def get_features(filename, features=features):
d = Dataset(filename)
ev = Evaluator()
ncols = len(features)
nrows = len(d['t'])
result = np.zeros( [nrows, ncols] )
for (idx, f) in enumerate(features):
print "Retrieving feature ", f
vec = ev.eval_expr(f, env = d)
if np.any(np.isnan(vec)):
print "Warning: NaN in", f
elif np.any(np.isinf(vec)):
print "Warning: inf in", f
result[:, idx] = vec
return result, d
def dataset_to_feature_matrix(d, features, start_idx=None, end_idx=None):
ev = Evaluator()
ncols = len(features)
t = d['t'][start_idx:end_idx]
nrows = len(t)
print "feature matrix shape:", [nrows, ncols]
result = np.zeros( [nrows, ncols] )
for (idx, f) in enumerate(features):
print "Retrieving feature ", f
vec = ev.eval_expr(f, env = d, start_idx=start_idx, end_idx=end_idx)
if np.any(np.isnan(vec)):
print "Warning: NaN in", f
elif np.any(np.isinf(vec)):
print "Warning: inf in", f
result[:, idx] = vec
return result