How ynput/ayon-substance-designer works
Project type and purpose
• This is an integration-plugin project: a Python-based Adobe Substance Designer addon that plugs into the AYON pipeline.
• Its goal is to let an artist inside Substance Designer create, load, and publish assets directly to an AYON server.
File structure
• Top-level directories:
• client/ayon_substancedesigner – the Substance Designer plugin and its API wrappers
• server/settings – default configuration for image I/O and main AYON settings
• docs – mkdocs‐based documentation
• Key Python packaging files: pyproject.toml, package.py, create_package.py
• CI/CD workflows under .github/workflows orchestrate linting, labeling, releases, and deployment to Ynput Cloud.
README
• Confirms the plugin’s purpose: “Adobe Substance Designer integration for AYON.”
• Points to the Substance 3D Designer scripting API, implying heavy use of Adobe’s Python scripting bridge.
Main components
a. Substance Designer host application
b. AYON plugin (client/ayon_substancedesigner)
• subdivided into API modules (lib, pipeline, plugin, project_creation)
• plugin hooks (addon.py) and deploy scripts
• sub-modules for create/load/publish steps
c. AYON Server
• Core AYON REST API endpoints
• Data store for assets, versions, and workfiles
• imageio service (server/settings/imageio.py)
d. External dependencies
• Adobe scripting API (External)
• CI/CD (GitHub Actions)
• Documentation site (MkDocs)
Relationships and interactions
• User triggers action in Substance Designer → the plugin’s addon.py
• Plugin calls into its internal API modules (client/api/*.py)
• API modules construct REST calls to the AYON Server for:
• Project creation
• Pipeline retrieval
• Uploading published assets (SBSAR, textures, workfiles)
• Server responds with job IDs, metadata, and storage URLs
• Plugin downloads or loads textures/workfiles when requested
• CI/CD workflows run on git events to lint, tag, and publish the plugin package
Architectural patterns and design principles
• Modular API wrapper: separate lib, pipeline, plugin, and project_creation concerns
• Plugin command pattern: each create/load/publish is its own module under plugins/
• Config-driven behavior: server/settings defines environment specifics
• Standard Python packaging and deployment (pyproject.toml, create_package.py)
Final advice
• Be meticulous: break the plugin into its API layer and command modules
• Full lifecycle: creation, loading, publishing, and CI/CD
• Don’t overthink—represent each logical component (api/lib, api/pipeline, plugins/create, etc.) as its own node for maximum clarity.