally

• Install Python 3.11. You can use pyenv to install the required Python version.

pyenv install 3.11.7

• Set it up as the global version.

pyenv global 3.11.7

• Create the virtual environment with the pyenv virtualenv plugin.

pyenv virtualenv 3.11.7 idoven-venv-3.11.7

• Activate it

pyenv activate idoven-venv-3.11.7

• Install the dependencies

pip install -r requirements.txt

• Launch the app using the built-in uvicorn server

uvicorn src.main:app --port 8000 --reload

or by executing the launch script

chmod u+x run.sh

./run.sh

• When done, you can clean the virtual environment using the following commands

pyenv deactivate

pyenv uninstall idoven-venv-3.11.7

Testing

To get and idea of a full user flow (from user registration to insights consulting), you can consult the test_integration_tests.py file, where there is an integration tests that mimics all the steps that a user would do to perform the mentioned actions.

With Docker

docker build --target run_tests -t idoven-backend-challenge-tests .

docker run --rm --name idoven-backend-challenge-tests-execution idoven-backend-challenge-tests

Locally

After installing dependencies, just run pytest to execute the tests.

pytest -v --cov tests