PFT Extractor

A project for managing pulmonary function test data using AWS.

PFT X uses Textract and regular expressions to filter pulmonary function test data out of PDF format and into structured JSON.

Note This project is for demonstration purposes only.

Architecture

Tech

• AWS
  • DynamoDB
  • Lambda
  • Textract
  • S3
  • KMS
  • SNS
  • CDK
  • CloudWatch
  • CloudFormation
• TypeScript
• SST + AWS CDK
• React
• Next.js

Opinions

• ⚙️ Reproducible cloud infrastructure
• 🕵️‍♂️ Audit friendly

Environment setup

src/server

0. Set up ~/.aws/credentials. This guide will assume you are using your crendetials as the "default" profile.
1. Ensure you have Node.js 14.x installed
2. cd src/server; npm install
3. npm run deploy
4. Create a user account to log into via the AWS Managment panel -> Cognito
5. The deployment will store output values necessary to run the client in outputs.cdk.json - this file will not be committed

src/client

The frontend application for the project can be run completely locally.

0. Ensure you have Node.js 12.x installed
1. cd src/client; npm install
2. Use the outputs.cdk.json values to build a file at src/client/.env.local that looks like:

NEXT_PUBLIC_API_ENDPOINT = ""
NEXT_PUBLIC_AWS_REGION = ""
NEXT_PUBLIC_USER_POOL_ID = ""
NEXT_PUBLIC_USER_POOL_WEB_CLIENT_ID = ""

3. npm run start

Development

src/server

We use AWS CloudFormation Stacks to describe the entirety of an individual deployment of the software. We use Serverless Stack (SST) - which is a superset framework on top of AWS CDK - to achieve this. One of the great features about SST is the ability to proxy Lambdas locally. When you're reading to begin work on the contents of src/server/src, which is where the APIs etc are defined, you can run a local development environment with npm run start.

src/development

No additional development notes. Make sure you set up your .env.local file.