eduardo simioni

pipeline

creating a dynamic menu system in Motionbuilder

by on Feb.16, 2014, under GUI, MotionBuilder, pipeline, python

There are two main tricky parts in doing this:

1) Organizing your tools
2) The event function

Prior to face this sort of task you need to be aware of the concepts explained on the previous post, python modules and import system and have a very very very good understanding of FBMenuManager.

Organizing your tools

In short, there are two ways of accomplishing this. One is having each menu entry as a simple .py script which is executed with execfile(), the other is to think of each tool as a module/folder. If you want to group stuff into different menus, it helps to think of each menu as a different empty module/folder as well.

So let’s suppose you want a menu structure inside Motionbuilder like this:

  • Scripts
    • Cinematics
      • Mocap Cleaner
      • Mocap Destroyer
    • Gameplay
      • Import Tool
      • Export Tool
    • Generic
      • Animate_Everything
      • Fix_Animation

Cinematics, Gameplay and Generic are the main menus, and they show on the right of “Help”. Mocap Cleaner, Mocap Destroyer, Import Tool, Export Tool are (imaginary) full featured tools, each one in its own folder with multiple files. Animate_Everything and Fix_Animation are simple .py scripts, and they are grouped in one single folder. Like this:

Files organization

The Generic scripts is quite simple to do and it’s covered down below. The way I did for full featured tools though is, obviously, creating one menu entry for each folder inside the Scripts folder. Each of these “menu folders” had a __init__.py (see below in initialization) with some variables describing their content and would become a main entry besides the “Help” menu. Each folder would then be added to sys.path and create one of these data structures:

class MenuData():
    '''
    This stores all necessary information for one new menu to be constructed it`s sub menu items.
    '''
    def __init__(self, sMenuName, sFullPath, mMenuModule):
        # menu information
        self.sMenuName   = sMenuName
        self.mMenuModule = mMenuModule
        self.sFullPath   = sFullPath
        # actual FBGenericMenu object for the menu
        self.oMenu       = None
        # the key one, list of tools in tuples, 0 is name, 1 is module or full script path (for generics)
        self.lmTools     = []

So you end up with a list of MenuData’s, one for Cinematic, Gameplay and Generic. Now you need to create the actual FBGenericMenu for each folder. You could simplify this into one single for loop, if you have all tools into the Scripts root you could even do away with the MenuData (which is to create one menu per folder), but I wanted to do something generic and support many different outcomes so I was doing it in two loops described below.

Initializing Tools:

For each MenuData I look for each subfolder (Scripts/Gameplay/Import, Scripts/Gameplay/ExportTool for example) and save each module (present in this subfolder) into MenuData.lmTools. The tricky part is how to do this:

for oMenu in listOfMenuData:
    lDir = os.listdir( oMenu.sFullPath )
    for sPath in lDir:
        sMenuPathWithSubFolder = oMenu.sFullPath + "/" + sPath
        if os.path.isdir( sMenuPathWithSubFolder ) and os.path.isfile( sMenuPathWithSubFolder+"/__init__.py" ):
            module = __import__( sPath )
            oMenu.lmTools.append( (sPath, module ) )

Line 6 works because each “Menu/Module” path was added to sys.path, so you can easily import each tool without its entire path. The important line though is the last one. It adds a list of tools to MenuData.lmTools with a tuple that contains (“string to the tool module”, “the tool module itself”). And here is why you need to add the menu folder to sys.path. You could still get the tool module with other means, but by having Gameplay in sys.path you can do import ImportTool on other tools and share code between them.

On the tool folder you need a __init__.py so python can import that as a package. You can (or need) to put variables in there, so the menu system can create the proper menu entry for the tool (to use a custom name, for example). This is an example of what I’m using on /Scripts/Gameplay/Destroyer/__init__.py:

sToolFullName = "Mocap Destroyer"
sStartFileToImport = "destroyerDialog"
sCustomFunction = "mainUI()"

It is important that there’s no code running in this __init__. This file is loaded and ran during Motionbuilder startup, so if there’s something crashing in there the entire menu system won’t load (unless you don’t trust your colleagues and populate it with try/catches).

And here is where python magic happens. Once you loaded the module in MenuData you can access all the contents from that __init__.py with for example:

print oMenu.lmTools[1].sToolFullName
print oMenu.lmTools[1].sTartFileToImport

And this feature you will need to use on the next two steps.

Create menu entries

You haven’t created anything yet because you need to know WHAT to create. That’s why I was importing each tool __init__.py before actually creating FBGenericMenuItems. This way the title of the menu item can be set on the tool __init__.py itself, instead of relying on a separate file altogether.

But when dealing with FBGenericMenu is best to just use code instead of describing it:

oMenuMngr = fb.FBMenuManager()
for oMenu in listOfMenuData:
    # FBGenericMenu
    oMenuMngr.InsertLast( None, oMenu.sMenuName )
    oFBMenu = oMenuMngr.GetMenu( oMenu.sMenuName )
    oFBMenu.OnMenuActivate.Add( EventMenu )
    oMenu.oFBMenu = oFBMenu

    # create menu contents (MenuItems) for tools
    print ('create tool: '+ oMenu.sMenuName)
    for i in range(len(oMenu.lmTools)):
        oTool = oMenu.lmTools[i][1]
        if hasattr( oTool, 'sToolFullName' ):
            sName = oTool.sToolFullName
        else:
            sName = oMenu.lmTools[i][0]
        oFBMenu.InsertLast( sName, i ) # you need the "i"ndex for the event function below (although you could use a string instead)

As you can see, a FBGenericMenu is stored in MenuData.oFBMenu. You will need it for the EventMenu function. After that we just iterate through each stored module, creating a menu entry for each one. If the module has a sToolFullName, we use that one for the menu entry, otherwise we use its folder name.

EventMenu

All menu entries call just one function. You don’t want to generate a new function for each menu entry, otherwise you will end up with some heavy meta-python coding. And you shouldn’t need to. EventMenu is called with two variables, control and event. Control is the actual FBGenericMenu that was clicked. Event.Id has the index which you used to add that menu entry, last line of the code snippet above. You also have a Event.Name, which is the actual string being displayed in the interface. With this cards in your pocket you can do this:

def EventMenu( control, event ):
    '''
    Runs the script selected from the menu. Each menu entry has a corresponding tool __init__ialized (but not loaded in memory)
    This is stored in MenuData.lmTools[1], and it should have variables/attributes point what should be imported and
    how the tool should be executed. Defaults to main() in /Scripts/MENU/TOOL/TOOL.py.
    '''
    #print control, event.Id, event.Name

    # find which menu was clicked and get it's MenuData from loMenus
    oMenu = None
    for oTM in loMenus:
        if control == oTM.oFBMenu:
            oMenu = oTM
            break

    sToolName = oMenu.lmTools[event.Id][0]
    oTool     = oMenu.lmTools[event.Id][1]

    # no sStartFileToImport, so we file to import is with same name as folder, as in: /Script/Tool/Tool.py
    if not hasattr( oTool, 'sStartFileToImport' ):
        sModule = oMenu.sMenuName+"."+sToolName+"."+sToolName
    # with sStartFileToImport, so we have /User/Script/Tool/myTool.py
    else:
        sModule = oMenu.sMenuName+"."+sToolName+'.'+oTool.sStartFileToImport

    #print "Importing:\n{0}".format(sModule)
    __import__(sModule)

    # if it doesn't have sCustomFunction we run main()
    if not hasattr( oTool, 'sCustomFunction' ):
        sModuleFunction = sModule + '.main()'
    # otherwise, we run the named function
    else:
        if oTool.sCustomFunction.find('(') == -1:
            sModuleFunction = sModule + '.' + oTool.sCustomFunction + '()'
        else:
            sModuleFunction = sModule + '.' + oTool.sCustomFunction

    #print "Evaluating:\n{0}".format( sModuleFunction )
    eval( sModuleFunction, sys.modules )
    return

First, we find which main menu (the ones beside Help) the “click” happened, by comparing all MenuData.oFBMenu to control. Once we have the MenuData, we have a list of tools inside, which we can quickly fetch with event.Id, since they were added in order and each have its own unique id. Another option would be to compare strings, trying event.Name in a list of tool names, but since we allow tools to have sToolFullName and comparing string is lower than using an index directly, index is better. This is the shitty FBMenuManager part, if you don’t get how this is working you need more time playing with FBMenuManager, FBGenericMenu and FBGenericMenuItem (and/or I haven’t worded it very clearly).

Once we have the actual tool module (/Scripts/Gameplay/ImporterTool/__init__.py) we can run the actual tool. I wanted to add some flexibility, so I allowed sStartFileToImport variable in the folder __init__.py, which points to the actual python file that will be imported. If this variable doesn’t exists, it will try to import a python file with the same name as the folder.

Now though is when python import magic actually gets in the way.

Python “import” command runs the imported script just once. When you do import again, it’s not re-evaluating the imported code. Even if you changed the script locally, you need to do reload module instead, which can cause a series of other problems.

Instead of making self-running scripts, I wanted to also add the flexibility for a function to be called for the tool to start (say you want to run something based on selected objects, for example). That’s why it’s also checking for a sCustomFunction variable. If there’s none, it will use just run (sModule + ‘.main()’).

Lastly, it does eval. Which runs said function.

Generic menu

For the generic menu, instead of loading a module I just save the path for the script in MenuData.lmTools[(“script name”, “full path for script”)] and run them with:

execfile( oMenu.lmTools[event.Id][1], {} )

You will just need to add something in /Scripts/Generic/__init__.py (like isGenericMenu=True) so you know that that folder is a generic menu and should run another exec procedure. The empty dictionary {} at the end is to specify a new empty set of locals and globals, otherwise you get scope problems when running scripts without functions.

Leave a Comment :, , , , , , , , , more...

sharing files between 3dsMax and Motionbuilder

by on Oct.28, 2013, under 3ds Max, ini, maxscript, MotionBuilder, pipeline, python, xml

Here’s a small tip on where to quickly put files (ini or xml for example) that you can easily fetch from both Max:

systemTools.getEnvVariable("LOCALAPPDATA") 

And Motionbuilder:

import os
os.getenv('LOCALAPPDATA')

Both will return “C:\Users\USER\AppData\Local”. And if you ever wondered what’s the difference between /AppData/Local and AppData/Roaming this might shed some light on it: http://superuser.com/questions/150012/what-is-the-difference-between-local-and-roaming-folders

Leave a Comment :, , , , , , more...

scale all the things!

by on Aug.16, 2012, under 3ds Max, maxscript, pipeline

Here’s a nice trick my mate at Funcom, Endre Eikrem, told me. You can scale everything, scene, mesh, skin, rig, animation, with this simple code:

units.SystemScale = 0.5
holdMaxFile()
units.SystemScale = 1.0
fetchMaxFile useFileUnits:false quiet:true

You just need to have “Respect System Unit in Files” activated, inside Customize/Unit Setup/System Unit Setup. That’s it!

Except! There’s always some exception. So far we found a couple of stuff that are not scaled with this trick:

  • Wire Parameters. They are strings, so it’s expected that Max doesn’t scale them.
  • Key Tangents. If tangents were manually edited they are not going to be scaled. On the other hand, animation keys are scaled correctly.
  • Float Limit Controller. It’s value is not scaled.

There might be other exceptions, but so far these are quite easy to workaround with Maxscript. Specially if your rig is simple. On Funcom rigs, we just needed to parse Orientation_Constraint controllers and edit their wire parameters to the FK/IK handle. Float_Limit controller is just a matter of finding it and setting a new value based on the units.SystemScale used. And key tangents are not to worry if you first bake/plot the animation on the handles. Keyframes being scaled, it works flawlessly.

If anyone trying this out find any more exceptions please let me know.

2 Comments :, , , , , , more...

xml with dotnet on maxscript

by on Apr.25, 2011, under 3ds Max, pipeline, xml

Paul Neale has a very good tutorial about dotnet and maxscript, specially reading and writing xml, but I decided to add some more information from a slightly different approach. First, let’s translate an XML to dotnetesque:

<rootElement>
    <element1 />
    <element2 attributeName="attribute content">
        <element2.1>innerXML of this element</element2.1>
    </element2>
</element>

And the entire document is a XmlDocument (instance of dotNetObject “System.Xml.XmlDocument“)

There are many ways to read and write XML with dotnet. If you need to read the whole document first, probably the most optimized is by using XmlParser, but it can be a bit of a pain. If you are using Maxscript you are not doing it in realtime, so a couple miliseconds won’t make that much of a difference.

Anyway, if you are looking for specific information inside tags you can easily do it using .GetElementsByTagName:

	XmlDoc = dotNetObject "System.Xml.XmlDocument"
	XmlDoc.load "D:\\Characters\\Rider\\exportEMFX.xml"

	-- I know <baseCharacter> has only one entry, so I get the first one
	tagBaseChar = XmlDoc.GetElementsByTagName "baseCharacter"
	baseCharacter = (tagBaseChar.item 0).InnerXML
	
	-- I don't know how many <file>s there is, so I get them into an array
	tagFiles = XmlDoc.GetElementsByTagName "file"
	fileList = (for i=0 to (tagFiles.count-1) collect (tagFiles.item i).InnerXML)

To read something like:

<mlExportEMFX>
    <baseCharacter>D:\\Characters\\Rider\\rider.max</baseCharacter>
    <filesToLoad>
        <file>D:\\Characters\\Rider\\FBX\\stopNeutral_L.fbx</file>
        <file>D:\\Characters\\Rider\\FBX\\stopNeutral_R.fbx</file>
    </filesToLoad>
</mlExportEMFX>

Note that GetElementsByTagName doesn’t care about hierarchy, it just finds the element you are looking for, no matter where in the tree it is or how many times it’s repeated. If you have a “file” under “filesToLoad” and also under “filesToSave” it won’t differentiate between them.

If you have a tree/graph structure and you don’t know each tag name one solution is to recourse through it all loading them on a dictionary or array.

Say you have an hierarchy structure you want to save, together with their layers, into an XML. You can store the data in two arrays with:

/*
this two are recursive functions to return the hierarchy in a single array
*/
fn GetHierarchyTree lParent =
(
	toRet = #(lParent.name)
	for o in lParent.children do
		append toRet (GetHierarchyTree o)
	toRet
)
fn GetLayerTree lParent =
(
	toRet = #(lParent.layer.name)
	for o in lParent.children do
		append toRet (GetLayerTree o)
	toRet
)

/*
this is just an example, you shouldn't use globals like this
*/
global parentTree = #()
global layerTree = #()

/*
gets the data into parentTree and layerTree
*/
fn getData =
(
	for o in objects where o.parent == undefined do
	(
		append parentTree (GetHierarchyTree o)
		append layerTree (GetLayerTree o)
	)
)

To save this into an XML you also go with recursion:


fn recurseTreeToXML baseNode xmlDocNew pTree lTree =
(
    for i=1 to pTree.count do
    (
		-- if it's a leaf, there's data, so we append
        if (isKindOf pTree[i] String) then
        (
            curNode = xmlDocNew.CreateElement pTree[i]
            curNode.SetAttribute "layer" lTree[i]
            baseNode.appendChild curNode
            baseNode = curNode
        )
        else -- it's a branch with more data underneath
        (
            recurseTreeToXML baseNode xmlDocNew pTree[i] lTree[i]
        )
    )
)
 
fn saveDataOnXML =
(
	-- you cannot append elements to an xmldoc
	-- you have to append elements to a single root element, child of xmldoc
    xmlDocNew = dotNetObject "System.Xml.XmlDocument"
    xmlRoot = xmlDocNew.CreateElement (getFilenameFile maxfilename)
    xmlDocNew.appendChild xmlRoot
 
    recurseTreeToXML xmlRoot xmlDocNew parentTree layerTree
 
    xmlDocNew.save ("C:\\temp\\test.xml")
)

And also to read an XML:


fn recurseXML docElement =
(
	tempPTree = #(docElement.name)
	tempLayerTree = #(docElement.getAttribute "layer")
	
	for i = 0 to (docElement.childNodes.count - 1) do
	(
		tmp = (recurseXML docElement.childNodes.itemOf[i])
		append tempPTree tmp[1]
		append tempLayerTree tmp[2]
	)
	return #(tempPTree, tempLayerTree)
)

global parentTreeXML = #()
global layerTreeXML = #()
fn readXML =
(
	xmlDoc = dotNetObject "System.Xml.XmlDocument"
	xmlDoc.load ("C:\\temp\\test.xml") 
	
	docElement = XmlDoc.documentElement
	
	-- if we don't do this here we end up with the root node on parentTreeXML
	-- which can be worked out in anoter way also, of course.
	for i = 0 to (docElement.childNodes.count - 1) do
	(
		tmp = (recurseXML docElement.childNodes.itemOf[i])
		append parentTreeXML tmp[1]
		append layerTreeXML tmp[2]
	)
)

Remember, besides dotnet help website, showproperties and showmethods are your best friends. You can also copy this code and paste on maxscript editor or jedit or somewhere with a syntax highlighter to better read it.

1 Comment :, , , , , , , more...

reading/writing xml with python and maxscript

by on Mar.09, 2011, under pipeline

If you know where to find the information it becomes quite straight forward if you are not doing anything complex. With python you can use xml.dom.minidom, while with maxscript you can use either .NET Object “System.Xml.XmlDocument” or Class “System.Xml.XmlReader“.

With python you create a root element and append child elements from there. A very good sample can be found on: http://www.postneo.com/projects/pyxml/ To read you have getElementsByTagName on both python and dotnet XmlDocument. XmlReader might be a bit faster, but you have to parse elements by yourself.

There’s an issue with minidom’s toprettyprintxml, it adds whitespace and tabs between tags, around textNodes. And they are obviously read afterwards. A couple of different solutions are discussed on Ron Rothman’s blog, the easiest one using xml.dom.ext.PrettyPrint.

Leave a Comment :, , , , more...

exporting mirrored animations from Motionbuilder

by on Mar.06, 2011, under MotionBuilder, pipeline

It’s quite easy. Basically you need to:

  1. plot to the skeleton;
  2. save one animation;
  3. turn on Mirror Animation for the character;
  4. plot back to the control rig, which then mirrors the animation;
  5. rotate the Character Reference model 180 degrees;
  6. plot back to the skeleton;
  7. save mirrored animation.

The tricky part is just number 5 where you need to do some matrix rotation. Python for this would be something like:

from pyfbsdk import *
app = FBApplication()
char = app.CurrentCharacter
savePath = r"C:\"
filename = app.FBXFileName
skeleton = FBCharacterPlotWhere.kFBCharacterPlotOnSkeleton
ctrlrig = FBCharacterPlotWhere.kFBCharacterPlotOnControlRig

# plot to skeleton, see bellow
plotAnim(char, skeleton)

# save left animation
sOptions = FBFbxOptions(False) # false = save options
sOptions.SaveCharacter = True
sOptions.SaveControlSet = False
sOptions.SaveCharacterExtension = False
sOptions.ShowFileDialog = False
sOptions.ShowOptionsDialog = False
app.SaveCharacterRigAndAnimation(savePath + "\\" + filename + "_L", char, sOptions)

# activate mirror and plot
char.MirrorMode = True
plotAnim(char, ctrlrig)

# get reference model
refModel = FBFindModelByName("Character_Ctrl:Reference")

# rotating 180, the tricky part
# http://www.j3d.org/matrix_faq/matrfaq_latest.html#Q28
rotateY180 = FBMatrix()
rotateY180[0] = math.cos((180*0.017453292519943295769236907684886))
rotateY180[2] = math.sin((180*0.017453292519943295769236907684886))
rotateY180[8] = -math.sin((180*0.017453292519943295769236907684886))
rotateY180[10] = math.cos((180*0.017453292519943295769236907684886))

refMT = FBMatrix()
refModel.GetMatrix(refMT)

refModel.SetMatrix( MatrixMult(rotateY180, refMT) )
scene.Evaluate()

# plot back to skeleton
plotAnim(char, skeleton)

# save again
app.SaveCharacterRigAndAnimation(savePath + "\\" + filename + "_R", char, sOptions)

The plot and multiplication functions are:

# This is from Neil3d: http://neill3d.com/mobi-skript-raschet-additivnoj-animacii?langswitch_lang=en
def MatrixMult(Ma, Mb):
    res = FBMatrix()

    for i in range(0,4):
        for j in range(0,4):
            sum=0
            for k in range(0,4):
                sum += Ma[i*4+k] * Mb[k*4+j]

            res[i*4+j] = sum
    return res

def plotAnim(char, where):
    if char.GetCharacterize:
        switchOn = char.SetCharacterizeOn(True)

    plotoBla = FBPlotOptions()
    plotoBla.ConstantKeyReducerKeepOneKey = True
    plotoBla.PlotAllTakes = True
    plotoBla.PlotOnFrame = True
    plotoBla.PlotPeriod = FBTime( 0, 0, 0, 1 )
    #plotoBla.PlotTranslationOnRootOnly = True
    plotoBla.PreciseTimeDiscontinuities = True
    #plotoBla.RotationFilterToApply = FBRotationFilter.kFBRotationFilterGimbleKiller
    plotoBla.UseConstantKeyReducer = False
    plotoBla.ConstantKeyReducerKeepOneKey  = True

    if (not char.PlotAnimation(where, plotoBla)):
        FBMessageBox( "Something went wrong", "Plot animation returned false, cannot continue", "OK", None, None )
        return False

    return char    

If you are exporting the character to a game, pay attention to the root node rotation. If you used it in the characterization, chances are it might also be rotated, and you might not want that to happen. You can also establish a convention on the name of the files, and easily detect if the animation currently open ends with _L or _R, and adapt the filename correctly.

FBApplication().SaveCharacterRigAndAnimation() is the equivalent of the Save Character Animation on the Character Controls window, which saves only the animation, without mesh. I prefer to use this when exporting only the animation from Motionbuilder to some other software, since it’s cleaner, faster and file sizes are smaller, but you could use any other function also.

Leave a Comment :, , , , , , more...

Looking for something?

Use the form below to search the site:

Still not finding what you're looking for? Drop a comment on a post or contact us so we can take care of it!