Loading this library from MatLab

MatLab supports importing .NET assemblies(*.dlls) through the built-in NET.addAssembly() command.

Note

Once an assembly is loaded in MatLab, it cannot be unloaded except by restarting the program.

Note

Not all .NET features are supported in MatLab a list of limitiations is given in https://se.mathworks.com/help/matlab/matlab_external/limitations-to-net-support.html

Step-by-step:

• Take a complete build of TimeSeriesAnalyis.dll and accompanying assemblies such as Accord.NET *.dlls and copy into your MatLab working directory such as C:\Appl\OneDrive\Documents\MATLAB (either build yourself or copy pre-built zip-file from https://github.com/equinor/TimeSeriesAnalysis/releases/ and unzip.)
• load the assembly with NET.addAssembly():
• import the methods of the .NET assembly using the MatLab command import
• you can now directly call methods in the "TimeSeriesAnalysis" namespace (consult the API reference)

A simple "hello-world" example:

assembly = NET.addAssembly('C:\Appl\OneDrive\Documents\MATLAB\TimeSeriesAnalysis.dll')
import TimeSeriesAnalysis.*
results = Vec.Add([1 2],[3 4])

will give as a result a Double[] where

>> result(1)

ans =

     4
>> result(2)

ans =

     6

Calling

>> assembly.Classes

reveals the classes that Matlab has loaded:

ans =

  30×1 cell array

    {'TimeSeriesAnalysis.RegressionResults'                    }
    {'TimeSeriesAnalysis.Array2D'                              }
    {'TimeSeriesAnalysis.Array2DExtensionMethods'              }
    {'TimeSeriesAnalysis.Matrix'                               }
    {'TimeSeriesAnalysis.Vec'                                  }
    {'TimeSeriesAnalysis.VecExtensionMethods'                  }
    {'TimeSeriesAnalysis.Utility.CSV'                          }
    {'TimeSeriesAnalysis.Utility.StringToFileWriter'           }
    {'TimeSeriesAnalysis.Utility.TimeSeriesCreator'            }
    {'TimeSeriesAnalysis.Utility.UnixTime'                     }
    {'TimeSeriesAnalysis.Utility.Plot'                         }
    {'TimeSeriesAnalysis.Utility.Plot4Test'                    }
    {'TimeSeriesAnalysis.Utility.ParserFeedback'               }
    {'TimeSeriesAnalysis.Utility.SignificantDigits'            }
    {'TimeSeriesAnalysis.Dynamic.DefaultProcessModel'          }
    {'TimeSeriesAnalysis.Dynamic.DefaultProcessModelIdentifier'}
    {'TimeSeriesAnalysis.Dynamic.DefaultProcessModelParameters'}
    {'TimeSeriesAnalysis.Dynamic.BandPass'                     }
    {'TimeSeriesAnalysis.Dynamic.HighPass'                     }
    {'TimeSeriesAnalysis.Dynamic.LowPass'                      }
    {'TimeSeriesAnalysis.Dynamic.PIDModel'                     }
    {'TimeSeriesAnalysis.Dynamic.PIDModelParameters'           }
    {'TimeSeriesAnalysis.Dynamic.PIDAntiSurgeParams'           }
    {'TimeSeriesAnalysis.Dynamic.PIDcontroller'                }
    {'TimeSeriesAnalysis.Dynamic.PIDgainScheduling'            }
    {'TimeSeriesAnalysis.Dynamic.PIDscaling'                   }
    {'TimeSeriesAnalysis.Dynamic.PIDtuning'                    }
    {'TimeSeriesAnalysis.Dynamic.SubProcessDataSet'            }
    {'TimeSeriesAnalysis.Dynamic.TimeDelay'                    }
    {'TimeSeriesAnalysis.Dynamic.MovingAvg'                    }

It is possible to load classes form namesapces like TimeSeriesAnalysis.Utility and TimeSeriesAnalysis.Dynamic as well, for instance:

>> import TimeSeriesAnalysis.*
>> import TimeSeriesAnalysis.Utility.*
>> TimeSeriesCreator.Step(25,50,0,1)

returns a Double[] of length 50.