vrymel/komyut

A pathfinding and jeepney routes listing project.

About

This is a personal project that involves me learning Elixir and Dijkstra's path-finding algorithm.

There are two main features right now:

• Jeepney route search - users can specify two points within the city, and the it will return the shortest jeepney route(s) to take.
• Jeepney route listing - users can view jeepney routes within the city.

Jeepney Route

In the Philippines, a Jeepney is a commonly used public transport. They have defined routes within the city so people can access different areas within the city by taking a single or multiple routes of a jeepney.

Screenshots

Jeepney route search

Jeepney route listing

Project Structure

The application code is setup to run directly on the host machine. The database layer is the only component hosted in docker for ease of installation.

Requirements

The following are the requirements for the application environment. See user_data.txt for the installation commands.

1. Erlang and Elixir
2. Node 6.12.x
3. Docker and Docker Compose (only used for the database)
4. Mix hex/phoenix
5. PostgreSQL (as a docker image)

Deployment

Deploying in a new box

1. Clone the application

2. Create the following directories and .env file:

   • ./database_docker_volume
   • ./database_dumps
   • ./.env

3. (When needed) If there's a database dump for the application to use, make sure to uncomment docker-entrypoint-initdb.d in docker-compose.yml so the application will be preloaded with the dump database in ./database_dumps.

   This should be done only for the initial docker image build so make sure to comment the docker-entrypoint-initdb.d again on succeeding deployments.

4. Run the database containers

   docker-compose up -d

5. Install phoenix dependencies

   ./run_app_init.sh

6. Run the server

   ./run_phxserver.py --detached

   This helper script just retrieves the environment variables defined in .env and prepend it as flags when running mix phx.server. So it is important that .env's is targeting development or production.

   If the application can't connect to the database layer. Run this command with sudo.

Succeeding deployments

Since the application code is running locally. Updating is as easy as pulling from the remote repo.

git fetch --all && git pull origin master

If replacing the database do:

1. Kill the docker database container.

   docker-compose down

2. Delete and recreate directory database_docker_volume

3. Repeat step 3 in the deployment steps.