Python matplotlib.pyplot 模块，twinx() 实例源码

def plot_res(test_err, train_batch_loss, benchmark_err, epoch):
    flatui = ["#9b59b6", "#3498db", "#95a5a6", "#e74c3c", "#34495e", "#2ecc71"]

    test_x_val = np.array(list(x * 3 for x in range(0, len(test_err))))

    plt.plot(train_batch_loss[0],train_batch_loss[1], label="Training error", c=flatui[1], alpha=0.5)
    plt.plot(test_x_val, np.array(test_err), label="Test error", c=flatui[0])
    plt.axhline(y=benchmark_err[1], linestyle='dashed', label="No-modell error", c=flatui[2])
    plt.axhline(y=0.098, linestyle='dashed', label="State of the art error", c=flatui[3])

    plt.suptitle("Model error - cold queries")
    plt.yscale('log', nonposy='clip')
    plt.xlim([0,epoch+1])
    # second_axes = plt.twinx() # create the second axes, sharing x-axis
    # second_axes.set_yticks([0.2,0.4]) # list of your y values
    plt.xlabel('epoch')
    plt.ylabel('error')
    plt.legend(loc='upper right')
    plt.show()

def plot_history(logger, title):
    df = pd.DataFrame(logger.history)
    df[['acc', 'val_acc']].plot()
    plt.ylabel("accuracy")
    df[['loss', 'val_loss']].plot(linestyle='--', ax=plt.twinx())
    plt.ylabel("loss")
    plt.title(title)
    plt.show()

def plot_history(logger):
    df = pd.DataFrame(logger.history)
    df[['acc', 'val_acc']].plot()
    plt.ylabel("accuracy")
    df[['loss', 'val_loss']].plot(linestyle='--', ax=plt.twinx())
    plt.ylabel("loss")

def freeze_color_cycle(ax):
    """A function that freezes the color cycle. This is useful for example when
    the twinx() command is used and the color cycle would normally be reseted.

    Usage:

    import matplotlib.pyplot as plt
    import numpy as np

    plt.close('all')
    for i in range(3):
        plt.plot(np.random.random(20))

    ax=plt.gca()
    cc=freeze_color_cycle(ax)
    plt.twinx()
    ax=plt.gca()
    ax.set_color_cycle(cc)

    #plot some more on the new twined axes
    for i in range(3):
        plt.plot(np.random.random(20))

    #When we set-up a new figure we start with blue again
    fig=figure()
    for i in range(3):
        plot(np.random.random(20))

    """
    import matplotlib as mpl
    if mpl.__version__  >= '1.5.1':
        next_color=ax._get_lines.prop_cycler.next()['color']
    else:
        next_color=next(ax._get_lines.color_cycle)

    ix=plt.rcParams['axes.color_cycle'].index(next_color)
    color_cycle=plt.rcParams['axes.color_cycle'][ix:]+plt.rcParams['axes.color_cycle'][:ix]

    return color_cycle

def draw_timeseries(packets_all_hosts, cutoff_time_ms, save_dir,
                    output_postfix):
    """
    Draw a timeseries of packet latencies over time, in order that the packets
    were sent in
    """
    plt.figure()
    plt.suptitle("Packet latency over time")

    n_hosts = len(packets_all_hosts)
    for board_n in range(n_hosts):
        (filename, packet_ns, latencies_ms,
         total_n_packets) = packets_all_hosts[board_n]
        # Make copies, so that we don't distort the data that the other plots make use of
        packet_ns = list(packet_ns)
        latencies_ms = list(latencies_ms)

        (packet_ns, latencies_ms, dropped_packet_nos) = \
            add_dropped_packets_and_sort(total_n_packets, packet_ns, latencies_ms)

        plt.subplot(n_hosts, 1, 1 + board_n)
        plt.title(filename)

        line = plt.plot(packet_ns, latencies_ms)[0]
        if board_n == (n_hosts - 1):
            plt.xlabel("Packet no.")
        plt.ylabel("Latency (ms)")

        plt.twinx()  # Set up a second y-axis
        (bin_starts, bin_width_packets, drops) = drops_or_delays_in_each_bin(
            packet_ns, latencies_ms, cutoff_time_ms)
        bars = plt.bar(bin_starts,
                       drops,
                       bin_width_packets,
                       alpha=0.5,
                       color='red')
        plt.ylabel("Pct. packets dropped\nor > %d ms" % cutoff_time_ms)
        plt.ylim([0, 100])

        if board_n == 0:
            plt.legend([line, bars], ['Latencies', 'Pct. delayed packets'])

    plt.tight_layout()
    plt.subplots_adjust(top=0.9)  # to make room for suptitle

    plot_filename = os.path.join(save_dir, 'udp_latency_timeseries%s.png' %
                                 output_postfix)
    plt.savefig(plot_filename)