debug.py

import os
import numpy as np
import matplotlib.pyplot as plt

PLOTS_PATH = "plots"

def display_images(images, num_frames, filename=None, title=None, nrows=4, show=False, in_db=True):
    
    if not os.path.exists(PLOTS_PATH):
        os.makedirs(PLOTS_PATH)
        
    filename = os.path.join(PLOTS_PATH, filename) if filename else None
    
    ncols = int(np.ceil(num_frames / nrows))
    
    fig, axes = plt.subplots(nrows=nrows, ncols=ncols, figsize=(16, 8))
    axes = np.atleast_1d(axes).flatten()

    
    axes = axes.flatten()
    if title:
        fig.suptitle(title, fontsize=16)
        
    for i in range(num_frames):
        if in_db:
            img = 20 * np.log10(np.abs(images[i]) + 1e-8)
        else:
            img = images[i]
        
        # masked_img = np.ma.masked_where(img == 0, img)

        im = axes[i].imshow(img, aspect='auto', origin='lower', cmap='viridis')
        axes[i].set_title(f"Image {i+1}")
        fig.colorbar(im, ax=axes[i])
    
    plt.tight_layout()
    
    n_empty_plots = nrows * ncols - num_frames
    for j in range(n_empty_plots):
        fig.delaxes(axes[-(j+1)])
        
    if filename:
        plt.savefig(filename)
    if show:
        plt.show()


# def compute_true_mask(Xq, N, F0, ):
    
    
#     absql = np.abs(Xq[F0, :, 0, :])
#     absNl = np.abs(N[F0, :, 0])
    
#     Tmask = np.argmax(np.concatenate((absql, Gn * absNl[:,:,None]), axis=2),axis=-1)
#     Tmask[20 * np.log10(np.abs(Xt[:, :, 0])) <= -10] = Q
#     Tmask[:, np.sum(Tmask == Q, axis=0) / (NFFT // 2 + 1) > 0.85] = Q

# def extract_feature_vector(stft_images):
#     # Assuming stft_images is a 3D numpy array with shape (num_frames, num_bins, num_channels)
#     num_frames, num_bins, num_channels = stft_images.shape
#     feature_vector = np.zeros((num_bins, num_channels, num_frames), dtype=np.float64)
    
#     for i in range(num_frames):
#         feature_vector[:, :, i] = stft_images[i]
    
#     return feature_vector