medics-extension-sdk 2026.5.19

SDK for creating extensions for the MedICS (Medical Image Computing and Segmentation) platform

MedICS Extension SDK

Python SDK for building extensions for the MedICS (Medical Image Computing and Segmentation) platform. Provides base classes, API utilities, and a CLI scaffolding tool.

---

Installation

pip install medics-extension-sdk

With Qt support (required for widget-based extensions):

pip install medics-extension-sdk[qt]      # PySide6 (recommended)
pip install medics-extension-sdk[pyqt6]   # PyQt6
pip install medics-extension-sdk[qt5]     # PyQt5 (legacy)

Development tools:

pip install medics-extension-sdk[dev]

---

Quick Start

1. Create an extension class

from medics_extension_sdk import BaseExtension

class MyExtension(BaseExtension):
    def __init__(self, parent=None):
        super().__init__(
            parent=parent,
            extension_name="My Extension",
            author_name="Your Name",
        )

    def get_version(self) -> str:
        return "1.0.0"

    def get_description(self) -> str:
        return "Does something useful with medical images"

    def get_category(self) -> str:
        return "Analysis"

    def create_widget(self, parent=None, **kwargs):
        from PySide6.QtWidgets import QLabel
        return QLabel("Hello from My Extension!")

The extension id is automatically derived: "your_name.my_extension".

2. Access platform services

# Inside any extension method:
self.log_message("Processing started")
value = self.get_config_value("my_ext", "threshold", 0.5)
self.set_config_value("my_ext", "threshold", 0.8)
self.send_event("processing_done", {"status": "ok"})
workspace = self.get_component("workspace_manager")

3. Expose a programmatic API

Decorate public methods with @api_method so they are auto-discovered at runtime:

from medics_extension_sdk import BaseExtension, api_method

class MyExtension(BaseExtension):
    ...

    @api_method(description="Run analysis and return results", category="Processing")
    def run_analysis(self, threshold: float = 0.5) -> dict:
        return {"threshold": threshold, "result": "ok"}

Retrieve the API from an instance:

ext = MyExtension()
api = ext.get_instance_api()   # returns apiDict of decorated callables
api.run_analysis(threshold=0.7)

---

Scaffolding

Generate a new extension template with the CLI:

medics-create-extension "My Segmentation Tool" \
    --author "Jane Doe" \
    --category "Segmentation" \
    --output ./extensions

This creates a fully-structured extension directory with a main class, widget, API methods, config file, and README.

---

API Reference

BaseExtension

Member	Type	Description
id	str (read-only)	Auto-generated from author_name.extension_name
extension_name	str (read-only)	Set in __init__
author_name	str (read-only)	Set in __init__
app_context	Any	Injected by the platform on initialize()
get_version()	abstract	Return version string
get_description()	abstract	Return short description
get_category()	str	Extension category (default: "General")
create_widget(parent, **kwargs)	QWidget	Override to provide Qt UI
initialize(app_context)	bool	Called by platform at load time
cleanup()	None	Called by platform at unload time
get_instance_api()	apiDict	Returns dict of @api_method callables

ID format: {author_name}.{extension_name} — lowercased, spaces removed, non-alphanumeric characters replaced with _.

@api_method(description, category)

Marks a method for auto-discovery via get_instance_api().

apiDict

dict subclass with dot-notation attribute access.

---

Extension Categories

Standard categories for organizing extensions in the platform UI:

• "Analysis" — measurement and analysis tools
• "Segmentation" — segmentation algorithms
• "Visualization" — rendering and display
• "Processing" — image filters and transforms
• "Import/Export" — data I/O utilities
• "Workflow" — automation and pipeline tools
• "Utilities" — general helpers
• "Research" — experimental tools

---

Testing

import pytest
from medics_extension_sdk import BaseExtension

class MyExtension(BaseExtension):
    def __init__(self):
        super().__init__(extension_name="My Ext", author_name="Author")
    def get_version(self): return "1.0.0"
    def get_description(self): return "Test"

def test_id_readonly():
    ext = MyExtension()
    assert ext.id == "author.my_ext"
    with pytest.raises(AttributeError):
        ext.id = "hacked"

Run:

pytest tests/

---

Project Layout

Recommended structure for a published extension package:

my_extension/
+-- __init__.py          # exports Extension = MyExtensionClass
+-- my_extension.py      # main extension class
+-- icons/               # optional icon assets
+-- config.ini           # default configuration
+-- README.md

---

License

MIT - see LICENSE.