Labeling By Mayority

label_files.py

import shutil
import os
import pandas as pd
import numpy as np

def type_label(x):
    labels = np.zeros(4)
    for i in x:
        labels[i] += 1
    return np.argmax(labels)

cluster_labels = {}
for filename in os.listdir('.'):
    if filename.startswith('test_sequential_clustering') and filename.endswith('.csv'):
        df = pd.read_csv(filename, sep=',')
        for _, row in df.iterrows():
            if row['cluster'] not in cluster_labels:
                cluster_labels[row['cluster']] = []
            cluster_labels[row['cluster']].append(row['type'])

for c in cluster_labels.keys():
    cluster_labels[c] = type_label(cluster_labels[c])
    cname = "test_sequential_seq_cluster_{}.wav".format(c)
    if cluster_labels[c] == 2:
        outname = "clicks_test_sequential_seq_cluster_{}.wav".format(c)
    else:
        outname = "shape_test_sequential_seq_cluster_{}.wav".format(c)
    if os.path.exists(cname):
        shutil.copyfile(cname, outname)