Plan My Hikes

Server taken down due to cost

About

I have been using Chat GPT when planning logistics for hiking national parks. Since national parks don't change much, I query'd & saved hiking advice from Chat GPT for over 100+ national parks around the world.

OpenAI Documentation

I wrote documentation to query OpenAI & save results to JSON, CSV, and Postgres DB. The model I queried was Chat GPT 3.5

Deployment

GitHub Actions + AWS Elastic Beanstalk + Docker

Deploy occurs on every merged pull request on the main branch. I send my zipped docker-compose.yml file to AWS & it builds my public Docker images. The docker-compose.yml file is set as the deployment_package in my ./.github/worfklows/main.yml. This is used by the Beanstalk Deploy action.

Nginx

1. I added a Nginx server to my React Frontend Container. This listens on port 3000 & serves production assets from npm run build.

2. I forced HTTPS over HTTP in my default.conf for my application Nginx server.

React + Express + PostgresDB (Update: I removed Postgres due to AWS RDS costs)

I implemented Semantic React UI on client side + an Express/PostgresDB API.

Commands

Build multi-container application locally (application will be available locally on http://localhost:81)

docker compose -f docker-compose-dev.yml up --build

SSH onto EC2 instance

ssh - ./Path/To/EC2/SSH/Key/Pair ec2-user@{Your_EC2_Public_IPv4_DNS_Here}

Drop DB

docker exec -it {backend_container_name_here} node seed/drop.js

Seed DB (reads from ./data/json/parks.json file that contains saved Chat GPT responses)

docker exec -it {backend_container_name_here} node seed/seed.js