This repo is a POC of Airflow and DBT for ELT pipelines on a CI environment.

The goal is a personal evaluation of Airflow and Dbt, exploring the following topics:

• Setup airflow for a ci environment, how to develop and execute locally.
• Implement custom operators/hooks, etc.
• Testing Airflow dags, operators, hooks, using a TDD Approach.
• Reusability of custom components.
• Error handling in pipelines
• ELT Comparison between airflow and airflow+dbt.
• Testing DBT (besides dbt testing), for instance: testing if a view has the information that should.
• How Dbt handles incremental views

The simple data pipeline example:

Read data from https://covid19api.com/, stage the data on a covid19 table, and transform the data into covid19_stats which has the stats per day per country.

Install local dev environment

Install virtual env

pip install virtualenv

Create env

virtualenv .venv

Activate env

source .venv/bin/activate

Install the same requirements as the docker image

pip install -r requirements-docker.txt

Install requirements for development

pip install -r requirements-dev.txt

Init docker

docker-compose up

Init Environment

Local development, testing and airflow local execution

source dev-env.sh

Test

python -m pytest

Run backfills

airflow backfill covid19_dbt -s 2020-10-02 -e 2020-10-02

airflow backfill covid19 -s 2020-10-02 -e 2020-10-02

Test Dag Tasks

airflow test covid19_dbt dbt_run 2020-10-02

airflow test covid19_dbt dbt_test 2020-10-02