def draw_figures(data_dir='/auto/tdrive/mschachter/data', fig_dir='/auto/tdrive/mschachter/figures/encoder+decoder'):
agg_file = os.path.join(data_dir, 'aggregate', 'acoustic_encoder_decoder.h5')
agg = AcousticEncoderDecoderAggregator.load(agg_file)
# ###### these two functions write a csv file for decoder weights and draw barplots for decoder performance
export_decoder_datasets_for_glm(agg)
decoder_perf_stats()
def draw_figures(data_dir='/auto/tdrive/mschachter/data', fig_dir='/auto/tdrive/mschachter/figures/encoder+decoder'):
agg_file = os.path.join(data_dir, 'aggregate', 'acoustic_encoder_decoder.h5')
agg = AcousticEncoderDecoderAggregator.load(agg_file)
# tuning_curve_stats(agg)
export_psd_encoder_datasets_for_glm(agg)
"""
def draw_lags_vs_perf(data_dir='/auto/tdrive/mschachter/data'):
pfile = os.path.join(data_dir, 'GreBlu9508M', 'transforms', 'PairwiseCF_GreBlu9508M_Site4_Call1_L_raw.h5')
hf = h5py.File(pfile, 'r')
full_lags_ms = hf.attrs['lags']
hf.close()
max_lag = full_lags_ms.max()
agg_file = os.path.join(data_dir, 'aggregate', 'pard.h5')
agg = AcousticEncoderDecoderAggregator.load(agg_file)
print 'keys=',agg.df.keys()
lag_bnds = [ (-1., 1.), (-6., 6.), (-13., 13.), (-18., 18), (-23., 23), (-28., 28.), (-34., 34), (None,None)]
perfs_by_bound = list()
for lb,ub in lag_bnds:
if lb is None:
decomp = 'self+cross_locked'
else:
decomp = 'self+cross_locked_lim_%d_%d' % (lb, ub)
i = agg.df.decomp == decomp
print 'decomp=%s, i.sum()=%d' % (decomp, i.sum())
wkeys = agg.df.wkey[i].values
dperfs = np.array([agg.decoder_perfs[wkey] for wkey in wkeys])
perfs_by_bound.append(dperfs)
perfs_by_bound = np.array(perfs_by_bound)
pbb_mean = perfs_by_bound.mean(axis=1)
pbb_std = perfs_by_bound.std(axis=1)
plist = list()
for k,aprop in enumerate(REDUCED_ACOUSTIC_PROPS):
plist.append({'aprop':aprop, 'mean':pbb_mean[:, k], 'std':pbb_std[:, k]})
def _plot_pbb(pdata, ax):
_lagw = [2*x[1] for x in lag_bnds if x[1] is not None]
_lagw.append(max_lag*2)
plt.plot(_lagw, pdata['mean'], 'k-', linewidth=3.0, alpha=0.7)
plt.xlabel('Lag Width (ms)')
plt.ylabel('R2')
plt.axis('tight')
plt.title(pdata['aprop'])
multi_plot(plist, _plot_pbb, nrows=3, ncols=4)
plt.show()
def draw_figures(data_dir='/auto/tdrive/mschachter/data'):
agg_file = os.path.join(data_dir, 'aggregate', 'pard.h5')
agg = AcousticEncoderDecoderAggregator.load(agg_file)
decomps = ['spike_rate', 'full_psds', 'spike_rate+spike_sync', 'full_psds+full_cfs']
r2_vals = dict()
for decomp in decomps:
i = agg.df.decomp == decomp
assert i.sum() > 0
enc_r2 = list()
enc_cv_r2 = list()
dec_r2 = list()
dec_cv_r2 = list()
for wkey in agg.df.wkey[i].values:
eperf = agg.encoder_perfs[wkey].ravel()
eperf_cv = agg.encoder_cv_perfs[wkey].ravel()
enc_r2.extend(eperf)
enc_cv_r2.extend(eperf_cv)
dperf = agg.decoder_perfs[wkey].ravel()
dperf_cv = agg.decoder_cv_perfs[wkey].ravel()
dec_r2.extend(dperf)
dec_cv_r2.extend(dperf_cv)
print '############ %s ###########' % decomp
for r2,cvr2 in zip(dec_r2, dec_cv_r2):
print '%0.6f, %0.6f' % (r2, cvr2)
r2_vals[decomp] = {'enc_r2':np.array(enc_r2), 'enc_cv_r2':np.array(enc_cv_r2),
'dec_r2': np.array(dec_r2), 'dec_cv_r2': np.array(dec_cv_r2),}
figsize = (23, 10)
plt.figure(figsize=figsize)
nrows = 2
ncols = len(decomps)
gs = plt.GridSpec(nrows, ncols)
for k,decomp in enumerate(decomps):
ax = plt.subplot(gs[0, k])
x = np.linspace(0, 1, 50)
plt.plot(x, x, 'k-', alpha=0.5)
plt.plot(r2_vals[decomp]['enc_cv_r2'], r2_vals[decomp]['enc_r2'], 'ko', alpha=0.7)
plt.xlabel('Generalization R2')
plt.ylabel('Full R2')
plt.axis('tight')
plt.title('Encoder: %s' % decomp)
ax = plt.subplot(gs[1, k])
x = np.linspace(0, 1, 50)
plt.plot(x, x, 'k-', alpha=0.5)
plt.plot(r2_vals[decomp]['dec_cv_r2'], r2_vals[decomp]['dec_r2'], 'ko', alpha=0.7)
plt.xlabel('Generalization R2')
plt.ylabel('Full R2')
plt.axis('tight')
plt.title('Decoder: %s' % decomp)
plt.show()
def build_graph(bird='GreBlu9508M', block='Site4', segment='Call1', hemi='L', aprop='q2'):
# map electrodes to positions
df = pd.read_csv('/auto/tdrive/mschachter/data/aggregate/electrode_data+dist.csv')
i = (df.bird == bird) & (df.block == block) & (df.hemisphere == hemi)
g = df[i].groupby('electrode')
electrode_props = dict()
for e,gdf in g:
assert len(gdf) == 1
reg = str(gdf.region.values[0])
if reg.startswith('CM'):
reg = 'CM'
elif 'NCM' in reg:
reg = 'NCM'
elif reg in ['L', 'L2A', 'L2B', 'L2', 'L1', 'L3']:
reg = 'L'
electrode_props[e] = {'dist_midline':float(gdf.dist_midline.values[0]),
'dist_l2a':float(gdf.dist_l2a.values[0]),
'region':reg
}
#TODO get the lags in a less hacky way
data_dir = '/auto/tdrive/mschachter/data'
pfile = os.path.join(data_dir, 'GreBlu9508M', 'transforms', 'PairwiseCF_GreBlu9508M_Site4_Call1_L_raw.h5')
hf = h5py.File(pfile, 'r')
lags_ms = hf.attrs['lags']
hf.close()
lags_absmax = 6.
lags_i = np.abs(lags_ms) < lags_absmax
lags_short = lags_ms[lags_i]
# read the aggregator file
agg_file = os.path.join(data_dir, 'aggregate', 'pard.h5')
agg = AcousticEncoderDecoderAggregator.load(agg_file)
i = (agg.df.decomp == 'self+cross_locked_lim_-%d_%d' % (int(lags_absmax), int(lags_absmax))) & (agg.df.bird == bird) & (agg.df.block == block) & \
(agg.df.segment == segment) & (agg.df.hemi == hemi)
print 'i.sum()=%d' % i.sum()
assert i.sum() == 1, 'i.sum()=%d' % i.sum()
# get decoder weights
electrode_order = ROSTRAL_CAUDAL_ELECTRODES_RIGHT
if hemi == 'L':
electrode_order = ROSTRAL_CAUDAL_ELECTRODES_LEFT
electrode_order = [int(x) for x in electrode_order]
# get segment/decomp props
wkey = agg.df[i].wkey.values[0]
iindex = agg.df[i].iindex.values[0]
# get decoder weights
decoder_weights = agg.decoder_weights[wkey]
W = decoder_weights[:, :, :, REDUCED_ACOUSTIC_PROPS.index(aprop)]
# create a directed graph
g = nx.DiGraph()
# add a node for each electrode
for e in electrode_order:
g.add_node(e, dist_midline=electrode_props[e]['dist_midline'], dist_l2a=electrode_props[e]['dist_l2a'],
region=electrode_props[e]['region'], electrode=e)
# connect up the nodes
for k,e1 in enumerate(electrode_order):
for j in range(k):
e2 = electrode_order[j]
# get the coherency function and reduce it to a left and right side
cf = W[k, j, :]
assert len(cf) == len(lags_short)
cleft = np.sum(cf[lags_short <= 0]) # goes from e2 to e1
cright = np.sum(cf[lags_short >= 0]) # goes from e1 to e2
# add edges to the graph
g.add_edge(e2, e1, weight=float((cleft)))
g.add_edge(e1, e2, weight=float((cright)))
# plot the edge weights
ewts = np.array([g.edge[e1][e2]['weight'] for e1,e2 in g.edges()])
plt.figure()
plt.hist(ewts, bins=20, color='c')
plt.xlabel('Edge Weight')
plt.show()
# threshold out the edges with low weights
wthresh = 0.3
for e1,e2 in [(e1,e2) for e1,e2 in g.edges() if abs(g.edge[e1][e2]['weight']) < wthresh]:
g.remove_edge(e1, e2)
# write the graph to a file
nx.write_gexf(g, '/tmp/%s_graph.gexf' % aprop)
def draw_encoder_perfs(data_dir='/auto/tdrive/mschachter/data', fig_dir='/auto/tdrive/mschachter/figures/encoder_pairwise'):
#TODO: hack
# read the lags
pfile = os.path.join(data_dir, 'GreBlu9508M', 'transforms', 'PairwiseCF_GreBlu9508M_Site4_Call1_L_raw.h5')
hf = h5py.File(pfile, 'r')
lags_ms = hf.attrs['lags']
hf.close()
agg_file = os.path.join(data_dir, 'aggregate', 'pard.h5')
agg = AcousticEncoderDecoderAggregator.load(agg_file)
i = agg.df.decomp == 'self+cross_locked'
for (bird,block,segment,hemi),gdf in agg.df[i].groupby(['bird', 'block', 'segment','hemi']):
assert len(gdf) == 1
electrode_order = ROSTRAL_CAUDAL_ELECTRODES_RIGHT
if hemi == 'L':
electrode_order = ROSTRAL_CAUDAL_ELECTRODES_LEFT
# get segment/decomp props
wkey = gdf.wkey.values[0]
iindex = gdf.iindex.values[0]
# get encoder performances
encoder_perfs = agg.encoder_perfs[wkey]
index2electrode = agg.index2electrode[iindex]
assert len(lags_ms) == encoder_perfs.shape[-1]
nelectrodes = len(index2electrode)
r2_max = 0.3
# get decoder weights
decoder_weights = agg.decoder_weights[wkey]
dperfs = agg.decoder_perfs[wkey]
for k,aprop in enumerate(REDUCED_ACOUSTIC_PROPS):
dW = decoder_weights[:, :, :, k]
wmin = dW.min()
wmax = dW.max()
# make a figure for encoder performance
figsize = (24, 13)
fig = plt.figure(figsize=figsize)
fig.subplots_adjust(top=0.95, bottom=0.06, right=0.97, left=0.03, hspace=0.25, wspace=0.25)
gs = plt.GridSpec(nelectrodes, nelectrodes)
for i in range(nelectrodes):
for j in range(i):
eperf = encoder_perfs[i, j, :]
w = dW[i, j, :]
ax = plt.subplot(gs[i, j])
for li,er2 in enumerate(eperf):
a = (er2 / r2_max)*0.85
plt.axvline(lags_ms[li], c='r', alpha=a, linewidth=2.0)
plt.axvline(0, c='k', alpha=0.9)
# plt.axvline(-5., c='k', alpha=0.5)
# plt.axvline(5., c='k', alpha=0.5)
plt.axhline(0, c='k', alpha=0.5)
plt.plot(lags_ms, w, 'k-', linewidth=3.0, alpha=0.8)
plt.axis('tight')
plt.ylim(wmin, wmax)
plt.xticks([])
plt.yticks([])
if j == 0:
ytks = [wmin, 0, wmax]
plt.yticks(ytks, ['%0.2f' % x for x in ytks])
plt.ylabel('E%d' % electrode_order[i])
if i == nelectrodes-1:
xtks = [-40., 0, 40]
plt.xticks(xtks, ['%d' % x for x in xtks])
plt.xlabel('E%d' % electrode_order[j])
fname = '%s_%s_%s_%s_decoder_%s' % (bird, block, segment, hemi, aprop)
plt.suptitle('%s (R2=%0.2f)' % (fname, dperfs[k]))
fpath = os.path.join(fig_dir, '%s.svg' % fname)
print 'Saving %s...' % fpath
plt.savefig(fpath, facecolor='w')
plt.close('all')
# make a figure for encoder performance
figsize = (24, 13)
fig = plt.figure(figsize=figsize)
fig.subplots_adjust(top=0.95, bottom=0.06, right=0.97, left=0.03, hspace=0.25, wspace=0.25)
gs = plt.GridSpec(nelectrodes, nelectrodes)
for i in range(nelectrodes):
for j in range(i):
perf = encoder_perfs[i, j, :]
ax = plt.subplot(gs[i, j])
plt.axvline(0, c='k', alpha=0.5)
plt.plot(lags_ms, perf, 'k-', linewidth=2.0, alpha=0.7)
plt.axis('tight')
plt.xticks([])
plt.yticks([])
if j == 0:
ytks = [r2_max/3., (2/3.)*r2_max, r2_max]
plt.ylim(0, r2_max)
plt.yticks(ytks, ['%0.2f' % x for x in ytks])
plt.ylabel('E%d' % electrode_order[i])
if i == nelectrodes-1:
xtks = [-40., 0, 40]
plt.xticks(xtks, ['%d' % x for x in xtks])
plt.xlabel('E%d' % electrode_order[j])
fname = '%s_%s_%s_%s_encoder' % (bird, block, segment, hemi)
plt.suptitle(fname)
fname = os.path.join(fig_dir, '%s.svg' % fname)
print 'Saving %s...' % fname
plt.savefig(fname, facecolor='w')
plt.close('all')
