mor.reduction.reduceModel.ReduceModel

class ReduceModel(originalScene, nodeToReduce, listObjToAnimate, tolModes, tolGIE, outputDir, packageName='myReducedModel', addToLib=False, verbose=False, addRigidBodyModes=False, nbrCPU=4, phaseToSave=None, saveVelocitySnapshots=None, listPathToAnimation=[])

Bases: object

Main class that will perform the reduction

argument	type	definition
originalScene	str	absolute path to original scene
nodeToReduce	str	Paths to models to reduce
listObjToAnimate	list(ObjToAnimate)	list conaining all the ObjToAnimate that will be use to shake our model
tolModes	float	tolerance applied to choose the modes
tolGIE	float	tolerance applied to calculated GIE
outputDir	str	absolute path to output directiry in which all results will be stored
packageName	str	Which name will have the final componant ( & package if the option addToLib is activated)
addToLib	Bool	If True will add in the python library of this plugin the finalized reduced component
verbose	Bool	display more or less verbose
addRigidBodyModes	list(int)	List of 3 of 0/1 that will allow translation along [x,y,z] axis of our reduced model
nbrCPU	int	Number of CPU we will use to generate/calculate the reduced model
phaseToSave	list(int)	List of 0/1 indicating during which phase to save the elements/X0 by default will save during first phase

Methods

generateTestScene
performReduction	Perform all the steps of the reduction in one function
phase1	The step will launch in parallel multiple Sofa scene (nbrCPU by nbrCPU number of scene) until it has run all the scene in the sequence.
phase2	With the previous result obtain in during :meth:`phase1` we compute the modes
phase3	This step will launch in parallel multiple Sofa scene (nbrCPU by nbrCPU number of scene) until it has run all the scene in the sequence.
phase4	The final step will gather all the results in 1 folder and build a reusable scene from it
setListSofaScene	Will generate a list containing dictionnaries, where each dictionnary is a set of argument for the execution of one SOFA scene.

setListSofaScene(phasesToExecute=None, phase=None)

Will generate a list containing dictionnaries, where each dictionnary is a set of argument for the execution of one SOFA scene.

argument	type	definition
phasesToExecute	list(int)	Allow to choose which phase to execute for the reduction by default will select all the phase

The number of dictionnaries generated depend upon either the number of action possibility (self.reductionAnimations.nbPossibility) or you can give with phasesToExecute specifically which possibility you want to execute.

example :

You have 2 ObjToAnimate (thing that will be animated during the execution). From self.reductionAnimations you will have 2^2 possibilities:

[0,0] | [0,1] | [1,0] | [1,1] –> where 0 mean no animation & 1 animation

• if you give no argument, phasesToExecute = [0,1,2,3]

  it will execute possibilty 0,1,2 & 3

• if you give phasesToExecute=[1,3]

  it will execute possibility 1 & 3

performReduction(phasesToExecute=None, nbrOfModes=None)

Perform all the steps of the reduction in one function

argument	type	definition
phasesToExecute	list(int)	Allow to choose which phase to execute for the reduction more details see setListSofaScene
nbrOfModes	int	Number of modes you want to keep by default will keep them all

If you are sure of all the parameters this way is recommended to gain time

phase1(phasesToExecute=None)

The step will launch in parallel multiple Sofa scene (nbrCPU by nbrCPU number of scene) until it has run all the scene in the sequence.

argument	type	definition
phasesToExecute	list(int)	Allow to choose which phase to execute for the reduction more details see setListSofaScene

To run the SOFA scene in parallele we use the sofa launcher utility

What does it do to each scene:

• Add animation to each ObjToAnimate we want for our model in the predifined sequence

• Add a componant to save the shaking resulting states (WriteState)

• Take all the resulting states files and combines them in one file put in the debug dir with a debug scene

phase2()

With the previous result obtain in during :meth:`phase1` we compute the modes

See ReadStateFilesAndComputeModes for the way the modes are determined.

It will set nbrOfModes to its maximum, but it can be changed has argument to the next step : phase3

phase3(phasesToExecute=None, nbrOfModes=None)

This step will launch in parallel multiple Sofa scene (nbrCPU by nbrCPU number of scene) until it has run all the scene in the sequence.

argument	type	definition
phasesToExecute	list(int)	Allow to choose which phase to execute for the reduction more details see setListSofaScene
nbrOfModes	int	Number of modes you want to keep by default will keep them all

To run the SOFA scene in parallele we use the sofa launcher utility

What does it do to each scene:

• Take the previous one and add the model order reduction component:

  • HyperReducedFEMForceField

  • MappedMatrixForceFieldAndMas

  • ModelOrderReductionMapping

• Produce an Hyper Reduced description of the model

• Produce files listing the different element to keep

• Take all the resulting states files and combines them in one file put in the debug dir with a debug scene

phase4(nbrOfModes=None)

The final step will gather all the results in 1 folder and build a reusable scene from it

argument	type	definition
nbrOfModes	int	Number of modes you want to keep by default will keep them all

Final step :

• compute the RID and Weigts with ReadGieFileAndComputeRIDandWeights

• finalize the package

• add it to the plugin library if option activated