Managing data transformation (e.g. dbt) projects at scale can be challenging: DAG violations, that one model that grew to 2K lines, duplicated logic.

Inspired by the book 'Building Evolutionary Architectures' I built tff to help data transformation projects evolve in a desired direction. It comes with a github bot that you can plugin to your CI pipeline.

It aims to let transformation projects evolve in the desired direction by: * ensuring layer boundary integrity; * detect duplicate CTE logic; * catch violations locally and in CI

Core concepts: rules apply to models; checks apply to architecture and DAG checks.

Curious to read what you think. What can be improved? What's missing?