topoml: Segmentation of MSC from Image, conversion to graph data structure graph sage learning

mscnn_seg.py: 
    - callable script e.g. python mscnn_seg.py
    - use of msc_segmentation class to obtain msc or geometric msc from images
	      converted to graph data structure
	- compute geometric msc from images for graph data structure and learning
	Uses:  topoml.mscnn_segmentation import mscnn_segmentation     
	        which uses: topoml.topology.read_msc import MSC and topoml.topology.geometric_msc import GeoMSC       

test_gnn: compute msc of passed image into graph structure or take msc_segementation class output.
	  run graph neural network

mscnn_segmentation: compute msc segmentation to use by graph neural nets
		    encodes in graph data structure needed and saves example to file
		            compute geometric morse smale segmentation
		            
topoml.arcneurontracer: encodes msc in graph structure of arcs nodes pixels
                        inherets from neuron tracer which constructs the msc or geomsc using topoml.topology.utils.build_msc (or build_geometric_msc)

topoml.neurontracer: parent class to constructing msc. uses topoml.topology.utils to construct msc

graph_learning.classify_feats_graphs_v2 class graph_learners: employs various graph learning strategies
    (CAMLP, Deep Walk, Confidence Label Prediction, Random Forests, locally and globally consistency) using msc saved as csv

topoml.ui.ArcSelector: compute or use computed MSC to create graph
    -  msc_selector.draw_binary_segmentation  draw MSC graph as binary image
    
 topoml.topology.geometric_msc import GeoMSC 
    - read from file

GradIntegrator: extract2dridgegraph.cxx: compute geometric msc of 2d input. write to files filename"_vertices.txt" contents: vertex_id , x , y filename"_edges.txt" contents: edge_id, adjacent vertex 1 id, vertex 2 id, x_1, y_1, ... , x_i , y_i

Unsupervised_UNet: compute msc of input image and use segmentation for training UNet.