carrier-risk-platform

Carrier Risk Platform is an open-source solution designed to provide point-in-time accurate risk scoring for carriers based on federal FMCSA (Federal Motor Carrier Safety Administration) data. The project utilizes a robust technology stack centered around Apache Airflow for workflow orchestration, dbt for data transformation, and Terraform for infrastructure as code. The repository includes comprehensive CI/CD workflows that ensure code quality and security. These checks run on pull requests and pushes to the main branch, utilizing GitHub-hosted runners. Key validation steps include redacted Gitleaks scanning for secrets, Ruff linting and formatting, strict type checking with mypy, and a full pytest suite with branch coverage enforcement. The system enforces a high code coverage baseline (approximately 88%) and requires at least 80% coverage for changed lines in critical modules like `ingest` and `orchestration`. Data integrity is maintained through dbt builds that enforce model contracts, idempotency regressions, and temporal tests. Infrastructure testing is handled via Terraform, which validates configurations and exercises mock-provider plans without deploying actual cloud resources. PostgreSQL is used for local database fixtures during testing, with credentials generated and masked securely. The platform is designed to be deployed on AWS, leveraging Terraform for resource management.