def get_dist(mtx, biom_data):
cols = [i['id'] for i in biom_data['columns']]
if args.create_splits:
dtvar_jc = dt.binary_dist_jaccard(mtx, strict=False)
dtvar_bc = dt.dist_bray_curtis(mtx, strict=False)
dtvar_mh = dt.dist_morisita_horn(mtx, strict=False)
dtvar_cb = dt.dist_canberra(mtx, strict=False)
dtvar_kz = dt.dist_kulczynski(mtx, strict=False)
# there are 10 combos
np_jc_l = np.tril(dtvar_jc)
np_bc_u = np.triu(dtvar_bc)
np_mh_u = np.triu(dtvar_mh)
np_mh_l = np.tril(dtvar_mh)
np_cb_u = np.triu(dtvar_cb)
np_cb_l = np.tril(dtvar_cb)
np_kz_u = np.triu(dtvar_kz)
np_kz_l = np.tril(dtvar_kz)
matrices = {
'jc_kz': np_jc_l + np_kz_u,
'jc_cb': np_jc_l + np_cb_u,
'jc_mh': np_jc_l + np_mh_u,
'jc_bc': np_jc_l + np_bc_u,
'kz_cb': np_kz_l + np_cb_u,
'kz_mh': np_kz_l + np_mh_u,
'kz_bc': np_kz_l + np_bc_u,
'cb_mh': np_cb_l + np_mh_u,
'cb_bc': np_cb_l + np_bc_u,
'mh_bc': np_mh_l + np_bc_u
}
#print(matrices['cb_mh'])
# convention: first(left) is lower left on heatmap
for spl in matrices:
df = pd.DataFrame(data=matrices[spl], index=cols, columns=cols)
#dict = df.to_dict()
out_file = os.path.join(args.basedir,
args.prefix + '_distance_' + spl + '.tsv')
df.to_csv(out_file, sep='\t', encoding='utf-8')
#print(dict)
#sys.exit()
return {}
else:
if args.metric == 'bray_curtis':
dtvar = dt.dist_bray_curtis(mtx, strict=False)
elif args.metric == 'morisita_horn':
dtvar = dt.dist_morisita_horn(mtx, strict=False)
elif args.metric == 'canberra':
dtvar = dt.dist_canberra(mtx, strict=False)
elif args.metric == 'jaccard':
dtvar = dt.binary_dist_jaccard(mtx, strict=False)
elif args.metric == 'kulczynski':
dtvar = dt.dist_kulczynski(mtx, strict=False)
else: # default
dtvar = dt.dist_bray_curtis(mtx, strict=False)
#sys.exit()
dist = distance.squareform(dtvar)
return dist
def test_dist_canberra_bug(self):
i = array([[0, 0, 1], [0, 1, 1]])
d = (1.0 / 2.0) * sum(
[abs(0.0 - 1.0) / (0.0 + 1.0),
abs(1.0 - 1.0) / (1.0 + 1.0)])
expected = array([[0.0, d], [d, 0.0]])
actual = dist_canberra(i)
assert_allclose(expected, actual)
def test_dist_canberra(self):
"""tests dist_canberra
tests inputs of empty mtx, zeros, and results compared with calcs done
by hand"""
self.assertFloatEqual(dist_canberra(self.zeromtx), zeros((4, 4), "d"))
mtx1expected = array([[0.0, 46.2 / 52.2], [46.2 / 52.2, 0.0]], "d")
self.assertFloatEqual(dist_canberra(self.mtx1), mtx1expected)
sparse1exp = ones((self.sparse1.shape[0], self.sparse1.shape[0]))
# remove diagonal
sparse1exp[0, 0] = sparse1exp[1, 1] = sparse1exp[2, 2] = sparse1exp[3, 3] = 0.0
sparse1exp[0, 1] = sparse1exp[1, 0] = (5.33 - 0.4) / (5.33 + 0.4)
self.assertFloatEqual(dist_canberra(self.sparse1), sparse1exp)
def get_dist(metric, mtx):
if metric == 'bray_curtis':
dtvar = dt.dist_bray_curtis(mtx, strict=False)
elif metric == 'morisita_horn':
dtvar = dt.dist_morisita_horn(mtx, strict=False)
elif metric == 'canberra':
dtvar = dt.dist_canberra(mtx, strict=False)
elif metric == 'jaccard':
dtvar = dt.binary_dist_jaccard(mtx, strict=False)
elif metric == 'kulczynski':
dtvar = dt.dist_kulczynski(mtx, strict=False)
else: # default
dtvar = dt.dist_bray_curtis(mtx, strict=False)
dist = distance.squareform(dtvar)
return dist
def get_dist(metric, mtx):
if metric == 'bray_curtis':
dtvar = dt.dist_bray_curtis(mtx, strict=False)
elif metric == 'morisita_horn':
dtvar = dt.dist_morisita_horn(mtx, strict=False)
elif metric == 'canberra':
dtvar = dt.dist_canberra(mtx, strict=False)
elif metric == 'jaccard':
dtvar = dt.binary_dist_jaccard(mtx, strict=False)
elif metric == 'kulczynski':
dtvar = dt.dist_kulczynski(mtx, strict=False)
else: # default
dtvar = dt.dist_bray_curtis(mtx, strict=False)
dist = distance.squareform( dtvar )
return dist
def get_dist(mtx, biom_data):
cols = [ i['id'] for i in biom_data['columns']]
if args.create_splits:
dtvar_jc = dt.binary_dist_jaccard(mtx, strict=False)
dtvar_bc = dt.dist_bray_curtis(mtx, strict=False)
dtvar_mh = dt.dist_morisita_horn(mtx, strict=False)
dtvar_cb = dt.dist_canberra(mtx, strict=False)
dtvar_kz = dt.dist_kulczynski(mtx, strict=False)
# there are 10 combos
np_jc_l = np.tril(dtvar_jc)
np_bc_u = np.triu(dtvar_bc)
np_mh_u = np.triu(dtvar_mh)
np_mh_l = np.tril(dtvar_mh)
np_cb_u = np.triu(dtvar_cb)
np_cb_l = np.tril(dtvar_cb)
np_kz_u = np.triu(dtvar_kz)
np_kz_l = np.tril(dtvar_kz)
matrices = {'jc_kz':np_jc_l + np_kz_u,'jc_cb':np_jc_l + np_cb_u,'jc_mh':np_jc_l + np_mh_u,'jc_bc':np_jc_l + np_bc_u ,'kz_cb':np_kz_l + np_cb_u,'kz_mh':np_kz_l + np_mh_u,'kz_bc':np_kz_l + np_bc_u,'cb_mh':np_cb_l + np_mh_u,'cb_bc':np_cb_l + np_bc_u,'mh_bc':np_mh_l + np_bc_u}
#print(matrices['cb_mh'])
# convention: first(left) is lower left on heatmap
for spl in matrices:
df = pd.DataFrame(data=matrices[spl],index=cols,columns=cols)
#dict = df.to_dict()
out_file = os.path.join(args.basedir, 'tmp', args.prefix+'_distance_'+spl+'.tsv')
df.to_csv(out_file, sep='\t', encoding='utf-8')
#out_file = os.path.join(args.basedir, 'tmp', args.prefix+'_distance_'+spl+'.json')
#out_fp = open(out_file,'w')
#out_fp.write(json.dumps(dict))
#out_fp.close()
# dtvar = np_jc_l + np_cb_u
# df = pd.DataFrame(data=dtvar,index=cols,columns=cols)
# out_file = os.path.join(args.basedir, 'tmp', args.prefix+'_distance_jc_cb.tsv')
# df.to_csv(out_file, sep='\t', encoding='utf-8')
#
#
# dtvar = np_jc_l + np_mh_u
# df = pd.DataFrame(data=dtvar,index=cols,columns=cols)
# out_file = os.path.join(args.basedir, 'tmp', args.prefix+'_distance_jc_mh.tsv')
# df.to_csv(out_file, sep='\t', encoding='utf-8')
#
#
# dtvar = np_jc_l + np_bc_u
# df = pd.DataFrame(data=dtvar,index=cols,columns=cols)
# out_file = os.path.join(args.basedir, 'tmp', args.prefix+'_distance_jc_bc.tsv')
# df.to_csv(out_file, sep='\t', encoding='utf-8')
#
#
# dtvar = np_kz_l + np_cb_u
# df = pd.DataFrame(data=dtvar,index=cols,columns=cols)
# out_file = os.path.join(args.basedir, 'tmp', args.prefix+'_distance_kz_cb.tsv')
# df.to_csv(out_file, sep='\t', encoding='utf-8')
#
#
# dtvar = np_kz_l + np_mh_u
# df = pd.DataFrame(data=dtvar,index=cols,columns=cols)
# out_file = os.path.join(args.basedir, 'tmp', args.prefix+'_distance_kz_mh.tsv')
# df.to_csv(out_file, sep='\t', encoding='utf-8')
#
#
# dtvar = np_kz_l + np_bc_u
# df = pd.DataFrame(data=dtvar,index=cols,columns=cols)
# out_file = os.path.join(args.basedir, 'tmp', args.prefix+'_distance_kz_bc.tsv')
# df.to_csv(out_file, sep='\t', encoding='utf-8')
#
#
# dtvar = np_cb_l + np_mh_u
# df = pd.DataFrame(data=dtvar,index=cols,columns=cols)
# out_file = os.path.join(args.basedir, 'tmp', args.prefix+'_distance_cb_mh.tsv')
# df.to_csv(out_file, sep='\t', encoding='utf-8')
#
#
# dtvar = np_cb_l + np_bc_u
# df = pd.DataFrame(data=dtvar,index=cols,columns=cols)
# out_file = os.path.join(args.basedir, 'tmp', args.prefix+'_distance_cb_bc.tsv')
# df.to_csv(out_file, sep='\t', encoding='utf-8')
#
#
# dtvar = np_mh_l + np_bc_u
# df = pd.DataFrame(data=dtvar,index=cols,columns=cols)
# dict = df.to_dict()
#df.to_csv(out_file, sep='\t', encoding='utf-8')
#print(dict)
#sys.exit()
return {}
else:
if args.metric == 'bray_curtis':
dtvar = dt.dist_bray_curtis(mtx, strict=False)
elif args.metric == 'morisita_horn':
dtvar = dt.dist_morisita_horn(mtx, strict=False)
elif args.metric == 'canberra':
dtvar = dt.dist_canberra(mtx, strict=False)
elif args.metric == 'jaccard':
dtvar = dt.binary_dist_jaccard(mtx, strict=False)
elif args.metric == 'kulczynski':
dtvar = dt.dist_kulczynski(mtx, strict=False)
else: # default
dtvar = dt.dist_bray_curtis(mtx, strict=False)
#sys.exit()
dist = distance.squareform( dtvar )
return dist