cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 

How to extract model data of 3d models in Vuforia Studio (without external Tools)?

Level 12

How to extract model data of 3d models in Vuforia Studio (without external Tools)?

1.)The first question it is possible to extract model data of 3d models in Vuforia Studio?

In this case we have a model widget with an assembly model without explicit modelitem  widget definitions. The question is if we can extract data for the components and if yes then what data we can extract.

Extracting of data is possible only in Preview mode. So we have in preview mode the method tml3dRenderer.GetObject() where we can access an model object (an component) example:

 

let comp_widget=tml3dRenderer.GetObject(selection).GetWidget()

 

where selection is some thing like "<modelname>-<compPath>" e.g. : "model-1-/0/0/3/2"

Then from the widget we can extract data:

 

 var loc=tml3dRenderer.GetObject(selection).GetWidget().GetLocation()
 console.error("DEBUG getObj.GetWidget()")
 console.warn(tml3dRenderer.GetObject(selection).GetWidget())

2019-03-19_13-07-57.gif

 

When we   explore there the different methods we will find methods to  get or set properties. To extract data, we can use the get... methods.

The methods tml3dRenderer.GetObject() seems currently not to work in Vuforia View on end devices so that we will not able for example to get the data of a component selection. So means we need a way to extract data in Preview mode and make it available in the Vuforia view on the end device. Here I did not find a methods to extract the original component name but I was able to create a list (json) with the position data ( I did not add color but this is possible e.g. tml3dRenderer.GetObject(selection).GetWidget().GetColor())

We can create a json  e.g. of  following data:

 

{"model-1-/0/0/0":{"valid":false,"orientation":{"x":0,"y":0,"z":0},
"size":{"x":1,"y":1,"z":1},"scale":{"x":1,"y":1,"z":1},
"position":{"x":9.999999998199587e-24,"y":9.999999998199587e-24,"z":9.999999998199587e-24}},
"model-1-/0/0":{"valid":false,"orientation":{"x":0,"y":0,"z":0},"size":{"x":1,"y":1,"z":1},
"scale":{"x":1,"y":1,"z":1},"position":{"x":0,"y":0,"z":0}},
"model-1-/0/0/2":{"valid":false,"orientation":{"x":0,"y":90,"z":0},"size":{"x":1,"y":1,"z":1},
"scale":{"x":1,"y":1,"z":1},"position":{"x":0,"y":0.029500000178813934,"z":-5.51091050576101e-18}},
...}

and assign it to variable e.g. $scope.COMP_LOCs

So later we can on end device read the current position data:

var selection_location=$scope.COMP_LOCs[l_currentSelection]  
//read the location data from json varible console.log("selection:"+l_currentSelection+"->X= "+ selection_location.position.x);
//print it to console selection_location.position.x= round(parseFloat(selection_location.position.x) + 0.005,4)
//add 0.005 shift and round to 4 dec

2019-03-19_14-01-55.gif

 

2.) For the scenario on point 1.)  the data should be extracted. Actually  we have all methods to extract the data but what we do not have is the selection itself . We have the model widget name e.g. “model-1” but   we do not have the component ID paths. To be able to construct a selection handle we need to construct the ID path of an component and then we need to check if it exist. This is some kind of graph search where we have an assembly with a components tree where the edges are the ids of the components.

e.g. /0/0/1/1 , /0/0/1/2, /0/0/1/4, … etc.

One possible algorithm is the deep first search:

 

2019-03-19_14-14-40.gif

 

To implement this I used the following javaScript code:

 

/////////////////////////////
var max_asm_depth=6; //this is the max depth in Creo Parametric
var max_numb_comp_asm=25;

/////////////////////////// ->deep first function check_comp_deep_first_recursively(target,path,arr) { //console.warn("called check_comp_deep_first_recursively(target="+target+",path="+path+")"); var selection = target+'-'+path var path_array = path.split('/') var depth = parseInt(path_array.length) var num = parseInt(path_array[depth -1]) var prev_num = parseInt(path_array[depth -2]) var prev_path = '' for (var i=1;i < depth -1;i++) {prev_path= prev_path +'/' + path_array[i]} if( check_for_valid_selection(selection) == 1) { arr[selection]=tml3dRenderer.GetObject(selection).GetWidget().GetLocation() if( (depth+1) < max_asm_depth) check_comp_deep_first_recursively(target, path + '/0', arr) else { if(num +1 < max_numb_comp_asm) check_comp_deep_first_recursively(target, prev_path + '/'+(num +1), arr)} } else { var right_num = num +1 if(right_num < max_numb_comp_asm) check_comp_deep_first_recursively(target, prev_path + '/'+right_num, arr) else if(!Number.isNaN(prev_num) ) {//console.log("--2") prev_path = '' for (var i=1;i < depth -2;i++) {prev_path = prev_path +'/' + path_array[i]} prev_path = prev_path +'/' + (prev_num +1) check_comp_deep_first_recursively(target, prev_path , arr) } } } //////////////////////////
///call of the function:
$scope.compJSON_loc_Data = new Object();
var target="model-1"
check_comp_deep_first_recursively(target,'/0',$scope.compJSON_loc_Data)
...

The code above has the following weak spot - I need to give the maximum depth (max_asm_depth) and the maximum possible branches (max_numb_comp_asm

The maximum depth currently in Creo assembly is 25 so that value which > 25 will not make a sense. 

The value of  max_numb_comp_asm  in a flat assembly (only one level of depth) corresponds to the number of the components - the maximum number of branches on particular level of depth

 

The another possible algorithm is the breadth first search:

 

2019-03-19_14-50-12.gif

To implment this  I used the following JavaScript code:

 

/////////////////////////////
var max_asm_depth=6; //this is the max depth in Creo Parametric
var max_numb_comp_asm=25;

/////////////////function check_comp_at_level(target,num,depth,arr) // ->breadth first function check_comp_at_level(selection,num,depth,arr) { var position =''; // console.log("call check_comp_at_level =>"+selection); try{ // console.log("====== check here ==========="); //console.warn(tml3dRenderer.GetObject(selection).GetWidget().GetLocation()); var loc=tml3dRenderer.GetObject(selection).GetWidget().GetLocation() if( (loc.scale.x == 0) || (loc.scale.y == 0) || (loc.scale.z == 0) ) return 0; // the scale could not be zero //position= tml3dRenderer.GetObject(selection).GetWidget().GetLocation().position //console.warn(position); //arr[selection]=position arr[selection]=loc return arr[selection]; } catch (e) {console.error("failsed with error="+e); return 0;} } /////////////////////////// function check_comp_at_level_recursively(selection,depth,arr) { //console.warn("called check_comp_at_level_recursively("+selection+",depth="+depth+")"); var num =0; if(depth >= max_asm_depth) { //console.log("maximum depth of max_asm_depth ="+max_asm_depth+" reached"); return 0;} for (num=0;num < max_numb_comp_asm; num++) { var currentSelection =selection+'/'+num if(depth <0) return 0; var pos = check_comp_at_level(currentSelection,num,depth,arr) if(pos ==0 ) { continue;} else {check_comp_at_level_recursively(currentSelection,(depth+1),arr) } } //end of for } ////////////////////////// //////////////////////////////// function check_for_valid_selection(selection) { //console.log(" check_for_valid_selection =>"+selection); try{ var loc=tml3dRenderer.GetObject(selection).GetWidget().GetLocation() if( (loc.scale.x == 0) || (loc.scale.y == 0) || (loc.scale.z == 0) ) return 0; return 1; } catch (e) {console.error("failsed with error="+e); return 0;} } ///////////////////////////
///call of the function:
$scope.compJSON_loc_Data = new Object();
var target="model-1"
check_comp_at_level_recursively(target,'/0',$scope.compJSON_loc_Data)
...

 

 

The code for the breadth first search  uses also the parameters for  maximum depth (max_asm_depth) and the maximum possible branches (max_numb_comp_asm)  - so means it have the some restriction. If we set a value wich is large this will increase the time until the search is completed so therefore depending of the assembly we need to set the both parameter properly (need to be able to scan the whole assembly but to minimize the search time)

For different assemblies the first deep or first breadth could lead to better results. For example, for flat assembly structures the better approach will be to use the first breadth algorithm 

But accutaly the performance is not so important here, because the serch will be called one time and  then the json list should be saved. 

With the current functionality  we can read a file (json file ) from the project  upload directory , but it seems not to be possible to save and information to a file there.

To read a json file form the upload folder we can use some code like this:

 

 

target='model-1'
$http.get('app/resources/Uploaded/' + jsonFile).success(function(data, status, headers, config) { $scope.compJSON_mod=data; // in this case the data is the received json object
angular.forEach(data , function(color, path_id){ $scope.compJSON_Data[path_id] =position; console.log("target="+target+" --> $scope.compJSON_Data["+path_id+"] = "+$scope.compJSON_Data[path_id]); });//end of the error function ////////// finish for each path_id }) .error(function(data, status, headers, config) {console.log("problem in the http will create a new ");

2019-03-19_15-36-40.gif

When we want to save data  (the generated json list) we need to use another workaround - we can use an thingworx repository. Following functions /events could be used to save and receive an json object to/from a twx repository:

 

// the methods SaveJSON and LoadJSON
// for the repository object should have 
//run permision for  es-public-access user
//////////////////////////////////////////////////////////////
$scope.SaveJsonToTwxRepository = function(path, content) {
 $scope.$applyAsync(function() {
    $rootScope.$broadcast('app.mdl.CAD-Files-Repository.svc.SaveJSON',
                          {"content": content, "path":path}

                         );}   ,500 );

};
//////////////////////////////////////////////////////////////
$scope.GetJsonFromTwxRepository = function(path) {
 $scope.$applyAsync(function() {
    
    $rootScope.$broadcast('app.mdl.CAD-Files-Repository.svc.LoadJSON',
      
                          {"path":path}

                         );}   ,500 );
   $scope.app.speak("after call of GetJsonFromTwxRepository")

//in the modelloaded listener register
// LoadJSON-complete event -> to laod the data into session
rootScope.$on('modelLoaded', function() {
////
$scope.$root.$on('LoadJSON-complete', function(event, args) {
console.log("LoadJSON-complete event");
$scope.COMP_LOCs=args.data
console.log(JSON.stringify( $scope.COMP_LOCs))
});
///
});

 

In  the code above I use the 'modelloaded' listener to register LoadJSON-complete event . Because the service is called asyncronously- we need this event to load the data into session when it is received from thingworx.

Here in this example the repository object is named "CAD-Files-Repository"

The Thingworx services should have run permission and it is required to be added in the external data panel :

 

2019-03-19_15-49-29.gif

So when we start the project in PREVIEW mode we can call the search for the assembly structure and save it then  to thingworx. In Vuforia View mode   then we can received json object from thingworx.

To check the current mode we can use 

 

if(twx.app.isPreview() == true)
...

 

it will  check if the current mode is preview mode or Vuforia View on the end device - here an example of the workflow:

 

if(twx.app.isPreview() == true) 
{// preview mode  
//calling   breadth first - test
check_comp_at_level_recursively(target+'-',0,$scope.compJSON_POS_Data)  
//console.warn($scope.compJSON_POS_Data)
  //calling   deep first a second test and generating a data - locations
check_comp_deep_first_recursively(target,'/0',$scope.compJSON_loc_Data)
console.log("========================================")
console.log("$scope.compJSON_POS_Data ->breadth first")
console.log("========================================")
console.log(JSON.stringify($scope.compJSON_POS_Data))
console.log("========================================")
console.log("")    
console.log("")
console.log("========================================")
console.log("$scope.compJSON_loc_Data ->deep first")
console.log("========================================")
console.log(JSON.stringify($scope.compJSON_loc_Data))
$scope.SaveJsonToTwxRepository('/CADFiles/json_lists/compJSON_loc_Data.json',$scope.compJSON_loc_Data)
$scope.GetJsonFromTwxRepository('/CADFiles/json_lists/compJSON_loc_Data.json')
console.log("========================================")
console.log("")   
 }
 else { //here is the part on mobile device
$scope.GetJsonFromTwxRepository('/CADFiles/json_lists/compJSON_loc_Data.json')    
                } 

 

I tested all points of  the described techniques above in a project which I want to provide here as zip file for the HoloLens (hideComponetsHoloLens .zip):

 

2019-03-19_16-37-11.gif

 

So to be able to test it you need to create in Thingworx a repository thing - means a thing which uses  the thing template "FileRepositroy" with the name "CAD-Files-Repository" and create a folder there "/CADFiles/json_lists/" (if you use another name and another folder (e.g. "/" no folder - the root repository folder) you have to adapt the javaScript code:

 

...
/CADFiles/json_lists/compJSON_loc_Data.json
...
app.mdl.CAD-Files-Repository.svc.SaveJSON'
...
app.mdl.CAD-Files-Repository.svc.LoadJSON'

 

2 REPLIES 2

Re: How to extract model data of 3d models in Vuforia Studio (without external Tools)?

Hello Roland,

once again, thank you so much for your contributions! Your articles are always very useful for so many users. Right now, it is very complicated to get the attributes from the CAD-Data back into Vuforia Studio. It would be great, if PTC could provide tools to select, which data should be filtered, and which data should be given to Vuforia. Like in this case, this­ could be orientation: Assign Errors, Warnings, live data, etc. to a part of the model.

It could also be helpful to have attributes like possible warnings. Then you could highlight all parts, that may be hot or that need to be handled with safety glasses on.

You wouldn’t have to assign these warnings by hand, which could safe so much time.

I hope, that my message might reach R&D to provide them helpful input.

 

Greetings

whity

Re: How to extract model data of 3d models in Vuforia Studio (without external Tools)?

Hello @whity ,

I am sorry for the delay, so will try now to provide a feedback to your request

 

it is very complicated to get the attributes from the CAD-Data back into Vuforia Studio.

 

>>> Yes, I agree it is complicated without to develop any tool with Creo View Toolkit

 

It would be great, if PTC could provide tools to select, which data should be filtered, and which data should be given to Vuforia.

 

>> so, depending on the requirements such tools could be developed via Toolkit: When data is coming form Creo Parametric without using of Illustrate we can extract all required data with Creo Parametric Toolkit Modules ( Toolkit C ansi and cpp, OTK CPP and OTK Java or J-Link and Web.link – also VBA API) The problem is that this are developments moduls which required a considerable invest in cost and in work. I do not believe that could be provided generally for all possible configurations.

For the case that we use Creo Illustrate we need to generate the model items list again because illustrate will change the path ids. We can use Creo View toolkit modus to extract data form .pvz models (also such edited with Creo Illustrate)

I did describe in some Studio Tech tips the possible usage scenarios.

 

 

Like in this case, this­ could be orientation: Assign Errors, Warnings, live data, etc. to a part of the model. It could also be helpful to have attributes like possible warnings.

 

>>> yes such functionality could be nice , but need to be discussed in the product management story board. Acutely such story board are created when customers has an projects with PTC – e.g. development / success plan with PTC – in this  case PTC Technical Support and consultants peoples will investigate the customers requirements and will try to report all possible enhancement points for further development.

 

Then you could highlight all parts, that may be hot or that need to be handled with safety glasses on.

 

>>> I think we can implement the most things also now with some invest in the javaScript (angular.js) But we need an exact requirements description what attribute, what behavior ….

 

You wouldn’t have to assign these warnings by hand, which could safe so much time.

I hope, that my message might reach R&D to provide them helpful input.

 

>>>The problem is that R&D will change functionality only if there is some development or enhancement request which is approved by Product management.

Basically such general description  where the info of the problem is more or less abstract is difficult to be handled. Enhancement could be reported if you contact some of the marking and PM Product specialists on some of the PTC Events (e.g. PTC liveworx)

For example I could  try to report an enhancement – but in this case I need to have exact definition of one particular problem definition( atom) where the development could good estimate the development costs and investment risk and where we have also a proof that this will increase the tool productivity considerably

Announcements

Surveys are coming to the Vuforia Community!  Your feedback is valuable to us!  When prompted, please take our 60 second survey!