def structure_plot(fig, molids, activations=None):
"""plot molecule structures"""
if activations != None:
assert len(activations[0]) == len(molids[0])
assert len(activations[1]) == len(molids[1])
id2name = defaultdict(str, rdl.get_id2name())
all_molids = molids[0] + molids[1]
all_activations = np.hstack(activations)
cr = utils.ceiled_root(len(all_molids))
for i, molid in enumerate(all_molids):
ax = fig.add_subplot(cr, cr, i+1)
try:
img = plt.imread(os.path.join(structures_path, molid + '.png'))
except Exception, e:
img = np.zeros(img.shape)
ax.imshow(img)
ax.set_xticks([])
ax.set_yticks([])
if i >= len(molids[0]):
for child in ax.get_children():
if isinstance(child, plt.matplotlib.spines.Spine):
child.set_color('#ff0000')
if activations != None:
ax.set_title('{}: {:.2f}'.format(id2name[molid], all_activations[i]),
rotation='20')
reload(plib)
plt.close('all')
active_thresh = 0.5
descriptor = 'saito_desc'
method = 'svr_ens'
selection = 'linear'
inpath = '/Users/dedan/projects/master/results/predict/search/conv_features'
outpath = '/Users/dedan/projects/master/results/predict'
data_path = "/Users/dedan/projects/master/data"
glomeruli = ['Or10a', 'Or42b', 'Or47b']
glom = 'Or42b'
glom = sys.argv[1]
# load stuff
id2name = defaultdict(str, rdl.get_id2name())
door2id = json.load(open(os.path.join(data_path, 'door2id.json')))
id2cas = defaultdict(str, {val[0]: key for key, val in door2id.items() if val})
daniel_set = json.load(open(os.path.join(data_path, 'daniel_set.json')))
daniel_set_molid = [door2id[cas][0] for cas in daniel_set]
# get the best parameters from the search result
config = rdl.get_best_params(inpath, descriptor, glom, method, selection)
features = run_lib.prepare_features(config)
data, targets, molids = run_lib.load_data_targets(config, features)
# fit model
print("use {} molecules for training".format(data.shape[0]))
sel_scores = run_lib.get_selection_score(config, data, targets)
data = flib.select_k_best(data, sel_scores, config['feature_selection']['k_best'])
