def testPhylogenyFromNewick(self):
"""
Tries to load a newick tree from file.
"""
phylo_file = os.path.dirname(os.path.realpath(__file__)) + '/example_data/Asp_protease_2.nhx'
phylo_graph = BioBayesGraph()
graph = phylo_graph.populate_from_newick(phylo_file)
self.assertTrue(("name" in graph.vertex_properties),
msg="Clade names not imported correctly.")
self.assertTrue(("branch_length" in graph.edge_properties),
msg="Branch lengths not imported correctly.")
bl_sum = 0.0
for e in graph.edges():
bl_sum += float(graph.edge_properties['branch_length'][e])
self.assertTrue(expr=abs(bl_sum - 168.58699) < 1e-6,
msg="Branch lengths not imported correctly "\
+"(sum is wrong)")
self.assertEqual(graph.num_vertices(), 608,
"Didn't get expected number of nodes from phylogeny.")
self.assertEqual(graph.num_edges(), 607,
"Didn't get expected number of nodes from phylogeny.")
class SIFTER(object):
def __init__(self):
"""
Constructor
"""
self.phylo_graph = BioBayesGraph()
self.evidence_processors = {}
def load_phylogeny(self, phylo_file, phylo_format='phyloxml'):
"""
"""
if phylo_format == 'phyloxml':
self.phylo_graph.populate_from_phyloxml(phylo_file)
elif phylo_format == 'newick':
self.phylo_graph.populate_from_newick(phylo_file)
else:
raise Exception, "Phylo format requested isn't supported."
def load_evidence_processor(self, evidence_type,
evidence_processor_class, processor_settings):
'''
Loads evidence processor to internal reference
'''
self.evidence_processors[evidence_type] = evidence_processor_class(processor_settings)
def parse_evidence(self, evidence_type, evidence_file,
evidence_constraints, evidence_format):
"""
"""
if not evidence_type in self.evidence_processors:
raise Exception, "Evidence type requested doesn't have a handler."
return self.evidence_processors[evidence_type].parse_evidence(\
evidence_file=evidence_file,
evidence_format=evidence_format,
evidence_constraints=evidence_constraints)
def setup_nodes(self, node_to_fcn_model_map):
'''
Node to fcn_model_map is a function mapping
"vertex_id" to {
'auxiliary_info':{'num_functions':num_fcns,
'max_num_simul':3},
'prob_dist_class':prob_dist_class
}
E.g.
node_to_fcn_model_map = \
lambda v: { \
'auxiliary_info':{'num_functions':num_fcns,
'max_num_simul':3},
'prob_dist_class':'FunctionModels.Sifter2.FunctionModel'
}
'''
dist_fcn_classes = {}
for n in self.phylo_graph.g.vertices():
node_index = int(n)
fcn_model_info = node_to_fcn_model_map(node_index)
# Store auxiliary info by node
self.phylo_graph.set_node_auxiliary_information(node_index=node_index,
auxiliary_info=fcn_model_info['auxiliary_info'])
# Make an instance of the custom prob dist function
dist_model = fcn_model_info['prob_dist_class']
if dist_model.__name__ not in dist_fcn_classes:
self.phylo_graph.add_prob_dist(prob_dist_class=dist_model)
dist_fcn_classes[dist_model.__name__] = dist_model(None,None,None,None)
dist_inst = dist_fcn_classes[dist_model.__name__]
# Query the number of variables from the custom function
self.phylo_graph.set_node_variable_count(node_index=node_index,
num_vars=1)
# Query the domain of each variable from the custom function
protein_states = [f for f in dist_inst.possible_protein_states(\
fcn_variants_cnt=fcn_model_info['auxiliary_info']['num_functions'],
max_fcn_cnt=fcn_model_info['auxiliary_info']['max_num_simul'])]
self.phylo_graph.set_node_variable_domains(node_index=node_index,
var_domains=[protein_states])
# Store the distribution function in the graph for the node.
self.phylo_graph.set_node_probability_dist(node_index=node_index,
prob_dist_class=fcn_model_info['prob_dist_class'].__name__)
def process_evidence(self, evidence_type, evidence_set, evidence_constraints):
'''
Incorporates evidence into the graph using the
appropriate processor
'''
if not evidence_type in self.evidence_processors:
raise Exception, "Evidence type requested doesn't have a handler."
return self.evidence_processors[evidence_type].process_evidence(\
evidence_set=evidence_set,
evidence_constraints=evidence_constraints)
