TIA ETL

ETL data pipeline project to ingest data from TIA Public API.

DAGS

• posts_pipeline
  • runs hourly to ingest 30 latest posts
• comments_pipeline
  • runs daily at 1AM to ingest yesterday posts comments

Development Setup

Requirements

• Docker and Docker-Compose
  • have 2GB RAM allocated for the VM and 2GB free disk space to run the services.

Setting Up The Project

1. Clone or download this repository to your machine.

2. Open your terminal/console and make sure docker and docker-compose commands are executable.

3. Go to this repository root directory in using your terminal.

4. Execute docker-compose build to build airflow, postgres, adminer as docker services.

   The command will consume a lot of network data, so prepare for the internet. If there is any error, try to execute command several times until it succeed. if it's no luck, raise an issue.

5. After all the service built, execute docker-compose up -d to start the services.

   You can inspect the running services by executing docker-compose ps to see if the services are up or failing. if one of the service is failing, you may want to inspect it using docker-compose logs <service_name> (eg. docker-compose logs airflow). another issue that may happen are the existing services with the same port used by our services. if that happens, try to stop the existing service(s) first and then start our project services.

6. After all the service running, you can start to open the Airflow UI at http://localhost:8080 to make sure the service really went up.

Start/Pause DAGs

1. Open Airflow UI at http://localhost:8080.
2. Login using admin as username and admin as password.
3. You will redirected to home page, you can find our paused DAGs here. Let's start with the posts_pipeline DAG.

4. Click on the posts_pipeline title to enter the DAG page or open http://localhost:8080/tree?dag_id=posts_pipeline.

5. On the post pipeline DAG page, you can click the blue switch to start the DAG.

   another thing you may want to check in this page is to see the Tree View and Graph View to check if the tasks are running, retrying or failed. you may want to check the DAG code by clicking the Code tab.

6. After all inspection done, you can pause the DAG by switch it back to off or leave it run on schedule.

7. You can do the same for the comments_pipeline

Monitoring The Data Using Adminer

1. Open Adminer web UI at localhost:32767.
2. Login using airflow as username and airflow as password. make sure use PostgreSQL as system, postgres as server, and airflow_db as database.

3. After logged in, you can start to monitor the posts data ingested by choosing DB: airflow_db, Schema: public and click on the posts table.

Modify DAGs

You can start to edit any availabe dags under docker/airflow/dags folder by using any code editor you prefered. the Docker volume system will map the changes to the associated dags in the container.

You may want to change any DAG schedule by changing the cron expression in the dag schedule_interval parameter.

After you modified the dags file, you can inspect it in the edited dag page on Airflow UI and see the changes applied on Code tab.

Test DAGs

You can test DAGs pipeline using the command:

docker-compose exec airflow airflow dags test <dag_id> <desired_execution_date>

Here is the example to test the posts_pipeline dag:

docker-compose exec airflow airflow dags test comments_pipeline 2022-01-01

and here is the example to test the comments_pipeline dag:

docker-compose exec airflow airflow dags test comments_pipeline 2022-01-01T01:00:00

Test DAGs Task

When you are developing a task, you may want to test it individually using the command:

docker-compose exec airflow airflow tasks test <dag_id> <task_id> <desired_execution_date>

Here is the example of testing one of posts_pipeline task:

docker-compose exec airflow airflow tasks test posts_pipeline is_tia_public_api_accessible 2022-01-01

here is the example of testing one of comments_pipeline task:

docker-compose exec airflow airflow tasks test comments_pipeline is_postgres_accessible 2022-01-01T01:00:00 

Stoping The Services

To stop all services, you can use the docker-compose command:

docker-compose down