How pyblish/pyblish-base works
Project type and purpose
• Pyblish-base is a Python library/framework: a modular, plugin-driven validation and publishing engine for creative pipelines.
• Its primary purpose is to provide a core API and plugin system. Other “pyblish-*” projects (e.g., integration with Maya or QML) build on top of this base.
File structure
• Top-level:
• setup.py, setup.cfg – packaging
• docs/ – Sphinx documentation
• pyblish/ – core package
• tests/ – unit and plugin compatibility tests
• run_testsuite.py, run_coverage.py – testing utilities
• Within pyblish/:
• __main__.py, cli.py – command-line entrypoints
• api.py – core API definitions (Context, Instance, etc.)
• logic.py – orchestration of plugin execution
• plugin.py – Plugin base class, discovery & registration
• lib.py, util.py, compat.py, error.py – supporting utilities
• version.py – package version
• plugins/ – example built-in plugins (collectors, validators)
• vendor/ – third-party vendored code
README for additional insights
• Confirms this is the “primary API” project.
• Mentions integration points (pyblish-maya, pyblish-qml) as external modules.
• Test-driven, collaborative, creative-pipeline focus.
a. Main components
• CLI Layer
– __main__.py, cli.py
• Core API & Engine
– api.py (data models: Context, Instance, Plugin)
– logic.py (pipeline orchestration)
• Plugin System
– plugin.py (base Plugin class, discovery)
– pyblish/plugins (built-in plugins)
– External Plugins (e.g., pyblish-maya)
• Utilities & Compatibility
– util.py, lib.py, compat.py
• Testing
– tests/ (unit tests, legacy pre-1.1, pre-1.3 compatibility)
• Documentation
– docs/ (Sphinx)
• Packaging & CI
– setup.py, .travis.yml, appveyor.yml, coveragerc, etc.
b. Relationships & data flows
1. User invokes CLI
→ Entry point (__main__.py) parses arguments
→ CLI module builds initial Context
2. Engine (logic.py) receives Context
→ Discovers plugins via plugin.py
→ Applies pipeline stages in order (collect → validate → extract → conform, etc.)
→ Each Plugin receives Context and modifies it or its Instances
3. Upon completion, results are reported back to CLI
4. Utilities modules support versioning, error handling, compat shims
5. Tests drive coverage and verify compatibility across releases
c. Architectural patterns & principles
• Plugin (inversion of control) pattern – core engine invokes user-provided plugins
• Pipeline / chain of responsibility – sequential stages that transform/validate data
• Separation of concerns – distinct layers (CLI, core API, plugin discovery, plugin code)
d. Key technologies & frameworks
• Python 3 (and 2) – language
• setuptools – packaging (setup.py)
• Sphinx – docs (docs/)
• pytest/unittest – tests
• CI: Travis CI, AppVeyor, Coveralls