A Model Class is the central configuration point of the Optimization module. It defines the structure and the look and behavior of Model Objects. The user workflow consist of sequence of Steps. In each Step, the user can run a Calculation and then explore and evaluate the results on multiple Tabs.
The Model Class definition is stored in a table with type code MC which has following fields:
uniqueName* – Unique name of the Model Class. It is used in Model Object to reference the Model Class they instantiate. Note that you can modify the uniqueName, but the references will not be updated automatically.
label – Human readable name of the Model Class, displayed to the users, when they select which Model Class they want to use for creating a new Model Object.
definition – Defines how the Model Class instances will look and behave. It stores definitions of Calculations, Steps and Evaluations in a JSON text format.
workflowFormulaName – Name of the workflow logic (with workflowType = “model”) which is responsible for building the approval workflow steps when the user clicks Submit. If not set, then the workflow feature is disabled for those Model Objects.
Model Class Definition
In a Model Class, the definition field contains a JSON document that has three main root elements: calculations, steps and evaluations.
Calculations (calculations)
A list of calculation. A Calculation defines a calculation process which can be started from a Step.
Example:
{
"calculations": [
{
"name": "calc1",
"type": "formula",
"formulaName": "POAI_Test_Calc1"
},
{
"name": "seq",
"type": "sequence",
"calculations": [
"calc2", "calc3"
]
},
{
"name": "calc2",
"type": "formula",
"formulaName": "POAI_Test_Calc2"
},
{
"name": "calc3",
"type": "formula",
"formulaName": "POAI_Test_Calc3"
},
{
"name": "calc4",
"type": "parallel",
"formulaName": "POAI_Test_ParallelCalc"
}
]
}
Fields:
name* – Name of the Calculation, unique within this Model Class.
type* – Accepted values:
formula – Simple calculation, all actions in the provided logic are done in a sequential order.
parallel – Parallel calculation. The calculations done by the logic can be split into more parts (called Items) and the calculation of those Items will be done in parallel.
sequence – This calculation does not have a logic, it is only reference to a list of other calculations which should be executed one by one.
formulaName* – Unique name of a logic. Used only for formula and parallel types.
For formula type – The logic must be of Nature
model_calculation. See Model Calculation Logic.For parallel type – The logic must be of Nature
model_parallel_calculation. See Model Parallel Calculation Logic.
sequence* – List of Calculation names defined in this Model Class. Used only for the sequence type.
Step (steps)
A list of Steps and their Calculations, Tabs and their Evaluation logics.
Example:
{
"steps": [
{
"name": "step1",
"label": "First Step",
"description": "This step has 1 calculation and 1 tab.",
"calculation": "calc1",
"tabs": [
{
"name": "tab1",
"label": "First Tab",
"type": "dashboard",
"formulaName": "POAI_Test_Dashboard1"
}
]
},
{
"name": "step2",
"label": "Second Step",
"description": "This step has 1 sequence calculation and 3 tabs.",
"calculation": "seq",
"tabs": [
{
"name": "tab1",
"label": "First Tab",
"type": "simple",
"formulaName": "POAI_Test_Dashboard2"
},
{
"name": "tab2",
"label": "Configurator Tab",
"type": "configurator",
"formulaName": "POAI_Test_Configurator"
},
{
"name": "tab3",
"label": "Visual Tree",
"type": "filtertree",
"treeName": "Segments",
"formulaName": "POAI_Test_FilterTreeNodeEval",
"selectedNodeIdsInputName": "selectedNodeIds"
}
]
}
]
}
Each Step has the following fields:
name* – Name of the step, unique in the Model Class.
label – Human readable name of the step, shown in the UI, it defaults to name if not specified.
description – Description of the step, shown in the UI.
calculation – Name of the calculation associated to this step, optional.
tabs – List of Tabs as defined below.
Each Tab has the following fields:
name* – Name of the tab, unique in the step.
label – Human readable name of the Tab, shown in the user interface. It defaults to name if not specified.
type* – Determines how the Tab will present the content to the user.
dashboard – The content of the tab will be rendered as a dashboard, i.e. input fields on left side, and portlets on the right side.
simple – The content of the tab will show the inputs generated by the logic on the left, and on the right side there will be results from the visible elements NOT as portlets, but rather one after another, as a simple list of values (e.g. similar to quote item calculation results).
configurator – The content of the tab will be rendered as the Configurator form, i.e. as a list of various input fields, whose values typically have dynamic dependencies.
filtertree – The content of the tab will be rendered as a tree (used for example to visualize the Segmentation Tree in the Negotiation Guidance Accelerator), and when you select/click one node, it will show details of the node on the right side, visualized as calculation results (e.g. similar to quote item calculation results).
formulaName* – Unique name of an existing logic. Depending on the field type (see above), the logic will serve different purposes, and it must be of different Natures.
For dashboard type – It must be Model Evaluation Logic which is expected to return input fields (filters) and content for portlets.
For simple type – It must be Model Evaluation Logic which is expected to return mostly simple types of values.
For configurator type – It must be Model Evaluation Configurator Logic which is expected to return the definition of form sections with input fields.
For filtertree type – It must be Model Evaluation Logic. It will be used when the user clicks a tree node, so the logic is expected to return detailed information about the selected node.
treeName* (only for filtertree type) – Name of the tree to be displayed in the Tab. The tree is expected to be stored in the model.
selectedNodeIdsInputName* (only for filtertree type) – Name of the input parameter passed to the evaluation logic, containing the list of selected node IDs.
Evaluation (evaluations)
This is used mostly for integration of the model with other modules. When another module (e.g. Quoting) needs to access the model results, it can either read the model data tables directly, or rather call the Evaluation to provide the data in a certain way. So, in a way, you can think of it as an API for reading the model.
Note that these Evaluations are not the same as the evaluations used to provide content for Tabs.
Example:
{
"evaluations": [
{
"name": "query_results",
"formulaName": "POAI_Test_QueryResults"
}
]
}
Each evaluation has the following fields:
name* – Name of the evaluation, unique in the Model Class.
formulaName* – Unique name of an existing logic of nature
model_evaluation, see also Model Evaluation Logic.
See Query Optimization Engine Results to execute an evaluation from another logic.