Version 0.1. Removed dependencies from binvox, trimesh, open3d. Auto-…

…install dependencies, mirroring, post-generation utilities

.gitignore

@@ -1 +1,2 @@
-__pycache__
+__pycache__
+_additional_modules

.gitmodules

@@ -1,3 +1,3 @@
 [submodule "RigNet"]
 	path = RigNet
-	url = https://github.com/zhan-xu/RigNet.git
+	url = https://github.com/pKrime/RigNet.git

README.md

@@ -1,8 +1,10 @@
-# bRigNet
+bRigNet
 ---------
 Neural Rigging for [blender](https://www.blender.org/ "Blender Home Page") using [RigNet](https://zhan-xu.github.io/rig-net/ "RigNet Home Page")
 
-Blender is the open source 3D application from the Blender Foundation. RigNet has been presented in the following papers
+Blender is the open source 3D application from the Blender Foundation. RigNet is the Machine Learning prediction
+for articulated characters. It has a dual license, GPL3 for open source projects, commercial otherwise.
+It was presented in the following papers
 
 ``` 
   @InProceedings{AnimSkelVolNet,
@@ -23,31 +25,147 @@ Blender is the open source 3D application from the Blender Foundation. RigNet ha
   }
 ```
 
-**Warning:** this addon is experimental, the install procedure is clumsy and the code could be better.
-Everything about it could change, hopefully in good.
 
 ## Setup
-Follow [RigNet instruction](https://github.com/zhan-xu/RigNet "RigNet repository")
-and create a Python environment with all required packages
 
-For compatibility reasons, [Pillow](https://python-pillow.org/) must be installed in blender.
-You can install pip following [this guide](http://www.codeplastic.com/2019/03/12/how-to-install-python-modules-in-blender/ "pip in blender"),
-then
+bRigNet requires SciPy, PyTorch and torch-geometric, along with torch-scatter and torch-sparse.
+
+## Installation
+
+Download the Neural Rigging add-on as a .zip file and install it from the blender addons window,
+or copy the code to the blender scripts path
+
+### Install dependencies via "Install" button
+
+At present, the CUDA toolkit from nVidia is required, it can be found at the
+[manifacturer website](https://developer.nvidia.com)
+
+A dependency installer is available in the preferences.
+
+* Install CUDA. At present prebuilt packages support versions 10.1, 10.2, 11.1
+* In the addon preferences, make sure that the Cuda version is detected correctly.
+* Hit the "Install" button. It can take time!
+
+Alternatively, Environment managers, like conda or virtualenv can be used to ease the install. 
+
+### Install dependencies using *conda*
+
+Anaconda is a data science platform from Anaconda Inc., it can be downloaded from the
+[company website](https://www.anaconda.com/).
+
+A lightweight version called [Miniconda](https://docs.conda.io/en/latest/miniconda.html) is available.
+Both versions include the package manager 'conda' used in the following steps.
+
+- Open a Miniconda or Anaconda prompt
+- Create a Conda Environment and activate it
 
 ```
-[BLENDER_DIR]\[BLENDER_VER]\python\bin\python -m pip install --upgrade Pillow
+conda create -n brignet python=3.7
+conda activate brignet_deps
 ```
 
-## Installation
+- Install PyTorch. If CUDA is installed, the CUDA version can be queried in a command prompt. For example
 
-Download bRigNet as a .zip file and install it from the blender addons window,
-or copy the code to the blender scripts path
+```
+nvcc --version
+```
+```
+nvcc: NVIDIA (R) Cuda compiler driver
+Copyright (c) 2005-2019 NVIDIA Corporation
+Built on Wed_Oct_23_19:32:27_Pacific_Daylight_Time_2019
+Cuda compilation tools, release 10.2, V10.2.89
+```
+
+In this case PyTorch can be installed in the command prompt via
+
+```
+conda install pytorch==1.8.1 cudatoolkit=10.2 -c pytorch
+```
+
+More complete information on the PyTorch command line can be found at the [PyTorch website](https://pytorch.org/)
+The install command on non-cuda devices is
+
+```
+conda install pytorch==1.8.1 cpuonly -c pytorch
+```
+
+- Install torch utilities. The syntax follows the pattern
+
+```
+pip install [package-name] -f https://pytorch-geometric.com/whl/torch-[version]+cu[cuda-version].html
+```
+
+```
+pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.8.1+cu102.html
+pip install torch-sparse -f https://pytorch-geometric.com/whl/torch-1.8.1+cu102.html
+pip install torch-cluster -f https://pytorch-geometric.com/whl/torch-1.8.1+cu102.html
+pip install torch-spline-conv -f https://pytorch-geometric.com/whl/torch-1.8.1+cu102.html
+pip install torch-geometric
+```
+
+Alternatively, pip can try and build the libraries. Even if part of torch-sparse fails without a proper environment,
+the rilevant modules are usually built
+
+```
+pip install torch-scatter
+pip install torch-sparse
+pip install torch-geometric
+```
+
+The directory of each environment can be obtained via
+
+```
+conda info --envs
+```
+
+The environment directory can be set in the "Additional Modules" setting of the bRigNet preferences
+
+### Install dependencies using virtualenv
+
+virtualenv can be used to create a Python environment with the required packages.
+First, python 3.7 must be installed on the system. It can be found at https://www.python.org/downloads/
+
+Make sure that **Add Python 3.7 to PATH** is checked in the setup options.    
+
+Usually, python comes with its package manager installed (pip). Please, refer
+to the [pip documentation](https://pypi.org/project/pip/) if pip is not present in your system.
+
+Next step is to install virtualenv. Open a command prompt and reach a folder where python packages will be kept
+please execute.
+
+```
+pip install virtualenv
+```
+
+then create the virtual environment and activate it
+
+```
+virtualenv brignet_deps
+cd brignet_deps
+Scripts\activate
+```
+
+now we can install the torch library. At present, version 1.8.1 is provided.
+torch-geometric provides prebuilt packages for CUDA 10.1, 10.2, 11.1
+
+CUDA 10.2 is used in this example:
+
+```
+pip install torch==1.8.1+cu102 -f https://download.pytorch.org/whl/torch_stable.html
+pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.8.1+cu102.html
+pip install torch-sparse -f https://pytorch-geometric.com/whl/torch-1.8.1+cu102.html
+pip install torch-cluster -f https://pytorch-geometric.com/whl/torch-1.8.1+cu102.html
+pip install torch-spline-conv -f https://pytorch-geometric.com/whl/torch-1.8.1+cu102.html
+pip install torch-geometric
+```
+
+the virtual environment directory can be set as the "Additional modules path" in the brignet preferences
 
 ## Usage 
 Enable *bRigNet* in the blender addons, the preferences will show up.
 Set the Modules path properties to the RigNet environment from the previous step
 
-RigNet requires a trained model. They have made theirs available at [this address](https://umass.box.com/s/l7dxfayrubf5qzxcyg7can715xnislwm)
+RigNet requires a trained model. They have made theirs available at [this address](https://umass-my.sharepoint.com/:u:/g/personal/zhanxu_umass_edu/EYKLCvYTWFJArehlo3-H2SgBABnY08B4k5Q14K7H1Hh0VA)
 The checkpoint folder can be copied to the RigNet subfolder.
 A different location can be set in the addon preferences.
 
@@ -63,6 +181,19 @@ on which to transfer the final weights
 Rigs generated using RigNet from the command line can be loaded via the **Load Skeleton** panel.
 Please select the *.obj and *.txt file and press the button **Load Rignet character**
 
+## Training
+
+The blender addon doesn't cover training yet. If you want to train your own model, please follow the instructions
+from the [RigNet project](https://github.com/zhan-xu/RigNet#training).
+
+## Disclaimer
+
+This blender implementation of RigNet and the author of this add-on are NOT associated with the University of
+Massachusetts Amherst.
+
+This add-on has received a research grant from the Blender Foundation.   
+
 ## License
 
-bRigNet is released under the GNU General Public License version 3. RigNet is licensed under the General Public License Version 3 (GPLv3), or under a Commercial License.
+This addon is released under the GNU General Public License version 3 (GPLv3).
+The RigNet subfolder is licensed under the General Public License Version 3 (GPLv3), or under a Commercial License.

__init__.py

@@ -17,58 +17,62 @@
 
 
 bl_info = {
-    "name": "bRigNet",
-    "version": (0, 0, 1),
+    "name": "Neural Rigging",
+    "version": (0, 1, 0),
     "author": "Paolo Acampora",
     "blender": (2, 90, 0),
     "description": "Armature and Weights prediction using RigNet",
-    "location": "Armature properties",
+    "location": "3D Viewport",
     "doc_url": "https://github.com/pKrime/brignet",
     "category": "Rigging",
 }
 
 import bpy
 
 from . import brignet, preferences, loadskeleton
+from . import postgen_utils
+from .ui import menus
 
 from importlib import reload
 try:
     reload(brignet)
     reload(preferences)
     reload(loadskeleton)
+    reload(postgen_utils)
 except NameError:
     pass
 
 from .brignet import BrignetPanel, BrigNetPredict, BrignetRemesh, BrignetCollection
-from .preferences import BrignetPrefs
+from .preferences import BrignetPrefs, BrignetEnvironment
 from .loadskeleton import LoadRignetSkeleton, LoadSkeletonPanel
-
+from .postgen_utils import NamiFy, ExtractMetarig, SpineFix, MergeBones
 
 
 # REGISTER #
 
 def register():
-    from importlib import reload
-    try:
-        reload(brignet)
-        reload(preferences)
-        reload(loadskeleton)
-    except NameError:
-        pass
-
     brignet.register_properties()
+    bpy.utils.register_class(BrignetEnvironment)
     bpy.utils.register_class(BrignetPrefs)
     bpy.utils.register_class(BrignetCollection)
     bpy.utils.register_class(BrignetRemesh)
     bpy.utils.register_class(BrigNetPredict)
+
+    bpy.utils.register_class(NamiFy)
+    bpy.utils.register_class(ExtractMetarig)
+    bpy.utils.register_class(SpineFix)
+    bpy.utils.register_class(MergeBones)
+
     bpy.utils.register_class(BrignetPanel)
     bpy.utils.register_class(LoadRignetSkeleton)
     bpy.utils.register_class(LoadSkeletonPanel)
 
-    if not BrignetPrefs.append_rignet():
-        print("RigNet not found, please set in bRigNet preferences")
-    if not BrignetPrefs.append_modules():
+    BrignetPrefs.check_cuda()
+    if not BrignetPrefs.add_module_paths():
         print("Modules path not found, please set in bRigNet preferences")
+    BrignetPrefs.check_modules()
+
+    bpy.types.VIEW3D_MT_pose_context_menu.append(menus.pose_context_options)
 
 
 def unregister():
@@ -79,11 +83,19 @@ def unregister():
         # if we have failed to load rignetconnect, we have no device to clear
         pass
 
+    bpy.types.VIEW3D_MT_pose_context_menu.remove(menus.pose_context_options)
+    BrignetPrefs.reset_module_paths()
+
     bpy.utils.unregister_class(BrignetPanel)
     bpy.utils.unregister_class(BrignetPrefs)
+    bpy.utils.unregister_class(BrignetEnvironment)
     bpy.utils.unregister_class(BrignetCollection)
     bpy.utils.unregister_class(BrignetRemesh)
     bpy.utils.unregister_class(BrigNetPredict)
+    bpy.utils.unregister_class(NamiFy)
+    bpy.utils.unregister_class(ExtractMetarig)
+    bpy.utils.unregister_class(SpineFix)
+    bpy.utils.unregister_class(MergeBones)
     bpy.utils.unregister_class(LoadSkeletonPanel)
     bpy.utils.unregister_class(LoadRignetSkeleton)
     brignet.unregister_properties()