Storing a part of the hierarchy

[pim]

I want to store a part of the hierarchy in a external (hf) file.
For example, an object with multiple levels of children, materials, etc.

I can put it in a new scene file and store that scene file, but perhaps there are other / better ways?

The ultimate goal is to store a part of the hierarchy in a tag plugin using node.write() and node.read().

[m_magalhaes]

hello,

you can use the function IsolateObjects than use SaveDocument or SaveProject

You can read it with LoadDocument to a temporary document inside memory.

Be careful, if you want to modify the active scene, you have to do it from the main thread otherwise you scene could be in an uncertain state leading to crash.
you can read information on how to deal with that case here

Cheers,
Manuel

[pim]

Thanks, again I learned another function (IsolateObjects).
If I use LoadDocument to store it in memory, how can I store that memory in the scene file, when the user saves the screen file?
I am going to use the tagplugin, so I guess that is all right?

-Pim

[m_magalhaes]

hello,

what kind of functionality are you trying to introduce ? Using a tag doesn't look like the right place to do what you are trying to do.

Cheers,
Manuel

[pim]

I will send you a pm with more details.

-Pim

[m_magalhaes]

you can send us an email at sdk_support@maxon.net

[m_magalhaes]

hello,

One possibility that i forgot to mention is MemoryFileStructure Manual

And ReadHyperFile or WriteHyperFile

You also have with a HyperFile the function WriteFilename

Hope this will help.

Cheers,
Manuel.

[pim]

I want to store the node to be hidden in the scene file, so I guess hf wil not work for me.
In the other thread I learned to use IsolateObjects and that works now.

So now I have a new doc in memory with the node I wanted to store.
I tried to put it in a subcontainer, but that did not work.
When reading back I just got None.

import c4d

PLUGIN_ID = 1234567
MyUniqueId = 456789

def main():
    #retrieves the document baseContainer
    docBC = doc.GetDataInstance()

    sel = doc.GetSelection()    # !!! only select the parent, not the children
    newDoc = c4d.documents.IsolateObjects(doc, sel)

    #create a sub-BaseContainer
    subBc = c4d.BaseContainer()
    subBc[1000] = "hello"
    subBc[2000] = newDoc

    # Add the container to the "main" Container
    docBC.SetContainer(MyUniqueId, subBc)

    # Updates the document container
    doc.SetData(docBC)

    # Print the values stored in our container.
    for cid, value in doc.GetDataInstance().GetContainer(MyUniqueId):
        print cid, value

    # Print Hello World
    print doc[MyUniqueId][1000], doc[MyUniqueId][2000]


if __name__=='__main__':
   main()

Output:

1000 hello
2000 <c4d.documents.BaseDocument object called '' with ID 110059 at 0x00000173601BD370>
hello <c4d.documents.BaseDocument object called '' with ID 110059 at 0x00000173601BD330>

Reading back the container:

import c4d

PLUGIN_ID = 1234567
MyUniqueId = 456789

def main():
    #retrieves the document baseContainer
    docBC = doc.GetDataInstance()


    # Print the values stored in our container.
    for cid, value in doc.GetDataInstance().GetContainer(MyUniqueId):
        print cid, value

    # Print Hello World
    print doc[MyUniqueId][1000], doc[MyUniqueId][2000]


if __name__=='__main__':
   main()

Output:

1000 hello
2000 None
hello None

[zipit]

Hi,

the following line

subBc[2000] = newDoc

will store a BaseLink (a reference, a pointer) to newDoc in your container, not the object itself. While Python's important syntax for bracket operators (__setitem__, __getitem__) usually is really convenient, in the case of BaseContainers it can obscure what you are actually doing. The verbose form of your call would be:

subBc.SetLink(2000, newDoc)

Which works because BaseDocument is derived from BaseList2D and therefor also from C4DAtom (the type BaseContainer.SetLink() does expect as its second argument). However, when this link is not properly serialized on saving the document (which is impossible with your floating document), this link would be a null pointer when the document is loaded again. You cannot store a C4DAtom in a BaseContainer. At least I am not aware of a way to do so. You can only store references, which need to be kept alive in order to resolve.

On a side note: This whole hijacking of the documents data container should not work in the first place, i.e. Cinema will just clear it out the container at your ID. I am a bit confused that you can actually read back your "hello" on a saved document.

Cheers
zipit

[zipit]

Hi,

One thing I forgot, which I already wanted to mention in your other threads, as it is a common technique to deal with similar problems in c4d. Are you aware that you can hide a GeListNode from the users view (so they won't show up in the Object Manager, the Material Manager, etc.) with their NBIT flags? This would be an easy way to store whatever you want to store. You won't believe all the hidden stuff that is floating around in your documents ;)

For details check GeListNode.ChangeNBit().

Cheers
zipit

[pim]

@zipit, I got the code from another thread: https://plugincafe.maxon.net/topic/11802/writing-data-to-the-c4d-file/19
And if you change

subBc[2000] = "world!"

to

subBc[2000] = op

it also works!

I tried

subBc.SetLink(2000, newDoc)

and indeed it does not work. It returns a None on reading back the subcontainer after storing the file.

[pim]

@zipit, yes I know that functionality (and use it often).
However, although it is hidden it still consumes cpu power.
We like to minimalize cpu power by converting a node using csto, hide the node and disabling generators and deformers in the node.

[zipit]

Hi,

subBc[2000] = "world!"

to

subBc[2000] = op

That is expected as op is also an C4DAtom.

it also works!

I tried

subBc.SetLink(2000, newDoc)

and indeed it does not work. It returns a None on reading back the subcontainer after storing the file.

The method SetLink() does not have a (boolean) return type. So this is to be expected too. I am not quite sure if you got my point, so here is a modified version of your first script - which you assume to be "working" as it printed back your document. I did modify it in such a way, that it does not work any more - I am really good at this specific task ;). I hope the code and the comments make more clear what I am trying to convey: That you are not storing your document, but a link to it.

import c4d

PLUGIN_ID = 1234567
MyUniqueId = 456789

def main():
    #retrieves the document baseContainer
    docBC = doc.GetDataInstance()

    sel = doc.GetSelection()    # !!! only select the parent, not the children
    newDoc = c4d.documents.IsolateObjects(doc, sel)

    #create a sub-BaseContainer
    subBc = c4d.BaseContainer()
    subBc[1000] = "hello"
    subBc[2000] = newDoc
    
    # Get the type of the ID we set
    print "ID 2000 is a link", subBc.GetType(2000) == c4d.DA_ALIASLINK
    
    # do the same thing in green. That SetLink() returns None
    # is expected as it has no return value (which is implictly
    # None in Python)
    subBc.SetLink(2000, newDoc)
    
    print "ID 2000 is a link", subBc.GetType(2000) == c4d.DA_ALIASLINK

    # Add the container to the "main" Container
    docBC.SetContainer(MyUniqueId, subBc)

    # Updates the document container
    doc.SetData(docBC)

    # Print the values stored in our container.
    for cid, value in doc.GetDataInstance().GetContainer(MyUniqueId):
        print cid, value
        
    print "\nDeleting newDoc...\n"
    # This part is a bit dodgy since Pythons garbage collector cannot
    # be trusted. But what I am trying to show is that you stored
    # a reference, not an object. For me it works here, but in some cases
    # an object can linger even after marking it for garbage collection.
    newDoc.Flush()
    del(newDoc)
    
    # Print the values stored in our container.
    for cid, value in doc.GetDataInstance().GetContainer(MyUniqueId):
        print cid, value


if __name__=='__main__':
   main()

this should print out something like:

ID 2000 is a link True
ID 2000 is a link True
1000 hello
2000 <c4d.documents.BaseDocument object called '' with ID 110059 at 0x00000135BC7F9030>

Deleting newDoc...

1000 hello
2000 None
>>> 

Cheers
zipit

[pim]

Thanks for the explanation.

Two questions:

1. the line <c4d.documents.BaseDocument object called '' with ID 110059 at 0x00000135BC7F9030> always show a strange name ''? Why is it not showing the name of the document?

2. Back to the main questions, what to do, to store the node in scene file.

[m_magalhaes]

hello,

1 - what kind of strange name ?

2 -

you can store the data itself in the tag and than with the function read and write store them in the Scenefine
I'm using a file on the disk on this example but that should work with the HyperFile link provide by the read and write functions.

Be aware that HyperFile.WriteMemory is storing byte sequences that will be platform dependent.

import c4d
from c4d import gui
# Welcome to the world of Python
import os


# Main function
def main():
    
    path =  c4d.storage.LoadDialog(c4d.FILESELECTTYPE_SCENES, flags=c4d.FILESELECT_DIRECTORY)
    path = os.path.join(path, "prout.txt")
    
    # using a MemoryFileStructure to store a document
    mfs = c4d.storage.MemoryFileStruct()
    mfs.SetMemoryWriteMode()
    
    newdoc = c4d.documents.IsolateObjects(doc , [op])
    # Save the document to the MemoryFileStructure
    c4d.documents.SaveDocument(newdoc, mfs, c4d.SAVEDOCUMENTFLAGS_NONE, c4d.FORMAT_C4DEXPORT)
    
    #Retrieve the data and store it somewhere, could be self.myData
    myData = mfs.GetData()
    
   
   
    #Save the data to a hyperfile
    myFile = c4d.storage.HyperFile()
    if not myFile.Open(0, path, c4d.FILEOPEN_WRITE, c4d.FILEDIALOG_NONE):
        raise RuntimeError("Failed to open the HyperFile in write mode.")
    
    # Store the size
    myFile.WriteInt32(myData[1])
    # Store the data itself
    if myFile.WriteMemory(myData[0]) is False:
        raise ValueError("can't write the file")
    myFile.Close()

    
    # Read the data from the HF
    if not myFile.Open(0, path, c4d.FILEOPEN_READ, c4d.FILEDIALOG_NONE):
        raise RuntimeError("Failed to open the HyperFile in read mode.")
    
    size = myFile.ReadInt32()
    data = myFile.ReadMemory()
    myFile.Close()
    
    #Set the MFS
    mfs2 = c4d.storage.MemoryFileStruct()
    mfs2.SetMemoryReadMode(data,size)
    
     
    c4d.documents.MergeDocument(doc, mfs2, c4d.SCENEFILTER_OBJECTS)
    c4d.EventAdd()

# Execute main()
if __name__=='__main__':
    main()

Once again, be aware of where you are modifying the scene (on main thread and not elsewhere)

by the way, are you going to use c++ or python at the end ? (just for the tags of this thread)

Cheers,
Manuel

[pim]

Great, thank you very much.
I am beginning to see the light.

And yes, my fault, apparently I indicated it as c++, but I am doing it in Python.

About your warning.
My plan was to do it all in a tag plugin.
There I can have the interface, do the isolate and the read/write in a hf.
Or is it better to do it in a Object plugin?

[zipit]

Hi,

@pim said in Storing a part of the hierarchy:

1. the line <c4d.documents.BaseDocument object called '' with ID 110059 at 0x00000135BC7F9030> always show a strange name ''? Why is it not showing the name of the document?

Not quite sure what you do mean by that:

1. The fact that the it prints the empty string for the document name is probably because you run the script on a unsaved document. This whole untitled_x.c4d stuff you see in c4d app is all smoke and daggers. An unsaved document will return the empty string for GetName().
2. If you mean how objects are printed out, this is because of the way it is convention to implement __repr__() for an object. I found that obsession with memory addresses also always quite bizarre, but, hey, everyone is doing it ;)

Cheers
zipit

[zipit]

@pim said in Storing a part of the hierarchy:

My plan was to do it all in a tag plugin.
There I can have the interface, do the isolate and the read/write in a hf.
Or is it better to do it in a Object plugin?

Hi, it does not really matter what kind of NodeData you use, what @m_magalhaes meant, was that you should be careful with the threaded context of methods like Execute() in TagData or GetVirtualObjects() in ObjectData. When you execute your code from a method that is executed from the main thread, you are fine (e.g. NodeData.Message()). You can always check there with c4d.threading.GeIsMainThread() if you are in the main thread to be extra sure.

Cheers
zipit

[m_magalhaes]

An ObjectData (generator) would be a bit better for that. But you are exploring possibilities, that's also the way to find new solutions / workflow.

About my warning it's just a reminder. You will have so sent message probably and use a MessageData to react to that message in order to do some action in the main thread.

Cheers,
Manuel

[pim]

Thanks for all the support.
I will use all the knowledge gained and start testing.
I am sure, I will be back with more questions.

-Pim