如何使用 Matplotlib 从预渲染图像创建子图

Question

目标是从某个程序产生的图形列表中创建一个子图。

这里，一个单独的图像由函数plot_conn()生成，并附加在all_fig中

最终，我想对这些数字进行细分，如下所示：

Matplotlib 或任何其他 Python 包是否可以做到这一点？

这 3 个数字是使用下面的代码生成的，并列在 all_figure 下

import numpy as np
import mne
from mne.connectivity import spectral_connectivity
from mne.viz import circular_layout, plot_connectivity_circle
import matplotlib.pyplot as plt


def generate_conn():
    # Generate data

    label_names = ['FP1', 'FP2', 'F3', 'F4', 'F7', 'F8', 'C3', 'C4',
                   'T3', 'T4', 'O1', 'O2']

    np.random.seed ( 42 )
    n_epochs = 5
    n_channels = len(label_names)
    n_times = 1000 
    data = np.random.rand ( n_epochs, n_channels, n_times )
    # Set sampling freq
    sfreq = 250  # A reasonable random choice


    # 10Hz sinus waves with random phase differences in each channel and epoch
    # Generate 10Hz sinus waves to show difference between connectivity
    # over time and over trials. Here we expect con over time = 1
    for i in range ( n_epochs ):
            for c in range ( n_channels ):
                wave_freq = 10
                epoch_len = n_times / sfreq
                # Introduce random phase for each channel
                phase = np.random.rand ( 1 ) * 10
                # Generate sinus wave
                x = np.linspace ( -wave_freq * epoch_len * np.pi + phase,
                                  wave_freq * epoch_len * np.pi + phase, n_times )
                data [i, c] = np.squeeze ( np.sin ( x ) )



    info = mne.create_info(ch_names=label_names,
                           ch_types=['eeg'] * len(label_names),
                           sfreq=sfreq)


    epochs = mne.EpochsArray(data, info)

    # Define freq bands
    Freq_Bands = {"delta": [1.25, 4.0],
                  "theta": [4.0, 8.0],
                  "alpha": [8.0, 13.0],
                  "beta": [13.0, 30.0],
                  "gamma": [30.0, 49.0]}


    n_freq_bands = len ( Freq_Bands )
    # Convert to tuples for the mne function
    fmin = tuple ( [list ( Freq_Bands.values () ) [f] [0] for f in range ( len ( Freq_Bands ) )] )
    fmax = tuple ( [list ( Freq_Bands.values () ) [f] [1] for f in range ( len ( Freq_Bands ) )] )

    # Connectivity methods
    connectivity_methods = ["plv"]
    n_con_methods = len ( connectivity_methods )

    # # Calculate PLV and wPLI - the MNE python implementation is over trials
    con, freqs, times, n_epochs, n_tapers = spectral_connectivity (
        epochs, method=connectivity_methods,
        mode="multitaper", sfreq=sfreq, fmin=fmin, fmax=fmax,
        faverage=True, verbose=0 )
    all_ch=epochs.ch_names

    return con,all_ch

def plot_conn(conmat,all_ch):
    lh_labels = ['FP1', 'F7', 'F3', 'C3', 'T3', 'O1']
    rh_labels = ['FP2', 'F8', 'F4', 'C4', 'T4', 'O2']
    node_order = lh_labels +rh_labels # Is this order tally with the con arrangement?
    node_angles = circular_layout ( all_ch, node_order, start_pos=90,
                                    group_boundaries=[0, len ( all_ch) // 2] )


    fig = plt.figure ( num=None, figsize=(8, 8), facecolor='black' )
    fig=plot_connectivity_circle ( conmat, all_ch, n_lines=300,
                               node_angles=node_angles,
                               title='All-to-All Connectivity '
                                     'Condition (PLI)_Delta', fig=fig )
    return fig


con,all_ch=generate_conn()
all_fig=[]
for idx in range (0,3):
    conmat = con [:, :, idx]
    fig=plot_conn(conmat,all_ch)
    all_fig.append(fig)

我会很感激任何提示。

Answer 1

其中一个肮脏的解决方案是将每个Figure 转换为Numpy array，然后垂直或水平堆叠数组。

1. 生成Numpy array

• 使用canvas.draw () 重绘plot_connectivity_circle 输出
• 通过使用np.frombuffer转换新的重绘图像来获取数组形式

    from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
    canvas = FigureCanvas ( fig )
    plot_connectivity_circle ( conmat, all_ch, n_lines=300,
                                    node_angles=node_angles,
                                    title=f'All-to-All Connectivity_ band_{bands}', fig=fig )
    
    canvas.draw ()
    s, (width, height) = canvas.print_to_buffer ()
    im0 = np.frombuffer ( s, np.uint8 ).reshape ( (height, width, 4) )

1. 通过堆叠数组创建子图

np.hstack ( all_fig ) # all_fig is a list of array

完整代码如下：

import mne
from mne.connectivity import spectral_connectivity
from mne.viz import circular_layout, plot_connectivity_circle
import numpy as np
from matplotlib import pyplot as plt
from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas

def generate_conn ():
    # Generate data

    label_names = ['FP1', 'FP2', 'F3', 'F4', 'F7', 'F8', 'C3', 'C4',
                   'T3', 'T4', 'O1', 'O2']

    np.random.seed ( 42 )
    n_epochs = 5
    n_channels = len ( label_names )
    n_times = 1000
    data = np.random.rand ( n_epochs, n_channels, n_times )
    # Set sampling freq
    sfreq = 250  # A reasonable random choice

    # 10Hz sinus waves with random phase differences in each channel and epoch
    # Generate 10Hz sinus waves to show difference between connectivity
    # over time and over trials. Here we expect con over time = 1
    for i in range ( n_epochs ):
        for c in range ( n_channels ):
            wave_freq = 10
            epoch_len = n_times / sfreq
            # Introduce random phase for each channel
            phase = np.random.rand ( 1 ) * 10
            # Generate sinus wave
            x = np.linspace ( -wave_freq * epoch_len * np.pi + phase,
                              wave_freq * epoch_len * np.pi + phase, n_times )
            data [i, c] = np.squeeze ( np.sin ( x ) )

    info = mne.create_info ( ch_names=label_names,
                             ch_types=['eeg'] * len ( label_names ),
                             sfreq=sfreq )

    epochs = mne.EpochsArray ( data, info )

    # Define freq bands
    Freq_Bands = {"delta": [1.25, 4.0],
                  "theta": [4.0, 8.0],
                  "alpha": [8.0, 13.0],
                  "beta": [13.0, 30.0],
                  "gamma": [30.0, 49.0]}

    n_freq_bands = len ( Freq_Bands )
    # Convert to tuples for the mne function
    fmin = tuple ( [list ( Freq_Bands.values () ) [f] [0] for f in range ( len ( Freq_Bands ) )] )
    fmax = tuple ( [list ( Freq_Bands.values () ) [f] [1] for f in range ( len ( Freq_Bands ) )] )

    # Connectivity methods
    connectivity_methods = ["plv"]
    n_con_methods = len ( connectivity_methods )

    # # Calculate PLV and wPLI - the MNE python implementation is over trials
    con, freqs, times, n_epochs, n_tapers = spectral_connectivity (
        epochs, method=connectivity_methods,
        mode="multitaper", sfreq=sfreq, fmin=fmin, fmax=fmax,
        faverage=True, verbose=0 )
    all_ch = epochs.ch_names

    return con, all_ch


def plot_conn (conmat, all_ch, idx, bands):
    lh_labels = ['FP1', 'F7', 'F3', 'C3', 'T3', 'O1']
    rh_labels = ['FP2', 'F8', 'F4', 'C4', 'T4', 'O2']
    node_order = lh_labels + rh_labels  # Is this order tally with the con arrangement?
    node_angles = circular_layout ( all_ch, node_order, start_pos=90,
                                    group_boundaries=[0, len ( all_ch ) // 2] )

    fig = plt.figure ( num=None, figsize=(8, 8), facecolor='black' )
    

    canvas = FigureCanvas ( fig )
    plot_connectivity_circle ( conmat, all_ch, n_lines=300,
                                    node_angles=node_angles,
                                    title=f'All-to-All Connectivity_ band_{bands}', fig=fig )
    
    canvas.draw ()
    s, (width, height) = canvas.print_to_buffer ()
    im0 = np.frombuffer ( s, np.uint8 ).reshape ( (height, width, 4) )
    return im0


con, all_ch = generate_conn ()

all_fig = [plot_conn ( con [:, :, idx], all_ch, idx, band ) for idx, band in enumerate ( ["delta", "theta", "alpha"] )]

SUBPLOT = np.hstack ( all_fig )



plt.imsave ( 'myimage.png', SUBPLOT )