How dpgalliance/publicgoods-candidates works
Project Type and Purpose
• This is an open-source data repository and pipeline for managing nominations of Digital Public Goods (DPGs).
• It hosts JSON “content” (nominee metadata, reviews, archived items), transformation scripts, documentation, and CI workflows.
• The output feeds a static site or downstream DPG application that publishes the list of approved and candidate public goods.
Key Architectural Components
a. Data Stores (Content Layer)
• /nominees – JSON files for each candidate project
• /digitalpublicgoods – JSON metadata for approved DPGs
• /archive, /excluded-nominees – JSON for historical and excluded items
• /reviews – reviewer feedback JSON
• nominee-schema.json, screening-schema.json – JSON Schema for validation
b. Processing Engine (Transformation Layer)
• /scripts – Node.js scripts (csvToJson.js, migrate-*.js, api.js, etc.) that:
– validate JSON against schemas
– normalize, merge, order fields
– generate or update JSON outputs
• package.json, package-lock.json – define dependencies and scripts
c. CI/CD Pipelines
• .github/workflows/*.yml – GitHub Actions that:
– run tests and validation on pull requests
– execute transformation scripts automatically
– deploy updated content to the public site (e.g., GitHub Pages or external endpoint)
d. Documentation & Policies
• /docs – development guides, overview, bug reporting
• /help-center – content guidelines, license policies, privacy practices
• CONTRIBUTING.md, CODE_OF_CONDUCT.md, LICENSE – process and governance
e. External Services
• digitalpublicgoods.net (static site or web application) – consumes generated JSON data
• bug-reporter.vercel.app – issue intake
• GitHub – source hosting, Actions, Pages
Relationships and Data Flows
• Authors commit JSON content to the repo.
• On push or pull request, GitHub Actions validate and run scripts.
• Scripts produce normalized JSON and may push built artifacts.
• CI then publishes the updated dataset to the DPG website or API endpoint.
• End users browse the static site or API to explore DPGs.
Architectural Patterns and Principles
• “Content as Code”: treating JSON data and schemas like source code under version control.
• “CI-driven pipeline”: automated validation, transformation, publication.
• Separation of concerns: distinct layers for data (JSON), processing (scripts), and deployment (CI).
• Schema-first design: enforce data quality via JSON Schema.