How cable-robots/cdpyr works
Project type and purpose
• This is a pure-Python, open-source library (cdpyr) for designing, analyzing, and simulating cable-driven parallel robots.
• There is no “frontend/backend” UI; it exposes an API to load robot specifications, perform kinematic/dynamic analyses, and render results.
File structure
• Top-level: packaging config (setup.py, pyproject.toml), CI (.travis.yml, gitlab-ci.yml, ci/), docs (docs/), tests (tests/).
• src/cdpyr/ contains core packages:
• cdpyr.robot: domain model of robot components (cables, pulleys, frames, motors)
• cdpyr.kinematics & cdpyr.mechanics: coordinate transforms, inertial calculations
• cdpyr.analysis: layered analysis engines (archetype, kinematics, dynamics, criterion, stiffness, workspace, force distribution)
• cdpyr.stream: parsers for JSON/XML/YAML and TwinCAT data
• cdpyr.schema & cdpyr.validator: (de)serialization and data validation
• cdpyr.visualization: plug-in engines (Mayavi, Plotly) to render models and results
README for additional insights
• Confirms it’s a Python 3 package published on PyPI.
• No built-in UI—users import the library and call its API.
• Testing via pytest and tox; docs on readthedocs.
a. Main components (group by package):
1. Data Input & Parsing
• XML/JSON/YAML parsers (cdpyr.stream.parser.*)
• TwinCAT parser
2. Domain Model
• Robot definitions (cdpyr.robot.*)
• Geometry primitives (cdpyr.geometry.*)
3. Core Computation
• Kinematics & Mechanics (cdpyr.kinematics, cdpyr.mechanics)
• Analysis engines (cdpyr.analysis.*)
4. Schema & Validation
• Pydantic-style schemas (cdpyr.schema.*)
• Validators (cdpyr.validator.*)
5. Visualization
• Engine abstractions (cdpyr.visualization.engine.*)
• Concrete backends (Mayavi, Plotly)
6. Infrastructure
• Helpers, mixins, base classes (cdpyr.helper, cdpyr.mixin, cdpyr.base)
• CI, tests, docs (peripheral)
b. Relationships and interactions
• Parsers read external specs → emit Domain Model objects
• Domain Model feeds Geometry & Kinematics modules → produce transformation matrices, inertias
• Analysis engines consume models plus kinematic/dynamic outputs → compute performance metrics (workspace, stiffness, singularities, force distribution)
• Results and models pass through Schema & Validator for export or API output
• Visualization engines subscribe to model + results to render graphs or 3D scenes
c. Architectural patterns & design principles
• Layered architecture (Input → Model → Compute → Output)
• Plugin-style extensibility: new analysis.archetype or criterion classes drop into an import folder
• Separation of concerns: parsing, computation, serialization, and rendering are isolated
d. Key technologies & frameworks
• Python 3, NumPy (linear algebra), SciPy (optional), pytest/tox for testing
• Data serialization (JSON, XML, YAML), Pydantic-style schema
• Visualization via Mayavi and Plotly