Пример #1
0
def plot_pulser_dac(x,
                    y,
                    y_err=None,
                    output_pdf=None,
                    title_suffix="",
                    atol_first_dev=1.0 * 1e-04,
                    atol_second_dev=1.0 * 1e-05):
    # plot result
    fig = Figure()
    FigureCanvas(fig)
    ax = fig.add_subplot(111)
    ax.errorbar(x, y, yerr=y_err, label='PlsrDAC', fmt='o')
    ax.set_title('PlsrDAC Measurement %s' % title_suffix)
    ax.set_xlabel("PlsrDAC")
    ax.set_ylabel('Voltage [V]')
    ax.grid(True)
    ax.set_xlim((0, max(x)))
    ax.legend(loc='upper left')
    if isinstance(output_pdf, PdfPages):
        output_pdf.savefig(fig)
    else:
        fig.show()

    # calculate 1st and 2nd deviation to estimate fit range
    tck = interpolate.splrep(x, y, k=3, s=0)
    xnew = np.linspace(min(x), max(x), num=100, endpoint=True)

    dev_1 = interpolate.splev(x, tck, der=1)
    dev_2 = interpolate.splev(x, tck, der=2)

    # calculate turning point
    turning_point_idx = np.where(
        np.logical_and(dev_1 > 0, dev_2 < -atol_second_dev))[0]
    if turning_point_idx.size:
        # take last index from array
        turning_point = turning_point_idx[-1]
    else:
        # select highest index
        turning_point = len(x) - 1

    # calculate slope
    slope_data_dev1_idx = np.where(dev_1 > 0)[0]
    slope_data_dev2_idx = np.where(np.isclose(dev_2, 0,
                                              atol=atol_second_dev))[0]
    slope_data_idx = np.intersect1d(slope_data_dev1_idx,
                                    slope_data_dev2_idx,
                                    assume_unique=True)
    slope_data_idx = slope_data_idx[slope_data_idx < turning_point]
    # index of slope fit values
    slope_idx = max(consecutive(slope_data_idx), key=len)

    # calculate plateau
    plateau_data_dev1_idx = np.where(np.isclose(dev_1, 0,
                                                atol=atol_first_dev))[0]
    plateau_data_dev2_idx = np.where(np.isclose(dev_2, 0,
                                                atol=atol_second_dev))[0]
    plateau_data_idx = np.intersect1d(plateau_data_dev1_idx,
                                      plateau_data_dev2_idx,
                                      assume_unique=True)
    plateau_data_idx = plateau_data_idx[plateau_data_idx > turning_point]
    # index of plateau fit values
    plateau_idx = max(consecutive(plateau_data_idx), key=len)

    fig = Figure()
    FigureCanvas(fig)
    ax1 = fig.add_subplot(311)
    ax1.set_title('PlsrDAC Fit Range %s' % title_suffix)
    ax1.plot(x, y, 'o', label='data')
    ax1.plot(x[slope_idx], y[slope_idx], 'ro', label='ramp')
    ax1.plot(x[plateau_idx], y[plateau_idx], 'go', label='plateau')
    ax1.plot(xnew, interpolate.splev(xnew, tck, der=0), label='B-spline')

    def slope_fit_fn(x, offset, slope):
        return offset + slope * x

    def plateau_fit_fn(x, offset):
        return offset

    try:
        slope_p_opt, slope_p_cov = optimize.curve_fit(
            slope_fit_fn,
            x[slope_idx],
            y[slope_idx],
            p0=[0.04, 0.0015],
            sigma=y_err[slope_idx] if y_err is not None else None,
            absolute_sigma=True)
    except (RuntimeError, TypeError):
        slope_p_opt = [np.nan, np.nan]
        slope_p_err = [np.nan, np.nan]
    else:
        slope_p_err = np.sqrt(np.diag(slope_p_cov))

    try:
        plateau_p_opt, plateau_p_cov = optimize.curve_fit(
            plateau_fit_fn,
            x[plateau_idx],
            y[plateau_idx],
            p0=[1.3],
            sigma=y_err[plateau_idx] if y_err is not None else None,
            absolute_sigma=True)
    # in case of failing fit or missing plateau
    except (RuntimeError, TypeError):
        plateau_p_opt = [np.nan]
        plateau_p_err = [np.nan]
    else:
        plateau_p_err = np.sqrt(np.diag(plateau_p_cov))


#     slope_p_opt = polyfit(x[slope_idx], y[slope_idx], 1)
#     slope_fit_fn = poly1d(slope_p_opt)
#     plateau_p_opt = polyfit(x[plateau_idx], y[plateau_idx], 0)
#     plateau_fit_fn = poly1d(plateau_p_opt)
    ax1.set_ylabel('Voltage [V]')
    ax1.set_xlabel("PlsrDAC")
    ax1.set_xlim((0, max(x)))
    ax1.set_ylim(bottom=0)
    ax1.legend(loc='best')

    ax2 = fig.add_subplot(312)
    ax2.plot(x, dev_1, label='1st dev')
    ax2.ticklabel_format(style='sci', axis='y', scilimits=(0, 0))
    ax2.set_xlabel("PlsrDAC")
    ax2.set_xlim((0, max(x)))
    ax2.legend(loc='best')

    ax3 = fig.add_subplot(313)
    ax3.plot(x, dev_2, label='2st dev')
    ax3.ticklabel_format(style='sci', axis='y', scilimits=(0, 0))
    ax3.plot(x[turning_point],
             dev_2[turning_point],
             'rx',
             label='Turning point')
    ax3.set_xlabel("PlsrDAC")
    ax3.set_xlim((0, max(x)))
    ax3.legend(loc='best')
    fig.tight_layout()
    if isinstance(output_pdf, PdfPages):
        output_pdf.savefig(fig)
    else:
        fig.show()

    # plot and fit result
    fig = Figure()
    FigureCanvas(fig)
    ax = fig.add_subplot(111)
    ax.errorbar(x, y, None, label='PlsrDAC', fmt='o', zorder=1)
    #     ax.plot(x[slope_idx], y[slope_idx], 'ro', label='PlsrDAC ramp', zorder=2)
    #     ax.plot(x[plateau_idx], y[plateau_idx], 'go', label='PlsrDAC plateau', zorder=2)
    ax.plot(x,
            np.vectorize(slope_fit_fn)(x, *slope_p_opt),
            '--k',
            label='%.5f+/-%.5f+\n%.5f+/-%.5f*x' %
            (slope_p_opt[0], slope_p_err[0], slope_p_opt[1], slope_p_err[1]),
            zorder=4)
    ax.plot(x,
            np.vectorize(plateau_fit_fn)(x, *plateau_p_opt),
            '-k',
            label='%.5f+/-%.5f' % (plateau_p_opt[0], plateau_p_err[0]),
            zorder=3)
    ax.set_title('PlsrDAC Calibration %s' % title_suffix)
    ax.set_xlabel("PlsrDAC")
    ax.set_ylabel('Voltage [V]')
    ax.grid(True)
    ax.set_xlim((0, max(x)))
    ax.set_ylim(bottom=0)
    ax.legend(loc='upper left')
    # second plot with shared axis
    divider = make_axes_locatable(ax)
    ax_bottom_plot = divider.append_axes("bottom", 1.0, pad=0.1, sharex=ax)
    ax_bottom_plot.bar(x,
                       (y - np.vectorize(slope_fit_fn)(x, *slope_p_opt)) * 1e3,
                       align='center')
    ax_bottom_plot.grid(True)
    ax_bottom_plot.set_xlabel("PlsrDAC")
    ax_bottom_plot.set_ylabel("$\Delta$Voltage [mV]")

    if isinstance(output_pdf, PdfPages):
        output_pdf.savefig(fig)
    else:
        fig.show()

    return slope_p_opt, slope_p_err, plateau_p_opt, plateau_p_err
Пример #2
0
    def analyze(self):
        with tb.open_file(self.output_filename + '.h5', 'r') as in_file_h5:
            output_pdf = PdfPages(self.output_filename + '.pdf')
            data = in_file_h5.root.plsr_dac_data[:]
            # Calculate mean PlsrDAC transfer function
            mean_data = np.zeros(shape=(len(self.pulser_dac_parameters), ),
                                 dtype=[("PlsrDAC", np.int32),
                                        ('voltage_mean', np.float),
                                        ('voltage_rms', np.float)])
            for index, parameter in enumerate(self.pulser_dac_parameters):
                mean_data["PlsrDAC"][index] = parameter
                mean_data['voltage_mean'][index] = data['voltage'][
                    data["PlsrDAC"] == parameter].mean()
                mean_data['voltage_rms'][index] = data['voltage'][
                    data["PlsrDAC"] == parameter].std()

            slope_fit, slope_err, plateau_fit, plateau_err = plot_pulser_dac(
                mean_data['PlsrDAC'],
                mean_data['voltage_mean'],
                mean_data['voltage_rms'],
                output_pdf,
                title_suffix="(DC " + ", ".join([
                    str(cols[0]) if len(cols) == 1 else
                    (str(cols[0]) + " - " + str(cols[-1]))
                    for cols in consecutive(self.colpr_addr_parameters)
                ]) + ")")
            self.register.calibration_parameters[
                'Vcal_Coeff_0'] = np.nan_to_num(slope_fit[0] *
                                                1000.0)  # store in mV
            self.register.calibration_parameters[
                'Vcal_Coeff_1'] = np.nan_to_num(slope_fit[1] *
                                                1000.0)  # store in mV/DAC

            # plot per double column
            # Calculate mean PlsrDAC transfer function
            # TODO: Store result in file
            dc_data_stability = np.zeros(
                shape=(len(self.colpr_addr_parameters) *
                       len(self.pulser_dac_parameters), ),
                dtype=[("colpr_addr", np.int32), ("PlsrDAC", np.int32),
                       ('voltage_mean', np.float), ('voltage_rms', np.float)])
            dc_data = np.zeros(shape=(len(self.colpr_addr_parameters), ),
                               dtype=[("colpr_addr", np.int32),
                                      ('Vcal_Coeff_0', np.float),
                                      ('Vcal_Coeff_1', np.float),
                                      ('Vcal_Coeff_0_err', np.float),
                                      ('Vcal_Coeff_1_err', np.float),
                                      ('Vcal_plateau', np.float),
                                      ('Vcal_plateau_err', np.float)])
            for dc_index, dc_parameter in enumerate(
                    self.colpr_addr_parameters):
                mean_data = np.zeros(shape=(len(self.pulser_dac_parameters), ),
                                     dtype=[("PlsrDAC", np.int32),
                                            ('voltage_mean', np.float),
                                            ('voltage_rms', np.float)])
                for index, parameter in enumerate(self.pulser_dac_parameters):
                    mean_data["PlsrDAC"][index] = parameter
                    mean_data['voltage_mean'][index] = data['voltage'][
                        np.logical_and(
                            data["PlsrDAC"] == parameter,
                            data['colpr_addr'] == dc_parameter)].mean()
                    mean_data['voltage_rms'][index] = data['voltage'][
                        np.logical_and(
                            data["PlsrDAC"] == parameter,
                            data['colpr_addr'] == dc_parameter)].std()

                dc_data_stability["colpr_addr"][
                    dc_index * len(self.pulser_dac_parameters):(dc_index + 1) *
                    len(self.pulser_dac_parameters)] = dc_parameter
                dc_data_stability["PlsrDAC"][
                    dc_index * len(self.pulser_dac_parameters):(dc_index + 1) *
                    len(self.pulser_dac_parameters)] = mean_data["PlsrDAC"]
                dc_data_stability["voltage_mean"][
                    dc_index * len(self.pulser_dac_parameters):(dc_index + 1) *
                    len(self.pulser_dac_parameters
                        )] = mean_data['voltage_mean']
                dc_data_stability["voltage_rms"][
                    dc_index * len(self.pulser_dac_parameters):(dc_index + 1) *
                    len(self.pulser_dac_parameters)] = mean_data['voltage_rms']

                slope_fit, slope_err, plateau_fit, plateau_err = plot_pulser_dac(
                    mean_data['PlsrDAC'],
                    mean_data['voltage_mean'],
                    mean_data['voltage_rms'],
                    output_pdf,
                    title_suffix="(DC " + str(dc_parameter) + ")")
                dc_data["colpr_addr"][dc_index] = dc_parameter
                dc_data['Vcal_Coeff_0'][
                    dc_index] = slope_fit[0] * 1000.0  # offset
                dc_data['Vcal_Coeff_1'][
                    dc_index] = slope_fit[1] * 1000.0  # slope
                dc_data['Vcal_Coeff_0_err'][
                    dc_index] = slope_err[0] * 1000.0  # offset error
                dc_data['Vcal_Coeff_1_err'][
                    dc_index] = slope_err[1] * 1000.0  # slope error
                dc_data['Vcal_plateau'][
                    dc_index] = plateau_fit[0] * 1000.0  # plateau
                dc_data['Vcal_plateau_err'][
                    dc_index] = plateau_err[0] * 1000.0  # plateau error

            for index, parameter in enumerate(self.pulser_dac_parameters):
                fig = Figure()
                FigureCanvas(fig)
                ax = fig.add_subplot(111)
                ax.set_title('PlsrDAC Voltage vs. DC at PlsrDAC %d' %
                             parameter)
                ax.errorbar(dc_data_stability["colpr_addr"][
                    dc_data_stability['PlsrDAC'] == parameter],
                            dc_data_stability['voltage_mean'][
                                dc_data_stability['PlsrDAC'] == parameter],
                            yerr=dc_data_stability['voltage_rms'][
                                dc_data_stability['PlsrDAC'] == parameter],
                            fmt='o',
                            label='PlsrDAC Voltage')
                ax.set_ylabel('Voltage [V]')
                ax.set_xlabel("Colpr_Addr")
                ax.set_xlim(-0.5, 39.5)
                output_pdf.savefig(fig)

            fig = Figure()
            FigureCanvas(fig)
            ax1 = fig.add_subplot(311)
            ax1.set_title('PlsrDAC Vcal_Coeff_0 vs. DC')
            ax1.errorbar(dc_data["colpr_addr"],
                         dc_data['Vcal_Coeff_0'],
                         yerr=dc_data['Vcal_Coeff_0_err'],
                         fmt='o',
                         label='Vcal_Coeff_0')
            ax1.set_ylabel('Vcal_Coeff_0 [mV]')
            ax1.set_xlabel("Colpr_Addr")
            ax1.set_xlim(-0.5, 39.5)

            ax2 = fig.add_subplot(312)
            ax2.set_title('PlsrDAC Vcal_Coeff_1 vs. DC')
            ax2.errorbar(dc_data["colpr_addr"],
                         dc_data['Vcal_Coeff_1'],
                         yerr=dc_data['Vcal_Coeff_1_err'],
                         fmt='o',
                         label='Vcal_Coeff_1')
            ax2.set_ylabel('Vcal_Coeff_1 [mV/DAC]')
            ax2.set_xlabel("Colpr_Addr")
            ax2.set_xlim(-0.5, 39.5)

            ax2 = fig.add_subplot(313)
            ax2.set_title('PlsrDAC Plateau vs. DC')
            ax2.errorbar(dc_data["colpr_addr"],
                         dc_data['Vcal_plateau'],
                         yerr=dc_data['Vcal_plateau_err'],
                         fmt='o',
                         label='Plateau')
            ax2.set_ylabel('Plateau [mV]')
            ax2.set_xlabel("Colpr_Addr")
            ax2.set_xlim(-0.5, 39.5)

            fig.tight_layout()
            output_pdf.savefig(fig)

            output_pdf.close()