Setup

Run on local machine

• Install pip3 and install needed dependency packages

pip3 install botocore boto3 dijkstar

• Install docker and run docker command to install dynamoDB locally then run the container before running the bash script.

docker run -p 8000:8000 amazon/dynamodb-local  

• Run bash script create-local-db.sh to create the dynamoDB tables airport and vehicle and populate them

./create-local-db.sh

• Check dynamoDB tables

aws dynamodb list-tables --endpoint-url http://localhost:8000
aws dynamodb describe-table --table-name airport --endpoint-url http://localhost:8000

• Inside lambda folder run bash script run.sh to run the lambda

./run.sh

• Example endpoints
  • http://localhost:8080/airport
  • http://localhost:8080/airport/AMS
  • http://localhost:8080/airport/AMS/to/LIS
  • http://localhost:8080/vehicle/2/100

Terraform with AWS

• Install openapi-merger for combining OpenAPI specifications (Inside openapi folder)

npm install openapi-merger -g 
openapi-merger -i main.yaml -o deploy-api.yaml

• Install Terraform with brew

brew tap hashicorp/tap
brew install hashicorp/tap/terraform

• Terraform commands. Run inside terraform folder

terraform init (initializes the current directory)
terraform init -upgrade (upgrade to latest acceptable provider version)
terraform validate (command validates the configuration files in a directory)
terraform fmt (format your configuration files into a canonical format and style)
terraform plan (dry run to see what Terraform will do)
terraform refresh (refreshes the state file)
terraform output (views Terraform outputs)
terraform apply (applies the Terraform code and builds stuff)
terraform graph (creates a DOT-formatted graph)
terraform destroy --target aws_lambda_function.lambda (destroys indivifual resource)
terraform destroy  (destroys what has been built by Terraform)

• Populate AWS DynamoDB after terraform apply

./update-db.sh

• AWS lambda sample test JSON

{
  "resource": "/vehicle/{people}/{distance}",
  "path": "/vehicle/2/100",
  "httpMethod": "GET",
  "queryStringParameters": {},
  "multiValueQueryStringParameters": {},
  "pathParameters": {
    "people": "2",
    "distance": "100"
  },
  "stageVariables": "None",
  "body": "None",
  "isBase64Encoded": "FALSE"
}

Github actions

• .github/workflow/terraform.yaml is the workflow file
• Create new IAM user for github actions and get the credentials.
• Add new New repository secrets (AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, BUCKET_TF_STATE) in Actions secrets of the Github repository.

Current Shortcomings

• Github actions: Can do terraform apply locally for the AWS but github action terrafom apply didn't trigger on merge/push to the main branch.
• API gateway: Manual steps involved.
  AWS Labmda is working (tested with above sample JSON). But to test API gateway Method execution, Integration Request function name needed to be manually updated for each GET.

Execute the 'Deploy API' button for the root resource ('/') from the AWS API gateway console.

• Public URL: The public URL throws Internal server error or Endpoint request timed out for some time and then it works (Need x-api-key which we can get this from Amazon API gateway -> API keys)

curl --location --request GET 'https://8v5w5mv4ze.execute-api.us-east-1.amazonaws.com/holiday-stage/vehicle/2/100' \
--header 'Content-Type: application/json' \
--header 'x-api-key: <x-api-key here>'

Reference

• Github actions
  • Managing Terraform with GitHub Actions
  • How to use GitHub Actions to automate Terraform
• Terraform
  • Openapi tf example
  • aws api-gateway sample
  • cliffdias/HelloworldAPI

Holiday Agency

Overview

A holiday agency would like to suggest the lowest travel cost for holiday journeys to their customers.
A return journey consists of the following parts:

1. Journey to the airport:

   • Taxi: £0.40/mile - (allows up to 4 people per taxi - e.g. require 2 taxis if 5 people travel together)
   • Car: £0.20/mile - (allows up to 4 people per car), additional £3 parking fee

2. An outbound and an inbound flight journey:

   • Flight: £0.10/passenger/mile

This flight agency keeps a table with its airports and their connections:

Airport ID	Airport Name	Latitude	Longitude	Connections and distance
IST	Istanbul Airport	41.262222	28.727778	ATH: 100mi AMS: 1200mi SVO: 800mi VIE: 750mi
CDG	Charles de Gaulle Airport	49.009722	2.547778	ORY: 10mi LHR: 100mi AMS: 120mi FRA: 130mi ZRH: 120mi FCO: 500mi LIS: 750mi MAD: 600mi
LHR	Heathrow Airport	51.4775	-0.461389	LGW: 10mi CDG: 50mi AMS: 100mi FRA: 230mi MAD: 500mi LIS: 350mi
AMS	Schiphol Airport	52.308056	4.764167	LHR: 100mi FRA: 90mi CDG: 110mi OSL: 500mi
SVO	Sheremetyevo International Airport	55.972778	37.414722	VKO: 10mi DME: 15mi LED: 200mi ATH: 1300mi
FRA	Frankfurt Airport	50.033333	8.570556	VIE: 300mi MUC: 90mi ZRH: 100mi CDG: 105mi LHR: 120mi AMS: 100mi DME: 1500mi
MAD	Adolfo Suárez Madrid-Barajas Airport	40.472222	-3.560833	LIS: 145mi FCO: 220mi LHR: 500mi BCN: 120mi
DME	Moscow Domodedovo Airport	55.408611	37.906111	ATH: 1105mi OSL: 980mi VKO: 15mi SVO: 20mi LED: 210mi ATH: 1000mi
BCN	Josep Tarradellas Barcelona-El Prat Airport	41.296944	2.078333	MAD: 50mi FCO: 140mi
VKO	Vnukovo International Airport	55.596111	37.2675	SVO: 5mi DME: 8mi
MUC	Munich Airport	48.353889	11.786111	VIE: 90mi ZRH: 80mi FRA: 100mi FCO: 200mi
LED	Pulkovo Airport	59.800278	30.2625	VKO: 190mi SVO: 170mi OSL: 300mi LIS: 2500mi
ORY	Orly Airport	48.723333	2.379444	CDG: 5mi AMS: 150mi
LGW	Gatwick Airport	51.148056	-0.190278	LHR: 10mi ORY: 125mi
FCO	Leonardo da Vinci-Fiumicino Airport	41.800278	12.238889	ATH: 200mi BCN: 190mi IST: 310mi AMS: 495mi
LIS	Lisbon Airport	38.774167	-9.134167	MAD: 100mi BCN: 300mi LHR: 300mi SVO: 2100mi
OSL	Oslo Airport - Gardermoen	60.202778	11.083889	LED: 320mi LHR: 390mi FRA: 720mi AMS: 690mi VKO: 980mi
ZRH	Zurich Airport	47.464722	8.549167	FCO: 150mi MUC: 85mi VIE: 190mi FRA: 100mi CDG: 180mi AMS: 310mi
ATH	Athens International Airport	37.936389	23.947222	IST: 150mi FCO: 200mi MAD: 500mi
VIE	Vienna International Airport	48.110833	16.570833	IST: 350mi VKO: 890mi OSL: 880mi MUC: 95mi FCO: 200mi

While the route A->B can exist, it doesn't necessarily mean that B->A exists. Additionally, the mileage may be different for each direction.

As a side-note, some of the mileages might not be at all accurate but imagine that someone make some fairly large one-way portals in the sky that can shorten the distance between two airports.

The current logic allows for the following endpoints:

• GET: /airports - returns a list of airports
• GET: /airports/{airportId} - return details of a single airport
• GET: /airports/{airportId}/to/{airportToId} - returns the cheapest quote for a journey from airportId to airportToId
• GET: /vehicle/{numberOfPeople}/{distance} - returns the cheapest vehicle to use (and the cost) for the given numberOfPeople and distance in miles