def setUp(self):
_, test_table = build_block_model(rank=2,
hoced=20,
hsced=20,
spar=2e3,
C_=2e3,
num_samples=50,
num_features=500,
mapping_on=False)
feat_ids = ['F%d' % i for i in range(test_table.shape[0])]
samp_ids = ['L%d' % i for i in range(test_table.shape[1])]
self.test_table = Table(test_table, feat_ids, samp_ids)
def create_test_table():
_, test_table = build_block_model(rank=2,
hoced=20,
hsced=20,
spar=2e3,
C_=2e3,
num_samples=50,
num_features=500,
mapping_on=False)
# the rclr is tested in other places
# this is just used as input into
# the OptSpace tests
test_table = np.array(test_table)
table_rclr = rclr(test_table)
return test_table, table_rclr
observed_table = niche_sort(otutabledf, mappingdf['host_num'])
mappingdf=mappingdf.T[observed_table.index].T
otutabledf=observed_table.copy()
otutabledf.to_dense().to_csv("cluster_models/base_model_keyboard_table.csv",sep=',', encoding='utf-8')
mappingdf.to_dense().to_csv("cluster_models/base_model_keyboard_meta.csv",sep=',', encoding='utf-8')
######### build the model #########
x0 = [3, 20, 20, 1e2, 1e2,1e1]
bnds = ((3,3),(0,1e2),(0,2e3),(0,1e10),(0,5e1),(1,10))
model_fit=minimize_model(x0,bnds,np.array(otutabledf.T[:104].T.as_matrix()))
base_truth,X_noise_sub=build_block_model(3, model_fit.x[1],
model_fit.x[2], model_fit.x[3]
, model_fit.x[4]
,otutabledf.shape[1]
,otutabledf.shape[0]
,overlap=model_fit.x[5]
,mapping_on=False)
save_base=[]
save_sub=[]
for rank_,overlap_ in zip([2],[20]):
#subsample_points=np.logspace(2,4,4)
seq_depth={500:3.05e2,
1000:6.1e2,
2000:1.25e3,
4000:2.5e3,
10000:6.05e3}
for sub_,model_peram in seq_depth.items():