Rego Data Filtering Playground

This is a simple dashboard to demonstrate the data filtering capabilities we're working on in the Enterprise OPA Platform.

Getting started

1. Clone the repo
2. docker compose up --build (respects, but doesn't require an EOPA_LICENSE_KEY env var)
3. Go to http://localhost:3000

When the Rego code in the "Data Policy" pane is edited, it is automatically sent to EOPA. That PUT request for the Policy API could fail -- if so, it's because of syntax errors which are then displayed in red. If the PUT succeeds, a request for evaluating the data policy is sent to EOPA, which either yields partial eval post-analysis errors, or a SQL query. Post-analysis errors are annotated in yellow. If a query is returned, it's used to subset the (hidden) SQL query collecting ordered products by user.

Notes

• Error locations provided by the EOPA API only contain starting points (row/col), and the editor component would really appreciate spans (start, end). To work around this, we're extending all the starting points to the end of the line. It's close enough in most cases.

• For local development,

  1. start EOPA on 127.0.0.1:8181 (no config or policy files needed)
  2. run npm run dev (npm install if you haven't before)
  3. run caddy run --config Caddyfile-local to proxy both Observable Framework and EOPA
  4. go to http://localhost:3001

• Open Dev Tools network tab for insights if something goes wrong: not all errors are currently surfaced in a usable way.

• If you want to test this against a development EOPA build, run make build-local on the EOPA source branch, and start the stack via

  EOPA_IMAGE=ko.local/enterprise-opa-private:edge EOPA_PULL_POLICY=missing docker compose up

Community

For questions, discussions and announcements related to Styra products, services and open source projects, please join the Styra community on Slack!