Viewer Modifier Porting Guide¶
Viewer Modifier plug-ins are a feature of Katana’s legacy, OpenSceneGraph-based Viewer tab, which allows custom drawing code to be executed for arbitrary scene graph location types. Viewer Modifiers are typically used to draw viewer-only representations of light and camera locations.
Viewer Modifiers are limited in their power and flexibility for various reasons, so with the development of the new Viewer API an alternative solution (now known as Locators) was developed. This guide will cover an approach for emulating the behaviour of legacy Viewer Modifier plug-ins, and how to convert any existing Viewer Modifiers to the new Viewer API’s Locators.
A new type of Python plug-in called a ViewerPluginExtension has been added which contains several callback functions that are triggered from the BaseViewerTab class in response to certain events. These functions are
onApplyTerminalOps(). Each function is passed relevant arguments, such as the ViewerDelegate or ViewportWidget objects. This allows a ViewerPluginExtension to modify a viewer’s behaviour by applying other plug-in types which may need to work together, such as ViewerDelegateComponents and ViewportLayers.
An example of a VPE may look like this:
from PluginAPI.BaseViewerPluginExtension import BaseViewerPluginExtension class BallViewerPluginExtension(BaseViewerPluginExtension): def onDelegateCreated(self, viewerDelegate, pluginName): viewerDelegate.addComponent('BallComponent', 'BallComponent') def onViewportCreated(self, viewportWidget, pluginName, viewportName): # Insert a layer named BallLayer at index 4 into the given viewport widget viewportWidget.insertLayer('BallLayer', 'BallLayer', 4) PluginRegistry = [ ("ViewerPluginExtension", 1, "BallViewerPluginExtension", BallViewerPluginExtension), ]
Viewport Layer & Viewer Delegate Component¶
These are two separate plug-ins that can be added to a viewer, usually by a ViewerPluginExtension. The ViewerDelegateComponent is used to track the scene graph and identify locations that require custom drawing and the ViewportLayer is responsible for drawing the locations that have been identified. These classes give you complete control over the drawing of your own geometry, but will require some bookkeeping in order to track changes to the scene graph and selection states.
Typically a viewer will have one ViewerDelegate, but possibly multiple viewports, so computationally heavy work is best suited in the VDC. Any code that requires a current OpenGL context should be called within one of the relevant ViewportLayer functions, or you can call
Viewport::makeGLContextCurrent() before executing.
An example of this style of plug-in is distributed with Katana and can be found at $KATANA_ROOT/plugins/Src/ViewerPlugins/BallLocationType.
Once you have a ViewportLayer and ViewerDelegateComponent plug-in you need to write a ViewerPluginExtension as seen above to append them to the ViewerDelegate and Viewport.
In order to accelerate the process of converting old ViewerModifier plug-ins for use in a Viewer API based viewer, we have provided a base implementation of a ViewportLayer and ViewerDelegateComponent pair and distribute their source code with Katana. The header for these plug-ins can be found in $KATANA_ROOT/plugin_apis/include/FnViewer/utils/FnBaseLocator.h. There is also an example that uses these classes to draw a spotlight cone in a similar manner to how you might have done in an older ViewerModifier. You can find the example in $KATANA_ROOT/plugins/Src/ViewerPlugins/ExampleSpotlightLocator/.
In order to get started you simply have to write a class which derives from
FnKat::ViewerUtils::FnBaseLocator which is then registered with the
FnBaseLocatorVDC class via the
Each locator needs to have four static functions before it can be successfully registered with the ViewerDelegateComponent. These are as follows:
// Determines whether this locator plug-in should be used to draw the // location described in the given viewer location event. static bool matches(const FnKat::ViewerAPI::ViewerLocationEvent& event); // Specifies whether this locator plug-in overrides the standard geometry // representation for matched locations. static bool overridesBaseGeometry(const FnKat::ViewerAPI::ViewerLocationEvent& event); // Gets the bounds of the given location. static FnAttribute::DoubleAttribute getBounds( const FnKat::ViewerAPI::ViewerLocationEvent& event); // Computes the extent of the given location. static FnAttribute::DoubleAttribute computeExtent( const FnKat::ViewerAPI::ViewerLocationEvent& event);
The VDC will then use the
matches() function of all Locators that have been registered with it to determine whether it should be used to draw each location. This function is similar to the same function found in ViewerManipulators, in that it allows your plug-in to match the location based on any conditions, rather then just the location type, as was necessary with old style ViewerModifiers. This implementation also allows multiple Locators to be active for a single location, which was not previously possible. This interface should allow you to convert old viewer modifiers into Locator plug-ins with relative ease. If more control over the plug-in behaviour is required, it is possible to write a class that derives from the base implementations of the VDC and ViewportLayer that we provide. A simple example of this is included in the ExampleSpotlightLocator plug-in.
Please note that the VDC and ViewportLayer classes should be compiled into the same dynamic library as your
FnKat::ViewerUtils::FnBaseLocator classes to ensure that you do not encounter issues related to passing complex objects across the boundary of a dynamic library.