def test_backfill_names_gap(self):
"""correctly backfill names"""
consensus_tree = DndParser("(((s1,s2)g1,(s3,s4)g2,(s5,s6)g3)f1)o1;")
rank_lookup = {'s':6,'g':5,'f':4,'o':3,'c':2,'p':1,'k':0}
for n in consensus_tree.traverse(include_self=True):
n.Rank = rank_lookup[n.Name[0]]
input = "((((1)s1,(2)s2),((3)s3,(4)s5)))o1;"
lookup = dict([(n.Name, n) for n in consensus_tree.traverse(include_self=True)])
#exp = "((((1)s1,(2)s2)g1,((3)'g2; s3',(4)'g3; s5')))'o1; f1'"
t = DndParser(input)
t.Rank = 3
t.Children[0].Rank = None
t.Children[0].Children[0].Rank = None
t.Children[0].Children[1].Rank = None
t.Children[0].Children[0].Children[0].Rank = 6
t.Children[0].Children[0].Children[1].Rank = 6
t.Children[0].Children[1].Children[0].Rank = 6
t.Children[0].Children[1].Children[1].Rank = 6
backfill_names_gap(t, lookup)
self.assertEqual(t.BackFillNames, ['o1'])
self.assertEqual(t.Children[0].BackFillNames, [])
self.assertEqual(t.Children[0].Children[0].BackFillNames, [])
self.assertEqual(t.Children[0].Children[1].BackFillNames, [])
self.assertEqual(t.Children[0].Children[0].Children[0].BackFillNames, ['f1','g1','s1'])
self.assertEqual(t.Children[0].Children[0].Children[1].BackFillNames, ['f1','g1','s2'])
self.assertEqual(t.Children[0].Children[1].Children[0].BackFillNames, ['f1','g2','s3'])
self.assertEqual(t.Children[0].Children[1].Children[1].BackFillNames, ['f1','g3','s5'])
def test_reroot(self):
"""Should correctly reroot a tree"""
t = DndParser("(((a,b)c,(d,e)f)g,(h,i)j);")
tips = ['a','b']
for n in t.traverse():
n.Length = 1.0
# note, g is lost because it has a single descendent and gets pruned off
exp = "((a:1.0,b:1.0)c:0.5,((d:1.0,e:1.0)f:1.0,(h:1.0,i:1.0)j:2.0):0.5);"
obs = reroot(t, tips)
self.assertEqual(obs.getNewick(with_distances=True), exp)
def test_commonname_promotion(self):
"""correctly promote names if possible"""
consensus_tree = DndParser("(((s1,s2)g1,(s3,s4)g2,(s5,s6)g3)f1)o1;")
rank_lookup = {'s':6,'g':5,'f':4,'o':3,'c':2,'p':1,'k':0}
for n in consensus_tree.traverse(include_self=True):
n.Rank = rank_lookup[n.Name[0]]
input = "((((1)s1,(2)s2),((3)s3,(4)s5)))o1;"
lookup = dict([(n.Name, n) for n in consensus_tree.traverse(include_self=True)])
exp = "((((1)s1,(2)s2)g1,((3)'g2; s3',(4)'g3; s5')))'o1; f1';"
t = DndParser(input)
t.Rank = 3
t.Children[0].Rank = None
t.Children[0].Children[0].Rank = None
t.Children[0].Children[1].Rank = None
t.Children[0].Children[0].Children[0].Rank = 6
t.Children[0].Children[0].Children[1].Rank = 6
t.Children[0].Children[1].Children[0].Rank = 6
t.Children[0].Children[1].Children[1].Rank = 6
backfill_names_gap(t, lookup)
commonname_promotion(t)
self.assertEqual(t.getNewick(with_distances=False), exp)
def test_gnodedata(self):
"""DndParser should assign Name to internal nodes correctly"""
t = DndParser(nodedata)
self.assertEqual(len(t), 2)
self.assertEqual(len(t[0]), 0) #first child is terminal
self.assertEqual(len(t[1]), 2) #second child has two children
self.assertEqual(str(t), '(abc:3.0,(def:4.0,ghi:5.0)jkl:6.0);')
info_dict = {}
for node in t.traverse():
info_dict[node.Name] = node.Length
self.assertEqual(info_dict['abc'], 3.0)
self.assertEqual(info_dict['def'], 4.0)
self.assertEqual(info_dict['ghi'], 5.0)
self.assertEqual(info_dict['jkl'], 6.0)
def test_gnodedata(self):
"""DndParser should assign Name to internal nodes correctly"""
t = DndParser(nodedata)
self.assertEqual(len(t), 2)
self.assertEqual(len(t[0]), 0) #first child is terminal
self.assertEqual(len(t[1]), 2) #second child has two children
self.assertEqual(str(t), '(abc:3.0,(def:4.0,ghi:5.0)jkl:6.0);')
info_dict = {}
for node in t.traverse():
info_dict[node.Name] = node.Length
self.assertEqual(info_dict['abc'], 3.0)
self.assertEqual(info_dict['def'], 4.0)
self.assertEqual(info_dict['ghi'], 5.0)
self.assertEqual(info_dict['jkl'], 6.0)
def test_backfill_names_dangling(self):
"""correctly fill in dangling missing ranks"""
consensus_tree = DndParser("(((s1,s2)g1,(s3,s4)g2,(s5,s6)g3)f1)o1;")
input = "((((1),(2)),((3),(4))))'o1; f1';"
lookup = dict([(n.Name, n) for n in consensus_tree.traverse(include_self=True)])
class OldPhyloNodeTests(TestCase):
"""Tests of the PhyloNode class -- these are all now methods of RangeNode."""
def setUp(self):
"""Make a couple of standard trees"""
self.t1 = DndParser('((a,(b,c)),(d,e))', RangeNode)
#selt.t1 indices: ((0,(1,2)5)6,(3,4)7)8
def test_makeIdIndex(self):
"""RangeNode makeIdIndex should assign ids to every node"""
self.t1.makeIdIndex()
result = self.t1.IdIndex
nodes = list(self.t1.traverse(self_before=True))
#check we got an entry for each node
self.assertEqual(len(result), len(nodes))
#check the ids are in the result
for i in nodes:
assert hasattr(i, 'Id')
assert i.Id in result
def test_assignQ_single_passed(self):
"""RangeNode assignQ should propagate single Q param down tree"""
#should work if Q explicitly passed
t = self.t1
Q = ['a']
t.assignQ(Q)
for node in t.traverse(self_before=True):
assert node.Q is Q
def test_assignQ_single_set(self):
"""RangeNode assignQ should propagate single Q if set"""
t = self.t1
Q = ['a']
assert not hasattr(t, 'Q')
t.Q = Q
t.assignQ()
for node in t.traverse(self_before=True):
assert node.Q is Q
def test_assignQ_single_overwrite(self):
"""RangeNode assignQ should overwrite root Q if new Q passed"""
t = self.t1
Q = ['a']
Q2 = ['b']
t.Q = Q
t.assignQ(Q2)
for node in t.traverse(self_before=True):
assert node.Q is Q2
assert not node.Q is Q
def test_assignQ_multiple(self):
"""RangeNode assignQ should propagate multiple Qs"""
t = self.t1
Q1 = ['a']
Q2 = ['b']
Q3 = ['c']
t.makeIdIndex()
t.IdIndex[7].Q = Q1
t.IdIndex[5].Q = Q2
t.assignQ(Q3)
result = [i.Q for i in t.traverse(self_after=True)]
assert t.Q is Q3
self.assertEqual(result, [Q3, Q2, Q2, Q2, Q3, Q1, Q1, Q1, Q3])
def test_assignQ_multiple_overwrite(self):
"""RangeNode assignQ should allow overwrite"""
t = self.t1
Q1 = ['a']
Q2 = ['b']
Q3 = ['c']
t.makeIdIndex()
t.IdIndex[7].Q = Q1
t.IdIndex[5].Q = Q2
t.assignQ(Q3, overwrite=True)
for i in t.traverse(self_after=True):
assert i.Q is Q3
def test_assignQ_special(self):
"""RangeNode assignQ should work with special Qs"""
t = self.t1
Q1 = 'a'
Q2 = 'b'
Q3 = 'c'
t.makeIdIndex()
special = {7: Q1, 1: Q2}
#won't work if no Q at root
self.assertRaises(ValueError, t.assignQ, special_qs=special)
t.assignQ(Q3, special_qs=special)
result = [i.Q for i in t.traverse(self_after=True)]
self.assertEqual(result, ['c', 'b', 'c', 'c', 'c', 'a', 'a', 'a', 'c'])
def test_assignP(self):
"""RangeNode assignP should work when Qs set."""
t = self.t1
for i in t.traverse(self_before=True):
i.Length = random() * 0.5 #range 0 to 0.5
t.Q = Rates.random(DnaPairs)
t.assignQ()
t.assignP()
t.assignIds()
for node in t.traverse(self_after=True):
if node.Parent is not None:
self.assertFloatEqual(average(1-diag(node.P._data), axis=0), \
node.Length)
def test_assignLength(self):
"""RangeNode assignLength should set branch length"""
t = self.t1
t.assignLength(0.3)
for i in t.traverse(self_before=True):
self.assertEqual(i.Length, 0.3)
def test_evolve(self):
"""RangeNode evolve should work on a starting vector"""
t = self.t1
t.Q = Rates.random(DnaPairs)
t.assignQ()
t.assignLength(0.1)
t.assignP()
start = array([
1, 0, 2, 1, 0, 0, 2, 1, 2, 0, 1, 2, 1, 0, 2, 0, 0, 3, 0, 2, 1, 0,
3, 1, 0, 2, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 3, 3, 3,
3, 3, 3
])
t.evolve(start)
for i in t.traverse():
self.assertEqual(len(i.Sequence), len(start))
self.assertNotEqual(i.Sequence, start)
#WARNING: Doesn't test base freqs etc. at this point, but those aren't
#really evolve()'s responsibity (tested as self.P.mutate(seq) once
#P is set, which we've already demonstrated works.)
def test_assignPs(self):
"""RangeNode assignPs should assign multiple scaled P matrices"""
t = self.t1
for i in t.traverse(self_before=True):
i.Length = random() * 0.5 #range 0 to 0.5
t.Q = Rates.random(DnaPairs)
t.assignQ()
t.assignPs([1, 0.5, 0.25])
t.assignIds()
for node in t.traverse(self_after=True):
if node.Parent is not None:
self.assertEqual(len(node.Ps), 3)
self.assertFloatEqual(average(1-diag(node.Ps[0]._data), axis=0), \
node.Length)
self.assertFloatEqual(average(1-diag(node.Ps[1]._data), axis=0), \
0.5*node.Length)
self.assertFloatEqual(average(1-diag(node.Ps[2]._data), axis=0), \
0.25*node.Length)
def test_evolveSeqs(self):
"""PhlyoNode evolveSeqs should evolve multiple sequences"""
t = self.t1
for i in t.traverse(self_before=True):
i.Length = 0.5
t.Q = Rates.random(DnaPairs)
t.assignQ()
t.assignPs([1, 1, 0.1])
t.assignIds()
orig_seqs = [array(i) for i in [randint(0,4,200), randint(0,4,200), \
randint(0,4,200)]]
t.evolveSeqs(orig_seqs)
for node in t.traverse(): #only look at leaves
if node.Parent is not None:
self.assertEqual(len(node.Sequences), 3)
for orig, new in zip(orig_seqs, node.Sequences):
self.assertEqual(len(orig), len(new))
self.assertNotEqual(orig, new)
assert sum(orig_seqs[1]!=node.Sequences[1]) > \
sum(orig_seqs[2]!=node.Sequences[2])
class OldPhyloNodeTests(TestCase):
"""Tests of the PhyloNode class -- these are all now methods of RangeNode."""
def setUp(self):
"""Make a couple of standard trees"""
self.t1 = DndParser('((a,(b,c)),(d,e))', RangeNode)
#selt.t1 indices: ((0,(1,2)5)6,(3,4)7)8
def test_makeIdIndex(self):
"""RangeNode makeIdIndex should assign ids to every node"""
self.t1.makeIdIndex()
result = self.t1.IdIndex
nodes = list(self.t1.traverse(self_before=True))
#check we got an entry for each node
self.assertEqual(len(result), len(nodes))
#check the ids are in the result
for i in nodes:
assert hasattr(i, 'Id')
assert i.Id in result
def test_assignQ_single_passed(self):
"""RangeNode assignQ should propagate single Q param down tree"""
#should work if Q explicitly passed
t = self.t1
Q = ['a']
t.assignQ(Q)
for node in t.traverse(self_before=True):
assert node.Q is Q
def test_assignQ_single_set(self):
"""RangeNode assignQ should propagate single Q if set"""
t = self.t1
Q = ['a']
assert not hasattr(t, 'Q')
t.Q = Q
t.assignQ()
for node in t.traverse(self_before=True):
assert node.Q is Q
def test_assignQ_single_overwrite(self):
"""RangeNode assignQ should overwrite root Q if new Q passed"""
t = self.t1
Q = ['a']
Q2 = ['b']
t.Q = Q
t.assignQ(Q2)
for node in t.traverse(self_before=True):
assert node.Q is Q2
assert not node.Q is Q
def test_assignQ_multiple(self):
"""RangeNode assignQ should propagate multiple Qs"""
t = self.t1
Q1 = ['a']
Q2 = ['b']
Q3 = ['c']
t.makeIdIndex()
t.IdIndex[7].Q = Q1
t.IdIndex[5].Q = Q2
t.assignQ(Q3)
result = [i.Q for i in t.traverse(self_after=True)]
assert t.Q is Q3
self.assertEqual(result, [Q3,Q2,Q2,Q2,Q3,Q1,Q1,Q1,Q3])
def test_assignQ_multiple_overwrite(self):
"""RangeNode assignQ should allow overwrite"""
t = self.t1
Q1 = ['a']
Q2 = ['b']
Q3 = ['c']
t.makeIdIndex()
t.IdIndex[7].Q = Q1
t.IdIndex[5].Q = Q2
t.assignQ(Q3, overwrite=True)
for i in t.traverse(self_after=True):
assert i.Q is Q3
def test_assignQ_special(self):
"""RangeNode assignQ should work with special Qs"""
t = self.t1
Q1 = 'a'
Q2 = 'b'
Q3 = 'c'
t.makeIdIndex()
special = {7:Q1, 1:Q2}
#won't work if no Q at root
self.assertRaises(ValueError, t.assignQ, special_qs=special)
t.assignQ(Q3, special_qs=special)
result = [i.Q for i in t.traverse(self_after=True)]
self.assertEqual(result, ['c','b','c','c','c','a','a','a','c'])
def test_assignP(self):
"""RangeNode assignP should work when Qs set."""
t = self.t1
for i in t.traverse(self_before=True):
i.Length = random() * 0.5 #range 0 to 0.5
t.Q = Rates.random(DnaPairs)
t.assignQ()
t.assignP()
t.assignIds()
for node in t.traverse(self_after=True):
if node.Parent is not None:
self.assertFloatEqual(average(1-diag(node.P._data), axis=0), \
node.Length)
def test_assignLength(self):
"""RangeNode assignLength should set branch length"""
t = self.t1
t.assignLength(0.3)
for i in t.traverse(self_before=True):
self.assertEqual(i.Length, 0.3)
def test_evolve(self):
"""RangeNode evolve should work on a starting vector"""
t = self.t1
t.Q = Rates.random(DnaPairs)
t.assignQ()
t.assignLength(0.1)
t.assignP()
start = array([1,0,2,1,0,0,2,1,2,0,1,2,1,0,2,0,0,3,0,2,1,0,3,1,0,2,0,0,0,0,0,1,1,1,1,1,2,2,2,2,2,3,3,3,3,3,3])
t.evolve(start)
for i in t.traverse():
self.assertEqual(len(i.Sequence), len(start))
self.assertNotEqual(i.Sequence, start)
#WARNING: Doesn't test base freqs etc. at this point, but those aren't
#really evolve()'s responsibity (tested as self.P.mutate(seq) once
#P is set, which we've already demonstrated works.)
def test_assignPs(self):
"""RangeNode assignPs should assign multiple scaled P matrices"""
t = self.t1
for i in t.traverse(self_before=True):
i.Length = random() * 0.5 #range 0 to 0.5
t.Q = Rates.random(DnaPairs)
t.assignQ()
t.assignPs([1, 0.5, 0.25])
t.assignIds()
for node in t.traverse(self_after=True):
if node.Parent is not None:
self.assertEqual(len(node.Ps), 3)
self.assertFloatEqual(average(1-diag(node.Ps[0]._data), axis=0), \
node.Length)
self.assertFloatEqual(average(1-diag(node.Ps[1]._data), axis=0), \
0.5*node.Length)
self.assertFloatEqual(average(1-diag(node.Ps[2]._data), axis=0), \
0.25*node.Length)
def test_evolveSeqs(self):
"""PhlyoNode evolveSeqs should evolve multiple sequences"""
t = self.t1
for i in t.traverse(self_before=True):
i.Length = 0.5
t.Q = Rates.random(DnaPairs)
t.assignQ()
t.assignPs([1, 1, 0.1])
t.assignIds()
orig_seqs = [array(i) for i in [randint(0,4,200), randint(0,4,200), \
randint(0,4,200)]]
t.evolveSeqs(orig_seqs)
for node in t.traverse(): #only look at leaves
if node.Parent is not None:
self.assertEqual(len(node.Sequences), 3)
for orig, new in zip(orig_seqs, node.Sequences):
self.assertEqual(len(orig), len(new))
self.assertNotEqual(orig, new)
assert sum(orig_seqs[1]!=node.Sequences[1]) > \
sum(orig_seqs[2]!=node.Sequences[2])
