def get_recurrence_go_no_rec_array_cluster_id_ls(self, curs, pattern_table, mcl_id2go_no_set):
"""
2006-09-26
from pattern_table and use mcl_id2go_no_set
go_no_list is the go_id Set
mcl_id2enc_recurrence is for get_recurrence_rec_array_bs_no_list()
"""
sys.stderr.write("Getting recurrence_go_no_rec_array_cluster_id_ls...\n")
no_of_datasets = 0
go_no2recurrence_cluster_id = {}
mcl_id2enc_recurrence = {}
curs.execute("DECLARE crs CURSOR FOR SELECT id, recurrence_array from %s"\
%pattern_table)
curs.execute("fetch 5000 from crs")
rows = curs.fetchall()
counter = 0
real_counter = 0
while rows:
for row in rows:
mcl_id, recurrence_array = row
if mcl_id in mcl_id2go_no_set:
#if this pattern has functions predicted
recurrence_array = recurrence_array[1:-1].split(',')
recurrence_array = map(float, recurrence_array) #this is not a binary 0/1 array
occurrence_cutoff_func = lambda x: int(x>=0.8) #0.8 is arbitrary
recurrence_array = map(occurrence_cutoff_func, recurrence_array)
if no_of_datasets == 0:
no_of_datasets = len(recurrence_array)
go_no_list = mcl_id2go_no_set[mcl_id]
encoded_recurrence = encodeOccurrenceBv(recurrence_array)
mcl_id2enc_recurrence[mcl_id] = encoded_recurrence #2006-09-26
for go_no in go_no_list:
if go_no not in go_no2recurrence_cluster_id:
go_no2recurrence_cluster_id[go_no] = [encoded_recurrence, Set([mcl_id])]
#use Set() because mcl_id has duplicates due to different p-values
else:
go_no2recurrence_cluster_id[go_no][0] = \
go_no2recurrence_cluster_id[go_no][0] | encoded_recurrence
go_no2recurrence_cluster_id[go_no][1].add(mcl_id)
real_counter += 1
counter += 1
if self.report:
sys.stderr.write("%s%s\t%s"%('\x08'*20, counter, real_counter))
curs.execute("fetch 5000 from crs")
rows = curs.fetchall()
curs.execute("close crs")
recurrence_go_no_rec_array_cluster_id_ls = []
for go_no in go_no2recurrence_cluster_id:
encoded_recurrence, mcl_id_set = go_no2recurrence_cluster_id[go_no]
recurrence_array = decodeOccurrence(encoded_recurrence) #not binary vector
recurrence = len(recurrence_array)
recurrence_go_no_rec_array_cluster_id_ls.append([recurrence, go_no, recurrence_array, mcl_id_set])
recurrence_go_no_rec_array_cluster_id_ls.sort()
sys.stderr.write("End getting recurrence_go_no_rec_array_cluster_id_ls.\n")
return recurrence_go_no_rec_array_cluster_id_ls, no_of_datasets, mcl_id2enc_recurrence
def get_recurrence_rec_array_bs_no_list(self, curs, cluster_bs_table, mcl_id2enc_recurrence):
"""
11-01-05
"""
sys.stderr.write("Getting recurrence_rec_array_bs_no_list...\n")
bs_no2enc_recurrence = {}
curs.execute("DECLARE crs CURSOR FOR select c.mcl_id, c.bs_no_list from %s c"%(cluster_bs_table))
curs.execute("fetch 5000 from crs")
rows = curs.fetchall()
counter = 0
real_counter =0
while rows:
for row in rows:
mcl_id, bs_no_list = row
if mcl_id in mcl_id2enc_recurrence:
encoded_recurrence = mcl_id2enc_recurrence[mcl_id]
bs_no_list = bs_no_list[1:-1].split(',')
bs_no_list = map(int, bs_no_list)
for bs_no in bs_no_list:
if bs_no not in bs_no2enc_recurrence:
bs_no2enc_recurrence[bs_no] = encoded_recurrence
else:
bs_no2enc_recurrence[bs_no] |= encoded_recurrence
real_counter += 1
counter += 1
if self.report:
sys.stderr.write("%s%s/%s"%('\x08'*20, counter, real_counter))
curs.execute("fetch 10000 from crs")
rows = curs.fetchall()
curs.execute("close crs")
recurrence_rec_array_bs_no_list = []
for bs_no, enc_recurrence in bs_no2enc_recurrence.iteritems():
recurrence_array = decodeOccurrence(enc_recurrence) #not binary vector
recurrence = len(recurrence_array)
recurrence_rec_array_bs_no_list.append([recurrence, recurrence_array, bs_no])
recurrence_rec_array_bs_no_list.sort()
sys.stderr.write("End getting recurrence_rec_array_bs_no_list.\n")
return recurrence_rec_array_bs_no_list
def get_core_vertex_set(self, vertex_list, recurrence_array, degree_cut_off):
"""
12-16-05
global structures used:
--self.edge2encodedOccurrence
12-18-05
expand to all datasets
--init_graph_from_vertex_set()
--decodeOccurrence()
--remove_loose_part_of_graph()
--remove_singleton_vertices()
--get_vertex_min_degree()
"""
no_of_datasets = len(recurrence_array)
#initialize all graphs
graph_list = [None]*no_of_datasets
anti_vertex_id_list = [None]*no_of_datasets
recurrence_set = Set()
for i in range(no_of_datasets):
graph_list[i], anti_vertex_id_list[i] = self.init_graph_from_vertex_set(vertex_list)
if recurrence_array[i] == 1:
recurrence_set.add(i)
no_of_vertices = len(vertex_list)
#vertex_list.sort() #presorted
#construct graphs for each 'on' dataset
for i in range(no_of_vertices):
for j in range(i+1, no_of_vertices):
edge_tuple = (vertex_list[i], vertex_list[j])
"""
if self.debug:
print "checking", edge_tuple
"""
if edge_tuple in self.edge2encodedOccurrence:
edge_recurrence = decodeOccurrence(self.edge2encodedOccurrence[edge_tuple]) #starting from 1
"""
if self.debug:
print "edge_recurrence", edge_recurrence
"""
for k in edge_recurrence:
index = k-1
v_descriptor1 = anti_vertex_id_list[index][vertex_list[i]]
v_descriptor2 = anti_vertex_id_list[index][vertex_list[j]]
graph_list[index].add_edge(v_descriptor1, v_descriptor2)
#remove loose part for each graph
on_dataset_index_ls = [0]*no_of_datasets
for i in range(no_of_datasets):
if graph_list[i].num_edges()>1: #at least the graph has two edges
degree_percentage = self.remove_loose_part_of_graph(graph_list[i], degree_cut_off)
if graph_list[i].num_vertices()>=4: #min graph size
"""
if self.debug:
print "graph %s has %s vertices remaining with degree_percentage: %s."%(i, graph_list[i].num_vertices(), degree_percentage)
"""
on_dataset_index_ls[i] = 1 #this dataset should be counted as 'on'
else:
on_dataset_index_ls[i] = 0
#find core vertex_list only in those recurrent 'on' datasets
vertex_id2occurrence = {}
recurrent_and_on_datasets_ls = []
on_but_not_recurrent_dataset2vertex_set = {}
for i in range(no_of_datasets):
if recurrence_array[i] == 1 and on_dataset_index_ls[i] == 1:
recurrent_and_on_datasets_ls.append(i)
for v in graph_list[i].vertices:
vertex_id = graph_list[i].vertex_properties['vertex_id'][v]
if vertex_id not in vertex_id2occurrence:
vertex_id2occurrence[vertex_id] = 0
vertex_id2occurrence[vertex_id] += 1
if recurrence_array[i] == 0 and on_dataset_index_ls[i] == 1:
on_but_not_recurrent_dataset2vertex_set[i] = Set()
for v in graph_list[i].vertices:
vertex_id = graph_list[i].vertex_properties['vertex_id'][v]
on_but_not_recurrent_dataset2vertex_set[i].add(vertex_id)
#only vertices in recurrent and 'on' datasets go into core_vertex_set
core_vertex_set = Set()
for vertex_id in vertex_id2occurrence:
if vertex_id2occurrence[vertex_id] == len(recurrent_and_on_datasets_ls):
core_vertex_set.add(vertex_id)
#find other on datasets from on_but_not_recurrent_dataset2vertex_set
for dataset_no, vertex_set in on_but_not_recurrent_dataset2vertex_set.iteritems():
intersection_set = core_vertex_set & vertex_set
if len(intersection_set)==len(core_vertex_set):
recurrent_and_on_datasets_ls.append(dataset_no)
core_vertex_ls = list(core_vertex_set)
core_vertex_ls.sort()
recurrent_and_on_datasets_ls.sort()
return core_vertex_ls, recurrent_and_on_datasets_ls