cdn/

README

Contents Delivery Network for Map Tiles

Here is a schematic diagram of the service.

The following instructions explain how to configure a map vector tile API step by step. These steps are integrated into an AWS CodePipeline that is described in continuous-delivery.md.

Map Tile Generator

A map tile generator is a Lambda function on AWS.

A Lambda function accepts the following parameters,

• zoom: zoom level of a tile
• x: x-position of a tile
• y: y-position of a tile

Building API

Prerequisites

You need the following software installed,

• AWS CLI
• AWS SAM

Creating an S3 bucket for code

1. Deploy api/code-bucket-template.yaml.

   aws cloudformation deploy --template-file api/code-bucket-template.yaml --stack-name imaginary-map-code-bucket
   

   You have to specify an appropriate credential.

2. Remember the S3 bucket name as CODE_REPOSITORY.

   CODE_REPOSITORY=`aws --query "Stacks[0].Outputs[?OutputKey=='CodeRepositoryName']|[0].OutputValue" cloudformation describe-stacks --stack-name imaginary-map-code-bucket | sed 's/^"//; s/"$//'`
   

   You have to specify an appropriate credential.

Creating an S3 bucket for GeoJSON files

1. Deploy api/geo-json-bucket-template.yaml.

   aws cloudformation deploy --template-file api/geo-json-bucket-template.yaml --stack-name imaginary-map-geo-json-bucket
   

   You have to specify an appropriate credential.

2. Remember the S3 bucket name for GeoJSON files as GEO_JSON_BUCKET.

   GEO_JSON_BUCKET=`aws --query "Stacks[0].Outputs[?OutputKey=='GeoJsonBucketName']|[0].OutputValue" cloudformation describe-stacks --stack-name imaginary-map-geo-json-bucket | sed 's/^"//; s/"$//'`
   

   You have to specify an appropriate credential.

Uploading GeoJSON files

Suppose the following variable is defined,

• ISLANDS_GEO_JSON_VERSION: version of the GeoJSON file for islands; e.g., ISLANDS_GEO_JSON_VERSION=v2
• PAPERS_GEO_JSON_VERSION: version of the GeoJSON file for papers; e.g., PAPERS_GEO_JSON_VERSION=v2

1. Upload a GeoJSON file of islands.

   aws s3 cp islands.json s3://$GEO_JSON_BUCKET/$ISLANDS_GEO_JSON_VERSION/islands.json
   

   You have to specify an appropriate credential.

2. Upload a GeoJSON file of papers.

   aws s3 cp papers.json s3://$GEO_JSON_BUCKET/$PAPERS_GEO_JSON_VERSION/papers.json
   

   You have to specify an appropriate credential.

Deploying API stack

1. Build Lambda functions.

   sam build --template api/api-template.yaml
   

   Go cannot be built with the --use-container option.

2. Package and deploy functions and API.

   sam deploy --stack-name imaginary-map-api --capabilities CAPABILITY_IAM --s3-bucket $CODE_REPOSITORY --parameter-overrides GeoJsonBucketName=$GEO_JSON_BUCKET
   

   You have to specify an appropriate credential.

NOTE: when you modify the template, you have to start over from the step 1 even if you have not modified Lambda functions.

Deploying API Gateway

The CloudFront asks a develop stage of the API for new map vector tiles. You have to manually deploy a develop stage of the API to enable the CDN.