Viz Plug-ins User Guide
Version 5.0 | Published December 20, 2022 ©
DataPool Plug-ins
The DataPool plug-ins in Viz are responsible for handling incoming data and for affecting the objects in the scene according to the DataFields values.
Overview
The DataPool plug-ins are found in Viz in the Function plug-ins pool, under a folder called Data. There are DataPool Container Plug-ins and DataPool Scene Plug-ins. The scene plug-ins manage incoming data from different sources, inserting it to DataFields and maintain the DataFields Table. For example, the DataPool plug-in receives information from external sources and inserts it to the different DataFields. The DataMouseSensor senses mouse position and events and stores them in predefined DataFields.
The container DataPool plug-ins trigger graphic changes in the container they are attached to, when the DataField registered to them changes. Most of the container plug-ins have an argument called Field Name. The field name value is the name of the DataField to which the plug-in is registered to. The Field Name can take the following forms:
<name>
<name>[number of items]
If the field name is VALUE then the DataField is defined as a single datum field. However, if the field name is VALUE [100] then the name of the DataField is VALUE and it has 100 data items (values). The field name, scene hierarchy and the way the information is sent to the DataPool plug-in determine the container plug-in behavior.
Example
This is a hierarchy of a graph scene:
The above figure shows a container A with three children 1, 2, and 3. A DataScale plug-in is attached to A. The DataScale plug-in controls the scaling of the container according to the data it receives.
If the Field Name in the DataScale plug-in is defined to be VALUES[3] (a vector of 3 items named VALUES), DataScale acts in two different ways:
-
If the data is entered without specifying an item index, DataScale affects container A.
Example: If the data is entered as VALUES=0.5 0.5 0.5; then the container A is scaled by 0.5 in each axis.
-
If the data specifies an item index, DataScale works on the respective child container of A.
Example: If the data is entered as VALUES[2] = 0.1 0.1 0.1;, then container3 is scaled by 0.1 in each axis (Note that the index value is zero based).