smm_eval.py

import sys
from pathlib import Path

sys.path.insert(0, "../")

import scipy.io

from leitmotifs.lama import *

from leitmotifs.competitors import *

import matplotlib
matplotlib.rcParams['pdf.fonttype'] = 42
matplotlib.rcParams['ps.fonttype'] = 42
import warnings

warnings.simplefilter("ignore")

dataset_names = [
    'physio',
    'Boxing',
    'Swordplay',
    'Basketball',
    'Charleston - Side By Side Female',
    'crypto',
    'birds',
    "What I've Done - Linkin Park",
    'Numb - Linkin Park',
    'Vanilla Ice - Ice Ice Baby',
    'Queen David Bowie - Under Pressure',
    'The Rolling Stones - Paint It, Black',
    'Star Wars - The Imperial March',
    'Lord of the Rings Symphony - The Shire']

scores = {}
# Load the .mat file

root_directory = "results/smm_benchmark/results/"
for dirpath, dirnames, filenames in os.walk(root_directory):
    for dirname in dirnames:
        path = os.path.join(dirpath, dirname)
        print ("-------------------------------")
        print(path)
        for i in range(1, 15):
            file_path = Path(os.path.join(path, "config.txt"))
            # print(file_path.read_text())

            series_path = "../datasets/benchmark/" + str(i) + ".csv"
            series = pd.read_csv(series_path, header=None).T
            df_gt = read_ground_truth(series_path)
            file = path + "/Motif_" + str(i) + "_DepO_2_DepT_2.mat"

            # only for plotting
            #ml = LAMA(
            #    dataset_names[i-1],
            #    series,
            #    dimension_labels=df.index,
            #)
            #ml.plot_dataset()

            # some dataset found no motifs
            if not os.path.exists(file):
                # print(f"The file {file} does not exist.")
                scores[i] = [dataset_names[i - 1], 0.0, 0.0]
                # print("\t", scores[i])
                continue

            # print("\t"+dataset_names[i - 1])

            mat_file = scipy.io.loadmat(file, struct_as_record=False, squeeze_me=True)
            motif_bags = mat_file["MotifBag"]

            if not isinstance(motif_bags, np.ndarray):
                motif_bags = [motif_bags]

            best_f_score = 0.0
            best_motif_set = []
            best_dims = []
            best_length = 0
            precision, recall = 0, 0

            for motif_bag in motif_bags:
                if motif_bag:
                    startIdx = motif_bag.startIdx

                    motif_set = startIdx
                    dims = motif_bag.depd[0] - 1  # matlab uses 1-indexing but python 0-indexing
                    if not isinstance(dims, np.ndarray):
                        dims = [dims]

                    length = motif_bag.Tscope[0]
                    if length == 0:
                        length = 1

                    precision, recall = compute_precision_recall(
                        np.sort(motif_set), df_gt.values[0, 0], length)

                    f_score = 2 * (precision * recall) / (precision + recall + 1e-8)
                    if f_score > best_f_score:
                        best_f_score = f_score
                        best_motif_set = motif_set
                        best_length = length
                        best_dims = dims
                        best_precision = precision
                        best_recall = recall

            precision, recall = compute_precision_recall(
                np.sort(best_motif_set), df_gt.values[0, 0], best_length)

            scores[i] = [dataset_names[i - 1], precision, recall]
            #print("\tMotifs:\t", len(motif_bags))
            #print("\tDims:\t",best_dims)
            #print("\tLength:\t",best_length)
            # print("\t", scores[i])

        df = pd.DataFrame(scores).T
        df.columns = "Dataset Precision Recall".split()
        df["Precision"] = df["Precision"].astype(float)
        df["Recall"] = df["Recall"].astype(float)

        # print(df)
        print(df.set_index("Dataset").describe().loc["mean"])

        df.to_csv('csv/smm_'+dirname+'_results.csv', index=False)
        # break