pyctbn-benchmarks/benchmark.py

#!/usr/bin/env python3

import glob
import os
import time
import argparse
from pathlib import Path

from pyctbn.legacy import JsonImporter
from pyctbn.legacy import SamplePath
from pyctbn.legacy import StructureConstraintBasedEstimator

def structure_constraint_based_estimation_example(network_file_path):
    Path("./data").mkdir(parents=True, exist_ok=True)
    # <read the json files in ./data path>
    read_files = glob.glob(os.path.join("./data/", "*.json"))
    # <initialize a JsonImporter object for the first file>
    importer = JsonImporter(
        file_path=network_file_path,
        samples_label='samples',
        structure_label='dyn.str',
        variables_label='variables',
        time_key='Time',
        variables_key='Name'
    )
    start_time = time.time()
    # <import the data at index 0 of the outer json array>
    importer.import_data(0)
    # construct a SamplePath Object passing a filled AbstractImporter object
    s1 = SamplePath(importer=importer)
    # build the trajectories
    s1.build_trajectories()
    # build the information about the net
    s1.build_structure()
    # construct a StructureEstimator object passing a correctly build SamplePath object
    # and the independence tests significance, if you have prior knowledge about
    # the net structure create a list of tuples
    # that contains them and pass it as known_edges parameter
    se1 = StructureConstraintBasedEstimator(
        sample_path=s1,
        exp_test_alfa=1e-6,
        chi_test_alfa=1e-4,
        known_edges=[],
        thumb_threshold=25
    )
    # call the algorithm to estimate the structure
    se1.estimate_structure()
    end_time = time.time()
    print("Elaspsed time: %d seconds" % (end_time - start_time))
    # obtain the adjacency matrix of the estimated structure
    #print(se1.adjacency_matrix())
    Path("./res").mkdir(parents=True, exist_ok=True)
    # save the estimated structure  to a json file
    # (remember to specify the path AND the .json extension)....
    se1.save_results("./res/results0.json")
    # ...or save it also in a graphical model fashion
    # (remember to specify the path AND the .png extension)
    #se1.save_plot_estimated_structure_graph("./res/result0.png")

parser = argparse.ArgumentParser()
parser.add_argument('filename', help="path of the network in json format")
args = parser.parse_args()

structure_constraint_based_estimation_example(args.filename)