How kanghengliu/tabula works
Project type and purpose
• This is an open‑source Python library (“Tabula”) for applying traditional reinforcement‑learning (RL) algorithms to simple games. It provides:
• Environments (boat, grid world, geo search)
• Solvers (Dynamic Programming, Monte Carlo, Temporal‑Difference)
• Utility routines (metrics collection, plotting, asset management)
• Example scripts and unit tests
File structure
• Top‑level files:
• README.md, LICENSE, pyproject.toml, requirements.txt → project metadata, dependencies, packaging
• docs/ → HTML/Jupyter docs for homework or demonstrations
• examples/ → example_usage.py illustrating CLI usage
• unit_tests/ → test suites for each module
• Core package: tabula/
• assets/ → images, plots, or other static resources
• environments.py → defines the game environments
• solvers.py → implements the three RL solvers
• utils.py → metrics collection, plotting functions, helper routines
README for additional insights
• Describes features (envs, solvers, metrics visualizations)
• Shows CLI usage pattern (python examples/example_usage.py --env … --algo …)
• Mentions outputs (convergence_plot.png, gameplay.gif, policy_visualization.png)
• Indicates how to build, install, and run tests
Detail without over‑splitting
• External files (pyproject.toml, requirements.txt) as part of the Packaging/Build component
• A separate Testing component that points at each core module