kaist-dmlab / fl-sim Goto Github PK

@YoojuShin

when numNodes=50 and numEdges=10
fl_data.groupByNode is ok
fl_data.groupByEdge(trainData_by1Nid, 't', 'q', NUM_NODES, NUM_EDGES) gets the following error

50 10 5.00 5.70 6.20 <class 'numpy.ndarray'> <class 'numpy.float32'> <class 'numpy.int32'>
50 10 5.00 6.20 15.60 <class 'numpy.ndarray'> <class 'numpy.float32'> <class 'numpy.int32'>

IndexError Traceback (most recent call last)
in
31 assert_groupByEdge(trainData_byNid, train_z, NUM_NODES, NUM_EDGES, NUM_NODES_PER_EDGE, numTotalClasses/10, numTotalClasses/4)
32
---> 33 (trainData_byNid, train_z) = fl_data.groupByEdge(trainData_by1Nid, 't', 'h', NUM_NODES, NUM_EDGES)
34 assert_groupByEdge(trainData_byNid, train_z, NUM_NODES, NUM_EDGES, NUM_NODES_PER_EDGE, numTotalClasses/10, numTotalClasses/2)
35

~/FL-Sim/fl_data.py in groupByEdge(data_by1Nid, nodeType, edgeType, numNodes, numEdges)
296 num_allocated_nodes = 0
297 while(len(class_left) != 0): # check whether all class are sampled at least once in the nodes for an edge
--> 298 if NUM_CLASS_NODE_SAMPLED[num_allocated_nodes] > len(class_left):
299 class_list = list(class_left)
300 class_left_complement_list = list(set(edge_class_list).difference(class_left))

IndexError: list index out of range