Document Extractor Proof of Concept

Warning

This application is no longer being developed or maintained.

Public Benefits Studio's Document Extractor to automate document data extraction with AI and OCR.

Using and Running

Using your browser of choice, navigate to the CloudFront distribution URL to start using.

Development

Requirements to Deploy

The requirements needed to deploy are...

• Python.
• uv.
• Terraform.
• Node.js.
• An AWS account.

Deploying

You can mimic the steps in our Continuous Delivery GitHub Action.

The basic steps to accomplish this are...

1. Build the backend.
2. Build the frontend.
3. Deploy using the Infrastructure as Code (IaC).

Authentication

One username and password is supported currently. To set this up, do a deployment and then do the following...

1. Generate an RSA private key. Use the PEM encoding. Upload it to AWS Secrets Manager with private-key in the name.
2. Generate a public key from the aforementioned private key. Use the PEM encoding. Upload it to AWS Secrets Manager with public-key in the name.
3. Come up with a username you want to use. Upload it to AWS Secrets Manager with username in the name.
4. Come up with a strong password you want to use. Do not store the plaintext version of the password in AWS Secrets Manager because it is not ideal and doesn't work anyway. From the backend folder, run echo 'import bcrypt;print(bcrypt.hashpw(b"<your password>", bcrypt.gensalt()).decode())' | uv run - and replace <your password> with the password you want to use. Upload what is printed out to AWS Secrets Manager with password in the name.

Building

Backend

To build the backend, execute...

cd ./backend/
uv sync
uv run build.py

The built artifact is backend/dist/lambda.zip.

Frontend

To build the frontend, execute...

cd ./ui/
npm ci
npm run build

The built artifact is in ui/dist/.

Additional Requirements to Develop

The additional requirements needed to contribute towards development are...

• Pre-Commit.

Running Locally

Frontend

cd ./ui/
npm install
npm run dev

1. Navigate to the UI folder
2. Install dependencies
3. Start the development server

And it runs on http://localhost:1234

Known issues and future improvements

• Replace browser system alert messages with user-friendly error handling.
• Improve page transitions for smoother user experience.
• Display the uploaded filename in the UI.

Future considerations

• Add frontend unit tests
• As the project grows, consider using USWDS Sass for better CSS organization and customization.

Pre-Commit Hooks

We use pre-commit to run some hooks on every commit. These hooks do linting to ensure things are in a good spot before a commit is made. Please install pre-commit and then install the hooks.

pre-commit install

Most of the time any errors encountered by pre-commit are automatically fixed. Run git status to see the fixed files, run git add . to add the fixes, and rerun the commit. You will need to manually fix any errors that are not automatically fixed.