main.py

import argparse
import graph_util as gu
import graph
import util
from evaluation import Validation
import numpy as np
import os


def main(args):
    cur_dir = os.getcwd()
    adjacency_folder = os.path.join(cur_dir, args.adj_folder)
    training_genes_file = os.path.join(cur_dir, args.train_genes_file)
    training_labels_file = os.path.join(cur_dir, args.train_labels_file)
    all_genes_file = os.path.join(cur_dir, args.all_genes_file)
    list_D = args.list_D
    list_C = args.list_C
    list_d = args.list_d
    list_r = args.list_r

    training_genes = util.load_list_from_file(training_genes_file)
    training_labels = [int(l) for l in util.load_list_from_file(training_labels_file)]
    all_genes = util.load_list_from_file(all_genes_file)

    # Creating list of graphs
    print("Unifying graphs...")
    if args.use_vec:
        graphs = gu.create_graphs(adjacency_folder_path=adjacency_folder, list_attr_path=args.node_vecs_file)
    else:
        graphs = gu.create_graphs(adjacency_folder_path=adjacency_folder)

    # Computing kernel matrices
    print("Computing graph kernels...")
    kernel_matrices = []
    for D in list_D:
        for C in list_C:
            g_union = gu.union_graphs(graphs=graphs, deg_threshold=D, cli_threshold=C)
            for d in list_d:
                for r in list_r:
                    vec = graph.CDNK_Vectorizer(d=d, r=r, L=len(graphs), n_nodes=len(graphs[0].nodes()),
                                                discrete=not args.use_vec)
                    kernel_matrices.append(vec.cdnk(g=g_union))

    print("Evaluating model...")
    if args.use_lou:
        val = Validation(kernels=kernel_matrices, all_genes=all_genes, training_genes=training_genes,
                         training_labels=training_labels)
        print('============')
        print('Performances')
        auc = val.validate_leave_one_out()
        print(auc)
    else:
        val = Validation(kernels=kernel_matrices, all_genes=all_genes, training_genes=training_genes, training_labels=training_labels, n_folds=5)
        print('============')
        print('Performances')
        aucs = val.validate_kfolds()
        for auc in aucs:
            print(auc)
        print('-----------')
        print(np.mean(aucs))


if __name__ == '__main__':
    parser = argparse.ArgumentParser(description='GCN')
    parser.add_argument("--adj_folder", type=str, default="", help="path to adjacency folder")
    parser.add_argument("--train_genes_file", type=str, default="", help="path to the file including train gene list")
    parser.add_argument("--train_labels_file", type=str, default="", help="path to the file including train label list")
    parser.add_argument("--node_vecs_file", type=str, default="",
                        help="path to the file including node feature vectors")
    parser.add_argument("--all_genes_file", type=str, default="", help="path to the file including all gene list")
    parser.add_argument("--list_D", type=int, nargs='+', help="List of degree thresholds")
    parser.add_argument("--list_C", type=int, nargs='+', help="List of clique thresholds")
    parser.add_argument("--list_d", type=int, nargs='+', help="List of maximum distance parameter in CDNK")
    parser.add_argument("--list_r", type=int, nargs='+', help="List of maximum radiua parameter in CDNK")
    parser.add_argument('--use_vec', dest='use_vec', action='store_true', default=False, help="Using node features")
    parser.add_argument('--use_lou', dest='use_lou', action='store_true', default=False, help="Using leave one out CV")
    args = parser.parse_args()

    print(args)
    main(args)