RL-Vizualization

Overview

This repository provides a graphical user interface (GUI) for generating figures from Reinforcement Learning (RL) evaluation data. It aims to simplify the visualization process by offering a modular approach using abstract base classes. Users can develop their own plotting scripts using these base classes to suit their specific needs.

Features

• GUI for interactive data visualization.
• Abstract base classes for defining custom plotting scripts.
• Supports various RL evaluation metrics and data formats.
• Export figures in multiple formats (e.g., PNG, SVG).

Installation

Requirements

• Python 3.x
• Dependencies listed in requirements.txt

Installation Steps

1. Clone the repository:

   git clone https://github.com/OSUrobotics/RL-Vizualization.git
   cd RL-Vizualization

2. Install dependencies:

   pip install -r requirements.txt

Usage

Plotting is implemented for a selection of environments linked in compatible environments.txt

Plotting for your specific environment will need to be developed (see Devloping Custom Plotting Scripts)

Running the GUI

To start the GUI for plotting evaluation data from the mojograsp environment:

python Data_analysis_gui.py

Developing Custom Plotting Scripts

1. Extend Base Classes: Implement your custom plotting logic by extending the provided abstract base classes (BasePlotter and DataLoader). This allows you to integrate your specific RL evaluation data and plotting requirements seamlessly into the GUI.

2. Example Script:

   from backend_base import BasePlotter
   from data_loading_base import DataLoader
   
   class CustomPlotter(BasePlotter):
       def draw_my_plot(self, data):
           # Custom plotting logic using matplotlib or other libraries
           pass
   
   class CustomDataLoader(DataLoader):
       def load_data(self, file_path):
           # Custom data loading logic for your RL evaluation data format
           return RL_dataframe

3. Integrating Custom Scripts: Integrate your custom plotting scripts into the GUI by modifying the Data_analysis_gui.py file or using as standalone scripts.

Contributing

Contributions are welcome! Please fork the repository and submit a pull request with your enhancements.

Contact

For questions or support, please contact [email protected].