add xml to graphml converter

Author: Vanessa Arndorfer

Simulator/Edges/Neuro/AllSpikingSynapses.h

@@ -203,7 +203,7 @@ class AllSpikingSynapses : public AllNeuroEdges {
    ///
    ///  @param  allEdgesDevice  GPU address of the AllSpikingSynapsesDeviceProperties struct
    ///                             on device memory.
-   virtual void copyDeviceEdgeWeightsToHost(void *allEdgesDevice) ;
+   virtual void copyDeviceEdgeWeightsToHost(void *allEdgesDevice);
 
 protected:
    ///  Allocate GPU memories to store all synapses' states,

Simulator/Edges/Neuro/AllSpikingSynapses_d.cpp

@@ -291,8 +291,8 @@ void AllSpikingSynapses::copyDeviceEdgeWeightsToHost(void *allEdgesDevice)
    // std::cout << "W_.data(): " << W_.data() << std::endl;
    // std::cout << "allEdgesDeviceProps.W_: " << allEdgesDeviceProps.W_ << std::endl;
 
-   HANDLE_ERROR(cudaMemcpy(W_.data(), allEdgesDeviceProps.W_, 
-                  maxTotalSynapses * sizeof(BGFLOAT), cudaMemcpyDeviceToHost));
+   HANDLE_ERROR(cudaMemcpy(W_.data(), allEdgesDeviceProps.W_,
+                           maxTotalSynapses * sizeof(BGFLOAT), cudaMemcpyDeviceToHost));
 }
 
 ///  Get summationCoord and in_use in AllEdges struct on device memory.

Tools/XMLToGraphML/README.md

@@ -0,0 +1,124 @@
+'''
+GETGRAPHEDGES Generate a graphml file from the weights matrix simulation output
+
+This function reads the Graphitti weight matrix output from the
+growth simulation and reformats it to graphml for input into the
+STDP simulation.
+
+Input:
+weights_file - Weight matrix output from Graphitti. The entire path can be used; for example
+               '/CSSDIV/research/biocomputing/data/2025/tR_1.0--fE_0.90_10000/weights-epoch-1.xml'
+graphml_file - graphml file to add the edges. This is typically the same file 
+               used as input to the Graphitti simulation.
+
+Output:
+  - weight_graph.graphml
+
+Author: Vanessa Arndorfer ([email protected])
+Last updated: 07/01/2025
+'''
+
+import numpy as np
+import os
+import pandas as pd
+import sys
+import xml.etree.ElementTree as ET 
+
+
+def xmlToNumpy(node, rows, cols):
+    print("Converting xml to matrix for: " + node.tag)
+
+    m = np.zeros(shape=(rows, cols))
+    r = 0
+    c = 0
+    for child in node:
+        m[r][c] = float(child.text)
+        c += 1
+
+        # new row
+        if c == cols:
+            r += 1
+            c = 0
+
+    return m
+
+
+def getWeightMatrix(file_name, src_root, weights_root):
+    print("Building weights matrix for: " + file_name)
+
+    tree = ET.parse(file_name)
+    root = tree.getroot()
+
+    idNum = 0
+    srcIdx = root.find(src_root)
+    weights = root.find(weights_root)
+
+    rows = 10000
+    cols = 200
+
+    srcIdx_np = xmlToNumpy(srcIdx, rows, cols)
+    weights_np = xmlToNumpy(weights, rows, cols)
+
+    print("Converting matrices into square format...")
+    edge_weights = np.zeros(shape=(rows, rows))
+
+    weight_count = 0
+
+    for r in range(weights_np.shape[0]):
+        for c in range(weights_np.shape[1]):
+            if weights_np[r][c] != 0:
+                weight_count += 1
+                src = int(srcIdx_np[r][c])
+                edge_weights[src][r] = weights_np[r][c]
+
+    print("Total weighted edges: " + str(weight_count))
+
+    return edge_weights
+
+
+def getEdgeGraphML(edge_weights, graphml_file, output_dir):
+    print("Adding edges to the graph file: " + graphml_file)
+
+    # register xml namespaces
+    ET.register_namespace('', "http://graphml.graphdrawing.org/xmlns")
+    ET.register_namespace('xsi', "http://www.w3.org/2001/XMLSchema-instance")
+    ET.register_namespace('xsi:schemaLocation', "http://graphml.graphdrawing.org/xmlns http://graphml.graphdrawing.org/xmlns/1.0/graphml.xsd")
+
+    tree = ET.parse(graphml_file)
+    root = tree.getroot()
+
+    # define weight attribute for edges
+    weight_key = ET.Element("key", attrib={"id":"weight", "for":"edge", "attr.name":"weight", "attr.type":"double"})
+    root.append(weight_key)
+
+    idNum = 0
+    graph = root.find('./{http://graphml.graphdrawing.org/xmlns}graph')
+
+    for r in range(edge_weights.shape[0]):
+        for c in range(edge_weights.shape[1]):
+            if edge_weights[r][c] != 0:
+                edgeId = "e" + str(idNum)
+                edge = ET.Element("edge", attrib={"id":edgeId, "source":str(r), "target":str(c)})
+                data = ET.SubElement(edge, "data", attrib={"key": "weight"})
+                data.text = str(edge_weights[r][c])
+                graph.append(edge)
+                idNum += 1
+
+    print(str(idNum) + " edges created")
+    ET.indent(tree, space="\t", level=0)
+    tree.write(output_dir + "/weight_graph.graphml", encoding="utf-8", xml_declaration=True)
+
+
+if __name__ == "__main__": 
+    # execution format: python3.9 ./getGraphEdges.py weights_file graphml_file
+    # example: python3.9 ./getGraphEdges.py /CSSDIV/research/biocomputing/data/2025/tR_1.0--fE_0.90_10000/weights-epoch-25.xml configfiles/graphs/fE_0.90_10000.graphml
+    weights_file = sys.argv[1]
+    graphml_file = sys.argv[2]
+
+    ht = os.path.split(weights_file)
+    output_dir = ht[0]
+    print("Output dir: " + output_dir)
+
+    df = getWeightMatrix(weights_file, "SourceVertexIndex", "WeightMatrix")
+    getEdgeGraphML(df, graphml_file, output_dir)
+