Vizualization module

MRS/viz.py

@@ -0,0 +1,88 @@
+"""
+
+Visualization functions
+
+"""
+
+import numpy as np
+import scipy.stats as stats
+import matplotlib.pyplot as plt
+
+def stacked_spectra(f, spectra, fig=None):
+
+    """
+    Plot stacked spectra
+
+    Parameters
+    ----------
+
+    """
+    if fig is None:
+        fig, ax = plt.subplots(1)
+    else:
+        fig = fig
+        ax = fig.get_axes()[0]
+
+    offset = 0
+    for s in spectra:
+        ax.plot(f, s.squeeze()+offset, color='k')
+        offset += np.std(s)/2.0
+    ax.invert_xaxis()
+    return fig
+
+def plot_spectra(f, spectra, n_boots=1000, fig=None, pct=68):
+    """
+    Plot average spectra with boot-strapped confidence intervals
+
+    Parameters
+    ----------
+    pct : how many percentiles to include in the CI (default to 95%)
+
+    """
+    if fig is None:
+        fig, ax = plt.subplots(1)
+    else:
+        fig = fig
+        ax = fig.get_axes()[0]
+
+    boots = np.zeros((n_boots, spectra.shape[-1]))
+    for b in xrange(n_boots):
+        boots[b] = np.random.random_integers(0,
+                                             spectra.shape[-1]-1,
+                                             spectra.shape[-1])
+
+    alpha = (100 - pct)/2.
+    CI = np.zeros(spectra.shape[-1])
+    for x in xrange(CI.shape[0]):
+        CI[x] = (stats.scoreatpercentile(boots[:, x], 100 - alpha) -
+                 stats.scoreatpercentile(boots[:, x], alpha))
+
+    y = np.mean(spectra, 0)
+    ax.plot(f, y, color='k')
+    ax.fill_between(f, y-CI, y+CI, alpha=0.2, color='b')
+    ax.invert_xaxis()
+    return fig
+
+
+def plot_gaba(G):
+    """
+    This is the basic three-panel plot: echo1/echo2/diff with model fits
+
+    """
+    G.fit_gaba()
+    fig, ax = plt.subplots(3)
+    ax[0].plot(G.f_ppm, np.mean(G.echo1,0))
+    ax[0].plot(G.f_ppm[G.cr_idx], stats.nanmean(G.creatine_model, 0), 'r')
+    Cr_text = 'Cr params \n freq0: %1.2f\n area: %1.4f \n hwhm: %1.2f \n phase: %1.2f \n offset: %1.2f \n drift %1.2f'%(tuple([w for w in np.mean(G.creatine_params, 0)]))
+    ax[0].text(0.9, 0.6, Cr_text, horizontalalignment='center', verticalalignment='center', transform = ax[0].transAxes)
+    ax[1].plot(G.f_ppm, np.mean(G.echo2,0))
+    ax[2].plot(G.f_ppm, np.mean(G.diff_spectra, 0))
+    ax[2].plot(G.f_ppm[G.gaba_idx], stats.nanmean(G.gaba_model, 0), 'r')
+    gaba_text = 'GABA params \n freq0: %1.2f\n sigma: %1.2f \n amp: %1.4f \n offset: %1.2f \n drift %1.2f'%(tuple([w for w in np.mean(G.gaba_params, 0)]))
+    ax[2].text(0.9, 0.5, gaba_text, horizontalalignment='center', verticalalignment='center', transform = ax[2].transAxes)
+    ax[2].set_xlabel('ppm')
+    fig.set_size_inches([10,8])
+    for a in ax:
+        a.invert_xaxis()
+
+    return fig