How revodatanl/revo-asset-bundle-templates works
Project type and purpose
• This is an open-source “scaffolding” or template-generator repository for Databricks Asset Bundles.
• Purpose: provide a turnkey project structure (code, pipelines, docs, devcontainer, tests, CI/CD) that users can initialize via databricks bundle init.
Top-level structure and key directories
• .github/ + .github/workflows/: GitHub Actions CI/CD templates
• modules/Makefile: reusable Make targets for building, testing, packaging
• release.config.mjs: semantic-release configuration
• template/: Jinja-style templates ({{.project_name}}, {{.package_name}}) for generating a new project
• .azuredevops/: Azure Pipelines YAML templates
• .devcontainer/: Dockerfile and devcontainer configuration
• docs/: markdown docs templates (getting started, coding standards)
• notebooks/, resources/: example DLT pipelines and jobs
• src/{{.package_name}}: stub Python package
• tests/: example Pytest tests
A. Template Repository
– Houses all scaffolding artifacts (CI/CD, devcontainer, docs, code stubs)
B. Template Engine (Databricks CLI plugin)
– databricks bundle init plugs into this repo to render templates into a new project
C. Generated Project Workspace
– The user’s new repo with concrete files, pipelines, docs, code, tests
D. Development Environment
– Local machine using DevContainer (Docker + VSCode)
– Pre-commit hooks, Mypy, Ruff linters configured via .pre-commit-config.yaml
E. CI/CD Platforms
– GitHub Actions (defined in .github/workflows/)
– Azure DevOps Pipelines (in .azuredevops/)
– semantic-release for automated versioning and publishing
F. Databricks Environment
– Target cloud environment (Azure or AWS)
– Databricks CLI profiles to authenticate & deploy asset bundles
Relationships and data flows
1. User invokes Databricks CLI → reads template/ → Template Engine renders files → outputs Generated Project
2. Developer opens Generated Project in DevContainer → writes code & tests → runs Make targets locally
3. Developer pushes to GitHub/Azure Repos → triggers CI/CD workflows (build, lint, test)
4. CI/CD pipelines invoke Databricks CLI → deploy notebooks, DLT pipelines, jobs to Databricks workspace
5. semantic-release auto-increments version → publishes package (e.g., to PyPI)
Architectural patterns and principles
• Template-based scaffolding (separation of concerns between template repo vs generated project)
• Pipeline-as-code (CI/CD defined alongside code stubs)
• Infrastructure automation (Databricks CLI + pipeline scripts)
• Modular Makefile for common tasks