A modular, plugin-based platform for parsing, modeling, visualizing, searching, and manipulating graph structures. Built with a clean plugin architecture that allows data sources and visualizers to be swapped or extended independently — the platform remains framework-agnostic and integrates with both Django and Flask.
- Multiple data source formats — JSON, XML, and CSV plugins for loading arbitrary graph data
- Two visualization modes — Simple (force-directed layout) and Block (attribute-rich rectangles)
- Three simultaneous views — Main View (interactive canvas), Tree View (collapsible hierarchy), and Bird View (minimap with synchronized viewport)
- Search & filtering — text-based search and attribute-level filtering (
age > 30) with support for chained operations; applied server-side on the graph model - Type-aware graph model — supports directed, undirected, acyclic, and cyclic graphs; attributes typed as
int,float,str, ordate - CLI — in-browser terminal for creating, editing, and deleting nodes/edges; filtering and searching via command syntax
- Multi-workspace — load multiple graphs simultaneously, each with its own active filters and search state
- Dual framework support — identical functionality available via both Django and Flask interfaces
 |
 |
 |
 |
- Language: Python 3.11+
- Web frameworks: Django, Flask
- Frontend: HTML, CSS, JavaScript, D3.js
- Plugin system: installable Python packages via
pip install -e - Templating: Jinja2 (visualizer plugins generate HTML string representations)
- Data formats: JSON, XML, CSV
The platform follows a strict plugin-based architecture. All components communicate through a shared API library — neither the platform nor the plugins depend on each other directly.
graph-explorer-platform/
├── api/ # Abstract interfaces and graph data model (shared dependency)
├── platform/ # GraphService, Workspace, CLI, PluginRegistry
├── json_plugin/ # Data source plugin — JSON
├── xml_plugin/ # Data source plugin — XML
├── csv_plugin/ # Data source plugin — CSV
├── simple_visualizer/ # Visualizer plugin — force-directed layout
├── block_visualizer/ # Visualizer plugin — attribute block layout
├── django/ # Django web application
└── flask/ # Flask web application
api— defines abstract base classes for data source and visualizer plugins, as well as the graph model (Node,Edge,Graph). Used as a dependency by all other components.platform— implementsGraphService(graph operations, search, filter),PluginRegistry(auto-discovers installed plugins),Workspace(per-session graph + filter state), and the CLI interpreter.- Data source plugins — parse input files and construct typed graph instances using the
apimodel. - Visualizer plugins — receive a
Graphobject and return an HTML string (with embedded CSS/JS) for rendering. - Web apps — Django and Flask applications that wire together the platform and plugins and expose them via a browser UI.
| Student | Index |
|---|---|
| Teodora Aleksic | SV7/2023 |
| Lenka Nikolic | SV16/2023 |
| Igor Maljik | SV37/2023 |
| Lazar Jovic | SV43/2023 |
| Vukasin Vujovic | SV11/2023 |
- Python 3.11+
- pip
Linux / macOS
chmod +x run_django.sh
./run_django.shWindows (PowerShell)
.\run_django.ps1Linux / macOS
chmod +x run_flask.sh
./run_flask.shWindows (PowerShell)
.\run_flask.ps1# 1. Create and activate virtual environment
python -m venv venv
# Windows
venv\Scripts\activate
# Linux / macOS
source venv/bin/activate
# 2. Install all components (order matters — api first)
pip install -e ./api
pip install -e ./platform
pip install -e ./json_plugin
pip install -e ./xml_plugin
pip install -e ./csv_plugin
pip install -e ./simple_visualizer
pip install -e ./block_visualizer
# 3a. Run Django
cd django
python manage.py migrate
python manage.py runserver
# 3b. Or run Flask
cd flask
python app.pyLinux / macOS
chmod +x test.sh
./test.shWindows (PowerShell)
.\test.ps1The test scripts automatically set up the virtual environment, install all dependencies, and run the full test suite.
New data source or visualizer plugins can be added without modifying the platform. A data source plugin only needs to implement the DataSourcePlugin interface from api, and a visualizer plugin implements VisualizerPlugin. Once installed into the active virtual environment, the platform's PluginRegistry detects them automatically.
Design patterns are applied throughout the codebase and were a deliberate part of the architecture. Their usage is discussed during the project defense.
The class diagram covering the graph model and plugin interfaces is available in the /diagrams folder.