3ecb4a4df1def4c9496bd676adad3564f746a02d,innerfidelity/innerfidelity_transformation.py,,main,#,19

Before Change

    //     legend.append(names[i])

    plt.plot(f, diff.raw)
    legend.append("Mean smoothed")

    plt.xlabel("Frequency (Hz)")
    plt.semilogx()
    plt.xlim([10, 30000])

After Change

def main():
    models = dict()
    for file_path in glob(os.path.join("transformation", "*.jpg")):
        print(file_path)
        file_path = os.path.abspath(file_path)
        file_name = os.path.split(file_path)[-1]
        file_name = file_name.split(".")[0]
        name, kind = file_name.split("_")
        if name not in models:
            models[name] = dict()

        with open(file_path, "rb") as f:
            model = os.path.split(file_path)[-1].split(".")[0]
            im = Image.open(file_path)

        // Add -1px to left and top, +1px to right and bottom of tight crop box in parse_innerfidelity
        fr = ImageGraphParser.parse_cropped(
            im,
            name=model,
            f_min=10,
            f_max=20000,
            a_min=-50,
            a_max=20
        )
        fr.interpolate()
        models[name][kind] = fr

    diffs = []
    names = []
    for model, frs in models.items():
        diffs.append(frs["before"].raw - frs["after"].raw)
        names.append(model)
    diffs = np.vstack(diffs)
    diff = np.mean(diffs, axis=0)
    f = models[list(models.keys())[0]]["before"].frequency
    diff = FrequencyResponse(name="Innerfidelity Transformation", frequency=f, raw=diff)
    diff.smooth(window_size=1/9, iterations=10)
    diff.raw = diff.smoothed
    diff.smoothed = np.array([])

    fig, ax = plt.subplots()
    legend = []

    for i, d in enumerate(diffs):
        fr = FrequencyResponse(name=names[i], frequency=f, raw=d)
        plt.plot(f, fr.raw)
        legend.append(names[i])

    plt.xlabel("Frequency (Hz)")

Instances