Node-Based Image Processor

This project is an implementation of a node-based image manipulation application in Python, inspired by tools like Substance Designer or Blender's node editors, and provides a basic graphical interface for interaction.

Overview

Users can visually create image processing pipelines by connecting different nodes. Each node performs a specific operation (loading, filtering, blending, saving, etc.). The graph automatically determines the execution order and updates previews in real-time (planned). The GUI allows users to interact with the node graph visually.

Current Status (Step 2: Basic GUI)

• Core Framework: Implemented the fundamental backend classes (Graph, Node, Port, Connection) focusing on structural representation (evaluation logic removed for now). Includes basic validation, addition/removal of elements. Unit tests cover this core structure.
• Basic GUI (PySide6): Implemented the initial visual framework:
  • Displays nodes (UINode), ports (UIPort), and connections (UIConnection) on a canvas (NodeGraphScene).
  • Allows moving nodes via drag-and-drop.
  • Context menu (right-click) to add simple placeholder nodes (Input, Process, Output).
  • Basic connection drawing by dragging between compatible output/input ports.
  • Limitations: Uses temporary placeholder backend nodes defined in the UI code, no properties panel for editing, basic styling, connection visual feedback only on completion, no save/load functionality.

Getting Started

1. Clone the repository.
2. Create a virtual environment: python -m venv venv
3. Activate the environment:
   • Windows: .\venv\Scripts\activate
   • macOS/Linux: source venv/bin/activate
4. Install dependencies: pip install -r requirements.txt (includes numpy and PySide6).
5. Run the application: python main.py

Next Steps

• Implement real image processing backend Node types (e.g., ImageInput, Blur, BrightnessContrast, Output) using OpenCV/Pillow, likely in a separate nodes/ directory.
• Integrate these real nodes into the GUI's NODE_TYPE_MAP in node_graph_scene.py.
• Develop the GUI properties panel to display and edit parameters of the selected node.
• Implement the evaluation engine (topological sort, cycle detection, caching, node.process() execution) in the backend node_graph package.
• Connect GUI actions (e.g., parameter changes) to trigger backend graph re-evaluation.
• Add visual feedback during connection dragging.
• Implement saving and loading of the graph structure (node positions, connections, parameters).
• Refine UI styling and overall user experience.
• Add proper error handling and display.

Project Structure