Plugin Overview

A Nessie plugin is a Python function decorated with @plugin("Name") that returns a dictionary describing which actions it handles, which actions it requires from other plugins, and what setup data it needs from the user.

Plugin Anatomy

from nessie_api.models import Action, Graph, plugin, SetupRequirementType
from nessie_api.protocols import Context

# 1. Define handler functions
def my_handler(action: Action, context: Context) -> Graph:
    path = action.payload["File Path"]
    # ... load data, build Graph, return it
    return graph

# 2. Register the plugin using the @plugin decorator
@plugin("My Plugin Name")
def my_plugin():
    return {
        "handlers": {
            "load_graph": my_handler,     # action → function
        },
        "requires": [                     # other actions this plugin needs
            "filter_graph",
        ],
        "setup_requires": {              # user-facing setup fields
            "File Path": SetupRequirementType.STRING,
            "Is Directed": SetupRequirementType.BOOLEAN,
        },
    }

Plugin Registration (pyproject.toml)

For the platform to auto-discover your plugin, declare it as an entry point in pyproject.toml:

[project.entry-points."nessie_plugins"]
my_plugin = "my_package.module:my_plugin"

After pip-installing your package (even with pip install -e .), the platform will discover and load it automatically.

SetupRequirementType Values

Official Plugins

Datasource Plugins

Datasource plugins handle the load_graph action and convert an external data source into a Nessie Graph.

Visualizer Plugins

Visualizer plugins handle the visualise_graph action. They receive a Graph and return an HTML string containing SVG nodes and edges. The web UI injects this HTML and then runs a D3 physics simulation on top.

Manipulation Plugins

UI Plugin

---