ParameterInfo

Overview

• ParameterInfo - Data structure for documenting algorithm parameterizations, translatable to python and inspectable in MeVisLab networks
• BaseWithParameterInfo - Base class for ParameterInfo supporting objects providing a getParameterInfo() call
• parameter_info_utils - Namespace with datatype conversion utilities (mostly from ML to QVariant-derivates)
• MLParameterInfoWrapper - Wrapper that allows to call .object().toDict() on a field holding a ParameterInfo object, \ providing a python dict (that can be converted to a true python ParameterInfo object via parameter_info.utils.toParameterInfo())
• MLBaseWithParameterInfoWrapper - Wrapper that allows to call .object().getParameterInfo() on a field holding a BaseWithParameterInfo object, \ also providing a python dict (conversion to true python ParameterInfo see above)

Example

Say, we want to disclose the parameterization of a class MyAlgorithm:

The header file of an example object we want to provide the parameter info of:

#include <mlParameterInfo.h>
 
ML_START_NAMESPACE
  
class MyAlgorithm : public BaseWithParameterInfo {
 
private:
  ParameterInfo _getParameterInfo() const override;
 
  int _exampleInt{ 4711 };
  double _exampleDouble{ 47.11 };
  std::string _exampleString{ "Some String" };
 
  // Complex data types
  ml::ImageVector _exampleImageVector{ 1, 2, 3, 4, 5, 6 };
  ml::Vector3 _exampleFloatVector{ 1.1, 2.2, 3.3 };
  ml::SubImageBox _exampleSubImageBox{ { 1, 2, 3, 4, 5, 6 }, { 1, 2, 3, 4, 5, 6 } };  
 
  ML_CLASS_HEADER( BaseWithParameterInfoExample )
};
 
ML_END_NAMESPACE

In the MeVisLab network, we would want the ParameterInfoInspector module to show for that object:

{
  "ExampleInt": 4711,
  "ExampleDouble": 47.11,
  "ExampleString": "Some String",
  "ComplexDatatypes": {
    "ExampleImageVector": [ 1, 2, 3, 4, 5, 6 ],
    "ExampleFloatVector": [ 1.1, 2.2, 3.3 ],
    "ExampleSubImageBox": [ [ 1, 2, 3, 4, 5, 6 ], [ 1, 2, 3, 4, 5, 6 ] ]
  }
}

We could achieve this with this implementation in the corresponding source file:

// [...]
 
void MyAlgorithm::_getParameterInfo() const {
  ParameterInfo pi;
  // Values need to be (convertible to) QVariant. This is how it works:
  pi[ "ExampleInt" ] = _exampleInt;
  pi[ "ExampleDouble" ] = _exampleDouble;
  pi[ "ExampleString" ] = QString::fromStdString( _exampleString );
  ParameterInfo complexPi;
  complexPi[ "ExampleImageVector" ] = parameter_info_utils::ToQVariantList( _exampleImageVector );
  complexPi[ "ExampleFloatVector" ] = parameter_info_utils::FloatVectorToQVariantList( _exampleFloatVector );
  complexPi[ "ExampleSubImageBox" ] = parameter_info_utils::ToQVariantList( _exampleSubImageBox );
  pi[ "ComplexDatatypes" ] = complexPi;
  return pi;
}
 
ML_CLASS_SOURCE( MyAlgorithm, BaseWithParameterInfo )
 
// [...]