def test_group_denovo_specific(self):
obs1, obs2 = group_denovo(self.testclusterslt10, ['.....((((.......))))...'], 10)
exp1 = {
'(((...)))...(((...)))': [(9, 10)],
'(((...)))': [(1, 2), (3, 4), (11, 12), (5, 6)],
'.....((((.......))))...': [(15, 16), (7, 8)],
'............': [(13, 14)]}
exp2 = ['.....((((.......))))...', '(((...)))...(((...)))', '(((...)))', '............']
self.assertEqual(obs1, exp1)
self.assertEqual(obs2, exp2)
def test_group_denovo_all(self):
obs1, obs2 = group_denovo(self.testclusterslt10, self.testclusterslt10.keys(), 10)
print "\n", obs1
print obs2
exp1 = {
'(((...)))...(((...)))': [(9, 10)],
'(((...)))': [(1, 2), (3, 4), (11, 12), (5, 6)],
'.....((((.......))))...': [(7, 8)],
'............': [(13, 14)]}
exp2 = ['.....((((.......))))...', '(((...)))...(((...)))', '(((...)))', '............']
self.assertEqual(obs1, exp1)
self.assertEqual(obs2, exp2)
#broken out by shapes. No comparison needed at first if not same shape,
#as most likely not simmilar enough
files = []
hold = {}
#pool = Pool(processes=args.c)
#run the pool over all shape groups to get final grouped structgroups
fout = open(otufolder + "shapesizes.txt", 'w')
groupnum = 1
for shapegroup in groups_shape.keys():
#write out each group to file for use in subprocess
groupinfo = {struct: structgroups[struct] for struct in groups_shape[shapegroup]}
#fout.write(shapegroup + "\t" + str(len(groupinfo)) + "\n")
#pool.apply_async(func=group_by_distance,
# args=(groupinfo, structscore, None, None, args.nr), callback=hold.update)
stime = time()
hold.update(group_denovo(groupinfo, groupinfo.keys(), structscore))
print len(groupinfo), "clusters:", str((time()-stime)/60), "min"
fout.close()
#memory saving wipe of structgroups, groups_shape, and groupinfo
groups_shape.clear()
del groups_shape
groupinfo.clear()
del groupinfo
structgroups.clear()
del structgroups
#pool.close()
#pool.join()
#hold should now be the combined dictionaries from all calls of
#group_by_forester, aka new structgroups
#do one more grouping with all remaining structs regardless of shape
#print len(hold), "clusters fgrouping"