def run(self): """ 10-07-05 10-09-05 input_node() add mcl_table 10-24-05 create new views for splat_table and mcl_table 10-28-05 no views, no new pattern_table, read from inputfile, write to outputfile 01-24-06 copy a whole block from MpiFromDatasetSignatureToPattern.py to read in edge sig matrix (rank==0) --get_no_of_datasets() --sendEdgeSigMatrix() elif free_computing_nodes: --PostFim() --receiveEdgeSigMatrix() mpi_synchronize() --input_node() --input_handler() --computing_node() --node_fire() --cleanup_handler() --output_node() --output_handler() """ communicator = MPI.world.duplicate() node_rank = communicator.rank free_computing_nodes = range(1,communicator.size-1) #exclude the last node #01-24-06 following block is directly copied from MpiFromDatasetSignatureToPattern.py block_size = 10000 MpiFromDatasetSignatureToPattern_instance = MpiFromDatasetSignatureToPattern() if communicator.rank == 0: no_of_datasets = MpiFromDatasetSignatureToPattern_instance.get_no_of_datasets(self.sig_vector_fname) #no_of_datasets is used in fillEdgeSigMatrix() and patternFormation() for node in free_computing_nodes: communicator.send(str(no_of_datasets), node, 0) MpiFromDatasetSignatureToPattern_instance.sendEdgeSigMatrix(communicator, free_computing_nodes, self.sig_vector_fname, \ no_of_datasets, self.min_sup, self.max_sup, block_size) elif communicator.rank in free_computing_nodes: data, source, tag = communicator.receiveString(0, 0) no_of_datasets = int(data) #take the data j_instance = johnson_sp(no_of_datasets) MpiFromDatasetSignatureToPattern_instance.receiveEdgeSigMatrix(communicator, j_instance, no_of_datasets, block_size) mpi_synchronize(communicator) if node_rank == 0: inf = csv.reader(open(self.inputfile,'r'), delimiter='\t') parameter_list = [inf] input_node(communicator, parameter_list, free_computing_nodes, self.size, self.report, input_handler=self.input_handler) del inf elif node_rank in free_computing_nodes: #exclude the last node parameter_list = [j_instance, self.parser_type] computing_node(communicator, parameter_list, self.node_fire, self.cleanup_handler, self.report) elif node_rank==communicator.size-1: writer = csv.writer(open(self.outputfile, 'w'), delimiter='\t') parameter_list = [writer] output_node(communicator, free_computing_nodes, parameter_list, self.output_handler, self.report) del writer
def run(self): """ 10-22-05 """ communicator = MPI.world.duplicate() node_rank = communicator.rank free_computing_nodes = range(1,communicator.size-1) print "this is node",node_rank if node_rank == 0: (conn, curs) = db_connect(self.hostname, self.dbname, self.schema) edge2occurrrence, no_of_datasets = get_edge2occurrence(curs, self.min_sup, self.max_sup) edge2occurrrence_pickle = cPickle.dumps((edge2occurrrence, no_of_datasets), -1) for node in free_computing_nodes: #send it to the computing_node communicator.send(edge2occurrrence_pickle, node, 0) del conn, curs elif node_rank in free_computing_nodes: #exclude the last node data, source, tag = communicator.receiveString(0, 0) edge2occurrrence, no_of_datasets = cPickle.loads(data) mpi_synchronize(communicator) if node_rank == 0: inf = csv.reader(open(self.inputfile,'r'), delimiter='\t') parameter_list = [inf] input_node(communicator, parameter_list, free_computing_nodes, self.message_size, self.report, input_handler=self.input_handler) del inf elif node_rank in free_computing_nodes: parameter_list = [self.min_size, self.alpha, edge2occurrrence, no_of_datasets] computing_node(communicator, parameter_list, self.node_fire, report=self.report) elif node_rank == communicator.size-1: writer = csv.writer(open(self.outputfile, 'w'), delimiter='\t') parameter_list = [writer] output_node(communicator, free_computing_nodes, parameter_list, self.output_handler, self.report) del writer
def run(self): """ 11-16-05 --computing_handler() --is_site_confirmed() --get_no_of_mismatches_allowed() --get_no_of_mismatches_for_consensus() --is_good_consensus() --get_no_of_mismatches_for_site() """ communicator = MPI.world.duplicate() node_rank = communicator.rank free_computing_nodes = range(1,communicator.size-1) #exclude the last node if node_rank == 0: (conn, curs) = db_connect(self.hostname, self.dbname, self.schema) if self.profile_filename: mt_id_set = get_mt_id_set_from_profile(self.profile_filename) else: mt_id_set = None mt_id2sites_ls = get_mt_id2sites_ls(curs, mt_id_set) mt_id2sites_ls_pickle = cPickle.dumps(mt_id2sites_ls, -1) for node in free_computing_nodes: #send it to the computing_node communicator.send(mt_id2sites_ls_pickle, node, 0) input_files = os.listdir(self.inputdir) for i in range(len(input_files)): #attach the directory path to the files input_files[i] = os.path.join(self.inputdir, input_files[i]) #the following infomation is just header info inserted into the top of the output_file match_output_header = self.get_match_output_header(input_files[0]) communicator.send(match_output_header, communicator.size-1, 0) elif node_rank in free_computing_nodes: data, source, tag = communicator.receiveString(0, 0) mt_id2sites_ls = cPickle.loads(data) #take the data elif node_rank==communicator.size-1: outf = open(self.output_file, 'w') match_output_header, source, tag = communicator.receiveString(0, 0) outf.write(match_output_header) mpi_synchronize(communicator) if node_rank == 0: aggregated_inf = fileinput.input(input_files) parameter_list = [0, aggregated_inf] input_node(communicator, parameter_list, free_computing_nodes, self.message_size, self.report, \ input_handler=self.input_handler) del aggregated_inf elif node_rank in free_computing_nodes: parameter_list = [mt_id2sites_ls, max_mis_match_perc, min_no_of_mismatches, max_esc_length] computing_node(communicator, parameter_list, self.computing_handler, report=self.report) elif node_rank==communicator.size-1: parameter_list = [outf] output_node(communicator, free_computing_nodes, parameter_list, self.output_handler, self.report) del outf
def run(self):
communicator = MPI.world.duplicate()
node_rank = communicator.rank
free_computing_nodes = range(1,communicator.size-1) #exclude the last node
if node_rank == 0:
(conn, curs) = db_connect(self.hostname, self.dbname, self.schema)
schema_instance = form_schema_tables(self.fname, self.acc_cutoff, self.lm_bit)
gene_id2no = get_gene_id2gene_no(curs)
gene2enc_array = self.get_gene2enc_array(self.gim_inputfname, gene_id2no)
gene2enc_array_pickle = cPickle.dumps(gene2enc_array, -1)
gene_no2id = get_gene_no2gene_id(curs)
gene_no2go_no = get_gene_no2go_no(curs)
gene_no2id_pickle = cPickle.dumps(gene_no2id, -1)
gene_no2go_no_pickle = cPickle.dumps(gene_no2go_no, -1)
for node in free_computing_nodes: #send it to the computing_node
communicator.send(gene2enc_array_pickle, node, 0)
communicator.send(gene_no2id_pickle, communicator.size-1, 0)
communicator.send(gene_no2go_no_pickle, communicator.size-1, 0)
elif node_rank in free_computing_nodes:
data, source, tag = communicator.receiveString(0, 0)
gene2enc_array = cPickle.loads(data) #take the data
elif node_rank==communicator.size-1:
schema_instance = form_schema_tables(self.fname, self.acc_cutoff, self.lm_bit)
data, source, tag = communicator.receiveString(0, 0)
gene_no2id = cPickle.loads(data)
data, source, tag = communicator.receiveString(0, 0)
gene_no2go_no = cPickle.loads(data)
mpi_synchronize(communicator)
if node_rank == 0:
curs.execute("DECLARE crs CURSOR FOR SELECT p.id, p.vertex_set, p.edge_set, p.recurrence_array,\
g.go_no_list from %s p, %s g where g.mcl_id=p.id"%(schema_instance.pattern_table, schema_instance.good_cluster_table))
input_node(communicator, curs, free_computing_nodes, self.message_size, self.report)
elif node_rank in free_computing_nodes:
parameter_list = [gene2enc_array, self.dataset_signature_set, self.p_value_cut_off]
computing_node(communicator, parameter_list, self.computing_node_handler, report=self.report)
elif node_rank==communicator.size-1:
if not os.path.isdir(self.pic_output_dir):
os.makedirs(self.pic_output_dir)
cluster_info_instance = cluster_info()
ofname = os.path.join(self.pic_output_dir, '%s_p%s'%(schema_instance.good_cluster_table, self.p_value_cut_off))
writer = csv.writer(open(ofname, 'w'), delimiter='\t')
parameter_list = [self.pic_output_dir, cluster_info_instance, gene_no2id, gene_no2go_no, writer]
output_node(communicator, free_computing_nodes, parameter_list, self.output_node_handler, self.report)
del writer
def run(self): """ 11-16-05 --computing_handler() --is_site_confirmed() --get_no_of_mismatches_allowed() --get_no_of_mismatches_for_consensus() --is_good_consensus() --get_no_of_mismatches_for_site() """ communicator = MPI.world.duplicate() node_rank = communicator.rank free_computing_nodes = range(1,communicator.size-1) #exclude the last node if node_rank == 0: (conn, curs) = db_connect(self.hostname, self.dbname, self.schema) known_data = self.get_known_data(curs, self.fname, self.filter_type, self.is_correct_type, self.need_cal_hg_p_value) known_data_pickle = cPickle.dumps(known_data, -1) for node in free_computing_nodes: #send it to the computing_node communicator.send(known_data_pickle, node, 0) elif node_rank in free_computing_nodes: data, source, tag = communicator.receiveString(0, 0) known_data = cPickle.loads(data) #take the data elif node_rank==communicator.size-1: writer = csv.writer(open(self.output_file, 'w'), delimiter='\t') #write down the header writer.writerow(['rpart_cp', 'loss_matrix', 'prior_prob', 'type', 'accuracy_avg','accuracy_std', 'no_of_predictions_avg',\ 'no_of_predictions_std', 'no_of_genes_avg', 'no_of_genes_std']) mpi_synchronize(communicator) if node_rank == 0: setting_ls = self.form_setting_ls(self.rpart_cp_ls, self.loss_matrix_ls, self.prior_prob_ls) self.input_node(communicator, setting_ls, free_computing_nodes, self.report) elif node_rank in free_computing_nodes: parameter_list = [known_data, self.no_of_buckets] computing_node(communicator, parameter_list, self.computing_handler, report=self.report) elif node_rank==communicator.size-1: parameter_list = [writer] output_node(communicator, free_computing_nodes, parameter_list, self.output_handler, self.report) del writer
def run(self):
"""
09-05-05
2006-09-21 add fuzzyDense_flag
2006-11-02 add tfbs_association_type
2006-11-02 differentiate good_cluster_table as pattern_xxx or good_xxx for pattern id
--db_connect()
--get_gene_no2bs_no_block()
--construct_two_dicts()
--input_node()
--fetch_cluster_block()
--computing_node()
--node_fire()
--cluster_bs_analysis()
--create_cluster_bs_table()
--output_node()
--submit_cluster_bs_table()
"""
communicator = MPI.world.duplicate()
node_rank = communicator.rank
free_computing_nodes = range(1,communicator.size-1)
print self.tfbs_association_type
if node_rank == 0:
(conn, curs) = db_connect(self.hostname, self.dbname, self.schema)
if self.tfbs_association_type==1: #2006-11-02
gene_no2bs_no_block = self.get_gene_no2bs_no_block(curs)
elif self.tfbs_association_type==2:
gene_no2bs_no_block = get_gene_no2bs_no_block_from_expt_tf_mapping(curs)
for node in range(1, communicator.size-1): #send it to the computing_node
communicator.send(gene_no2bs_no_block, node, 0)
if self.fuzzyDense_flag: #2006-09-21 add fuzzyDense_flag
#12-18-05 get edge2encodedOccurrence
MpiCrackSplat_instance = MpiCrackSplat()
edge2encodedOccurrence = {}
min_sup = 5 #need to expose them
max_sup = 40
total_vertex_set = self.return_total_vertex_set(curs, self.good_cluster_table)
edge2encodedOccurrence, no_of_datasets = self.fill_edge2encodedOccurrence(\
self.sig_vector_fname, min_sup, max_sup, total_vertex_set)
edge2encodedOccurrence_pickle = cPickle.dumps(edge2encodedOccurrence, -1)
for node in free_computing_nodes: #send it to the computing_node
communicator.send(edge2encodedOccurrence_pickle, node, 0)
elif node_rank>0 and node_rank<communicator.size-1:
data, source, tag, count = communicator.receive(Numeric.Int, 0, 0)
gene_no2bs_no_set, bs_no2gene_no_set = self.construct_two_dicts(node_rank, data)
if self.fuzzyDense_flag: #2006-09-21
#12-18-05
data, source, tag = communicator.receiveString(0, 0)
edge2encodedOccurrence = cPickle.loads(data)
elif node_rank==communicator.size-1: #establish connection before pursuing
(conn, curs) = db_connect(self.hostname, self.dbname, self.schema)
#12-20-05 for darwin output
gene_id2symbol = get_gene_id2gene_symbol(curs, self.tax_id)
dataset_no2desc = get_dataset_no2desc(curs)
mpi_synchronize(communicator)
if node_rank == 0:
if self.good_cluster_table.find('pattern')!=-1: #2006-11-02 it's pattern_xxx table, use id as pattern_id
curs.execute("DECLARE crs CURSOR FOR select distinct id, vertex_set, recurrence_array\
from %s "%(self.good_cluster_table))
else: #2006-11-02 it's good_xxx table, use mcl_id as pattern_id
curs.execute("DECLARE crs CURSOR FOR select distinct mcl_id, vertex_set, recurrence_array\
from %s "%(self.good_cluster_table))
input_node(communicator, curs, free_computing_nodes, self.size, self.report)
curs.execute("close crs")
elif node_rank<=communicator.size-2: #exclude the last node
if self.fuzzyDense_flag: #2006-09-21
fuzzyDense_instance = fuzzyDense(edge2encodedOccurrence)
else:
fuzzyDense_instance = None
parameter_list = [gene_no2bs_no_set, bs_no2gene_no_set, self.ratio_cutoff, \
self.top_number, self.p_value_cut_off, fuzzyDense_instance, self.degree_cut_off, self.fuzzyDense_flag]
computing_node(communicator, parameter_list, self.computing_node_handler, report=self.report)
elif node_rank==communicator.size-1:
#12-20-05 comment out
if self.new_table:
self.create_cluster_bs_table(curs, self.cluster_bs_table)
parameter_list = [curs, self.cluster_bs_table]
output_node(communicator, free_computing_nodes, parameter_list, self.submit_cluster_bs_table, report=self.report)
if self.commit:
curs.execute("end")
"""
def run(self):
"""
09-05-05
10-23-05
create views from old schema
result goes to the new schema's p_gene_table
(input_node)
--db_connect()
--form_schema_tables()
--form_schema_tables()
--get_gene_no2go_no_set()
--get_go_no2depth()
(pass data to computing_node)
(computing_node)
(take data from other nodes, 0 and size-1)
(judge_node)
--gene_stat()
--db_connect()
--gene_p_map_redundancy()
(output_node)
--db_connect()
--form_schema_tables()
--form_schema_tables()
--MpiPredictionFilter()
--MpiPredictionFilter_instance.createGeneTable()
--get_go_no2edge_counter_list()(if necessary)
(pass go_no2edge_counter_list to computing_node)
(input_node)
--fetch_cluster_block()
(computing_node)
--get_no_of_unknown_genes()
--node_fire_handler()
--cleanup_handler()
--judge_node()
--gene_stat_instance.(match functions)
--output_node()
--output_node_handler()
--MpiPredictionFilter_instance.submit_to_p_gene_table()
"""
communicator = MPI.world.duplicate()
node_rank = communicator.rank
if node_rank == 0:
(conn, curs) = db_connect(self.hostname, self.dbname, self.schema)
"""
#01-02-06
old_schema_instance = form_schema_tables(self.input_fname)
new_schema_instance = form_schema_tables(self.jnput_fname)
"""
gene_no2go_no = get_gene_no2go_no_set(curs)
gene_no2go_no_pickle = cPickle.dumps(gene_no2go_no, -1) #-1 means use the highest protocol
go_no2depth = get_go_no2depth(curs)
go_no2depth_pickle = cPickle.dumps(go_no2depth, -1)
go_no2gene_no_set = get_go_no2gene_no_set(curs)
go_no2gene_no_set_pickle = cPickle.dumps(go_no2gene_no_set, -1)
for node in range(1, communicator.size-2): #send it to the computing_node
communicator.send(gene_no2go_no_pickle, node, 0)
communicator.send(go_no2depth_pickle, node, 0)
communicator.send(go_no2gene_no_set_pickle, node, 0)
elif node_rank<=communicator.size-3: #WATCH: last 2 nodes are not here.
data, source, tag = communicator.receiveString(0, 0)
gene_no2go_no = cPickle.loads(data) #take the data
data, source, tag = communicator.receiveString(0, 0)
go_no2depth = cPickle.loads(data)
data, source, tag = communicator.receiveString(0, 0)
go_no2gene_no_set = cPickle.loads(data)
data, source, tag = communicator.receiveString(communicator.size-1, 0) #from the last node
go_no2edge_counter_list = cPickle.loads(data)
#choose a functor for recurrence_array
functor_dict = {0: None,
1: lambda x: int(x>=self.recurrence_x),
2: lambda x: math.pow(x, self.recurrence_x)}
functor = functor_dict[self.recurrence_x_type]
elif node_rank == communicator.size-2: #judge node
gene_stat_instance = gene_stat(depth_cut_off=self.depth)
(conn, curs) = db_connect(self.hostname, self.dbname, self.schema)
gene_stat_instance.dstruc_loadin(curs)
from gene_p_map_redundancy import gene_p_map_redundancy
node_distance_class = gene_p_map_redundancy()
elif node_rank==communicator.size-1: #establish connection before pursuing
(conn, curs) = db_connect(self.hostname, self.dbname, self.schema)
"""
#01-02-06, input and output are all directed to files
old_schema_instance = form_schema_tables(self.input_fname)
new_schema_instance = form_schema_tables(self.jnput_fname)
MpiPredictionFilter_instance = MpiPredictionFilter()
MpiPredictionFilter_instance.view_from_table(curs, old_schema_instance.splat_table, new_schema_instance.splat_table)
MpiPredictionFilter_instance.view_from_table(curs, old_schema_instance.mcl_table, new_schema_instance.mcl_table)
MpiPredictionFilter_instance.view_from_table(curs, old_schema_instance.pattern_table, new_schema_instance.pattern_table)
if self.new_table:
MpiPredictionFilter_instance.createGeneTable(curs, new_schema_instance.p_gene_table)
"""
if self.go_no2edge_counter_list_fname:
go_no2edge_counter_list = cPickle.load(open(self.go_no2edge_counter_list_fname,'r'))
else:
if self.eg_d_type==2:
go_no2edge_counter_list = None
else:
gene_no2go_no = get_gene_no2go_no_set(curs)
go_no2edge_counter_list = get_go_no2edge_counter_list(curs, gene_no2go_no, self.edge_type2index)
go_no2edge_counter_list_pickle = cPickle.dumps(go_no2edge_counter_list, -1)
for node in range(1, communicator.size-2): #send it to the computing_node
communicator.send(go_no2edge_counter_list_pickle, node, 0)
mpi_synchronize(communicator)
free_computing_nodes = range(1,communicator.size-2) #exclude the last node
if node_rank == 0:
"""
curs.execute("DECLARE crs CURSOR FOR SELECT id, vertex_set, edge_set, no_of_edges,\
connectivity, unknown_gene_ratio, recurrence_array, d_matrix from %s"%(old_schema_instance.pattern_table))
"""
self.counter = 0 #01-02-06 counter is used as id
reader = csv.reader(open(self.input_fname, 'r'), delimiter='\t')
parameter_list = [reader]
input_node(communicator, parameter_list, free_computing_nodes, self.message_size, \
self.report, input_handler=self.input_handler)
del reader
elif node_rank in free_computing_nodes:
no_of_unknown_genes = get_no_of_unknown_genes(gene_no2go_no)
GradientScorePrediction_instance = GradientScorePrediction(gene_no2go_no, go_no2gene_no_set, go_no2depth, \
go_no2edge_counter_list, no_of_unknown_genes, self.depth, self.min_layer1_associated_genes, \
self.min_layer1_ratio, self.min_layer2_associated_genes, self.min_layer2_ratio, self.exponent, \
self.score_list, self.max_layer, self.norm_exp, self.eg_d_type, self.debug)
parameter_list = [GradientScorePrediction_instance, functor]
computing_node(communicator, parameter_list, self.node_fire_handler, self.cleanup_handler, self.report)
elif node_rank == communicator.size-2:
self.judge_node(communicator, curs, gene_stat_instance, node_distance_class)
elif node_rank==communicator.size-1:
#01-02-06 output goes to plain file, not database
writer = csv.writer(open(self.jnput_fname, 'w'), delimiter='\t')
parameter_list = [writer]
output_node(communicator, free_computing_nodes, parameter_list, self.output_node_handler, self.report)
del writer
def run(self):
"""
11-16-05
11-19-05
use no_of_validations to multiply the setting(separate the one setting's validations
to different nodes)
the extra setting copy is for a non-validation real model fitting
--computing_handler()
--is_site_confirmed()
--get_no_of_mismatches_allowed()
--get_no_of_mismatches_for_consensus()
--is_good_consensus()
--get_no_of_mismatches_for_site()
"""
communicator = MPI.world.duplicate()
node_rank = communicator.rank
free_computing_nodes = range(1,communicator.size-1) #exclude the last node
if node_rank == 0:
(conn, curs) = db_connect(self.hostname, self.dbname, self.schema)
unknown_data, known_data = self.get_data(curs, self.fname, self.filter_type, self.is_correct_type, self.need_cal_hg_p_value)
known_data_pickle = cPickle.dumps(known_data, -1)
for node in free_computing_nodes: #send it to the computing_node
communicator.send(known_data_pickle, node, 0)
unknown_data_pickle = cPickle.dumps(unknown_data, -1)
for node in free_computing_nodes: #send it to the computing_node
communicator.send(unknown_data_pickle, node, 0)
elif node_rank in free_computing_nodes:
data, source, tag = communicator.receiveString(0, 0)
known_data = cPickle.loads(data) #take the data
"""
#11-19-05 shuffle data to check
index_ls = range(len(known_data))
random.shuffle(index_ls)
for i in range(len(index_ls)):
index_ls[i] = known_data[i]
known_data = index_ls
"""
data, source, tag = communicator.receiveString(0, 0)
unknown_data = cPickle.loads(data) #take the data
"""
#11-19-05 shuffle data to check
index_ls = range(len(unknown_data))
random.shuffle(index_ls)
for i in range(len(index_ls)):
index_ls[i] = unknown_data[i]
unknown_data = index_ls
"""
elif node_rank==communicator.size-1:
writer = csv.writer(open(self.output_file, 'w'), delimiter='\t')
#write down the header
writer.writerow(['rpart_cp', 'loss_matrix', 'prior_prob', 'type', 'accuracy_avg','accuracy_std', 'no_of_predictions_avg',\
'no_of_predictions_std', 'no_of_genes_avg', 'no_of_genes_std'])
mpi_synchronize(communicator)
if node_rank == 0:
if self.type==1:
setting_ls = self.form_setting_ls(self.rpart_cp_ls, self.loss_matrix_ls, self.prior_prob_ls, self.no_of_validations)
elif self.type==2:
#randomForest replaces rpart_cp_ls with mty_ls, others are ignored later
setting_ls = self.form_setting_ls(self.mty_ls, self.loss_matrix_ls, self.prior_prob_ls, self.no_of_validations)
else:
sys.stderr.write("type %s not supported.\n"%self.type)
sys.exit(3)
self.input_node(communicator, setting_ls, free_computing_nodes, self.report)
elif node_rank in free_computing_nodes:
parameter_list = [unknown_data, known_data, self.training_perc, self.no_of_validations, self.type, self.bit_string] #03-17-06 add type, bit_string
computing_node(communicator, parameter_list, self.computing_handler, report=self.report)
elif node_rank==communicator.size-1:
setting2validation_stat = {}
setting2unknown_known_acc_ls = {}
parameter_list = [writer, setting2validation_stat, setting2unknown_known_acc_ls, self.no_of_validations]
output_node(communicator, free_computing_nodes, parameter_list, self.output_handler, self.report)
#cPickle.dump([setting2validation_stat, setting2unknown_known_acc_ls], open('/home/yuhuang/MpiRpartValidation.setting2result.pickle','w')) #11-23-05
del writer
def run(self):
"""
2007-04-16
(rank==0)
--get_chr_start_ls()
elif free_computing_nodes:
-- (receive data)
--mpi_synchronize()
(rank==0)
--input_node()
--input_handler()
elif free_computing_nodes:
--computing_node()
--computing_node_handler()
--identify_ancestry_with_min_jumps()
--initialize_score_trace_matrix()
--is_child_heterozygous_SNP_compatible_with_parents()
(for loop)
--identify_ancestry_of_one_chr_with_DP()
--is_child_heterozygous_SNP_compatible_with_parents()
--trace()
--recursive_trace()
else:
--output_node()
--output_node_handler()
"""
node_rank = self.communicator.rank
free_computing_nodes = range(1, self.communicator.size - 1) # exclude the 1st and last node
if node_rank == 0:
FilterStrainSNPMatrix_instance = FilterStrainSNPMatrix()
header, strain_acc_list, category_list, data_matrix = FilterStrainSNPMatrix_instance.read_data(
self.input_fname
)
snp_acc_list = header[2:]
data_matrix = Numeric.array(data_matrix)
no_of_strains = data_matrix.shape[0]
(conn, curs) = db_connect(self.hostname, self.dbname, self.schema, password="******", user="******")
# 2007-09-17 send strain_acc_list to the output_node
strain_acc_list_pickle = cPickle.dumps(strain_acc_list, -1)
self.communicator.send(strain_acc_list_pickle, self.communicator.size - 1, 0)
chr_start_ls = self.get_chr_start_ls(curs, snp_acc_list, self.snp_locus_table)
chr_start_ls_pickle = cPickle.dumps(chr_start_ls, -1) # -1 means use the highest protocol
data_matrix_pickle = cPickle.dumps(data_matrix, -1)
for node in free_computing_nodes: # send it to the computing_node
self.communicator.send(chr_start_ls_pickle, node, 0)
self.communicator.send(data_matrix_pickle, node, 0)
elif node_rank in free_computing_nodes:
data, source, tag = self.communicator.receiveString(0, 0)
chr_start_ls = cPickle.loads(data) # take the data
data, source, tag = self.communicator.receiveString(0, 0)
data_matrix = cPickle.loads(data)
else:
data, source, tag = self.communicator.receiveString(0, 0)
strain_acc_list = cPickle.loads(data)
mpi_synchronize(self.communicator)
if node_rank == 0:
parameter_list = [no_of_strains]
self.input_node(self.communicator, parameter_list, free_computing_nodes, self.message_size, self.report)
elif node_rank in free_computing_nodes:
trio_arrangement_ls = [
[0, 1, 2],
[1, 2, 0],
[2, 0, 1],
] # three different ways to pick the parent-set and the child
parameter_list = [data_matrix, chr_start_ls, trio_arrangement_ls]
computing_node(self.communicator, parameter_list, self.computing_node_handler, report=self.report)
else:
writer = csv.writer(open(self.output_fname, "w"), delimiter="\t")
parameter_list = [writer, strain_acc_list]
output_node(self.communicator, free_computing_nodes, parameter_list, self.output_node_handler, self.report)
del writer
def run(self): """ 08-06-05 08-24-05 read all edge data into matrix 08-31-05 the integer returned by encodeOccurrenceBv() could be 138-bit(human no_of_datasets) And Numeric.Int is only 32 bit. So Change edge_sig_matrix format. 12-31-05 no database connection any more 2 threads on computing node 01-08-06 no threads back to edge_sig_matrix 01-11-06 use the cc module, PostFim 01-15-06 add min_line_number and max_line_number (rank==0) --get_no_of_datasets() --sendEdgeSigMatrix() elif free_computing_nodes: --PostFim() --receiveEdgeSigMatrix() --mpi_synchronize() (rank==0) --input_node() --input_handler() elif free_computing_nodes: --computing_node() --computing_node_handler() else: --output_node() --output_node_handler() --mpi_synchronize() (rank==0) --receive node_outputfile --netmine_wrapper() --collect_and_merge_output() --uniqueSort() else: --return node_outputfile """ communicator = MPI.world.duplicate() free_computing_nodes = range(1,communicator.size-1) #exclude the 1st and last node block_size = 10000 if communicator.rank == 0: no_of_datasets = self.get_no_of_datasets(self.sig_vector_fname) #no_of_datasets is used in fillEdgeSigMatrix() and patternFormation() for node in free_computing_nodes: communicator.send(str(no_of_datasets), node, 0) self.sendEdgeSigMatrix(communicator, free_computing_nodes, self.sig_vector_fname, \ no_of_datasets, self.min_sup, self.max_sup, block_size=10000) elif communicator.rank in free_computing_nodes: data, source, tag = communicator.receiveString(0, 0) no_of_datasets = int(data) #take the data offset = communicator.rank - 1 node_outputfile = '%s.%s'%(self.outputfile, offset) PostFim_instance = PostFim(self.no_cc, no_of_datasets, self.min_cluster_size, node_outputfile) self.receiveEdgeSigMatrix(communicator, PostFim_instance, no_of_datasets, block_size) mpi_synchronize(communicator) if communicator.rank == 0: reader = csv.reader(open(self.inputfile, 'r'), delimiter=' ') parameter_list = [reader, self.min_line_number, self.max_line_number] #01-15-06 self.line_number = 0 #01-15-06 used in input_handler() input_node(communicator, parameter_list, free_computing_nodes, self.queue_size, \ self.report, input_handler=self.input_handler) del reader elif communicator.rank in free_computing_nodes: parameter_list = [PostFim_instance] computing_node(communicator, parameter_list, self.computing_node_handler, report=self.report) else: parameter_list = [] output_node(communicator, free_computing_nodes, parameter_list, self.output_node_handler, self.report) mpi_synchronize(communicator) if communicator.rank == 0: #12-31-05 wait until of_name_list is full of_name_list = [] while len(of_name_list)<len(free_computing_nodes): data, source, tag = communicator.receiveString(None, 1) of_name_list.append(data) #collecting intermediateFile = '%s.unsorted'%self.outputfile #intermediateFile to store concatenated results netmine_wrapper_instance = netmine_wrapper() netmine_wrapper_instance.collect_and_merge_output(of_name_list, intermediateFile) self.uniqueSort(intermediateFile, self.outputfile, self.tmpdir) elif communicator.rank in free_computing_nodes: communicator.send(node_outputfile, 0, 1) #send back the outputfile
def run(self):
"""
2007-04-16
(rank==0)
--get_chr_start_ls()
elif free_computing_nodes:
-- (receive data)
--mpi_synchronize()
(rank==0)
--input_node()
--input_handler()
elif free_computing_nodes:
--computing_node()
--computing_node_handler()
--identify_ancestry_with_min_jumps()
--initialize_score_trace_matrix()
--is_child_heterozygous_SNP_compatible_with_parents()
(for loop)
--identify_ancestry_of_one_chr_with_DP()
--is_child_heterozygous_SNP_compatible_with_parents()
--trace()
--recursive_trace()
else:
--output_node()
--output_node_handler()
"""
node_rank = self.communicator.rank
free_computing_nodes = range(1, self.communicator.size -
1) #exclude the 1st and last node
if node_rank == 0:
FilterStrainSNPMatrix_instance = FilterStrainSNPMatrix()
header, strain_acc_list, category_list, data_matrix = FilterStrainSNPMatrix_instance.read_data(
self.input_fname)
snp_acc_list = header[2:]
data_matrix = Numeric.array(data_matrix)
no_of_strains = data_matrix.shape[0]
(conn, curs) = db_connect(self.hostname,
self.dbname,
self.schema,
password='******',
user='******')
#2007-09-17 send strain_acc_list to the output_node
strain_acc_list_pickle = cPickle.dumps(strain_acc_list, -1)
self.communicator.send(strain_acc_list_pickle,
self.communicator.size - 1, 0)
chr_start_ls = self.get_chr_start_ls(curs, snp_acc_list,
self.snp_locus_table)
chr_start_ls_pickle = cPickle.dumps(
chr_start_ls, -1) #-1 means use the highest protocol
data_matrix_pickle = cPickle.dumps(data_matrix, -1)
for node in free_computing_nodes: #send it to the computing_node
self.communicator.send(chr_start_ls_pickle, node, 0)
self.communicator.send(data_matrix_pickle, node, 0)
elif node_rank in free_computing_nodes:
data, source, tag = self.communicator.receiveString(0, 0)
chr_start_ls = cPickle.loads(data) #take the data
data, source, tag = self.communicator.receiveString(0, 0)
data_matrix = cPickle.loads(data)
else:
data, source, tag = self.communicator.receiveString(0, 0)
strain_acc_list = cPickle.loads(data)
mpi_synchronize(self.communicator)
if node_rank == 0:
parameter_list = [no_of_strains]
self.input_node(self.communicator, parameter_list, free_computing_nodes, self.message_size, \
self.report)
elif node_rank in free_computing_nodes:
trio_arrangement_ls = [[0, 1, 2], [1, 2, 0], [
2, 0, 1
]] #three different ways to pick the parent-set and the child
parameter_list = [data_matrix, chr_start_ls, trio_arrangement_ls]
computing_node(self.communicator,
parameter_list,
self.computing_node_handler,
report=self.report)
else:
writer = csv.writer(open(self.output_fname, 'w'), delimiter='\t')
parameter_list = [writer, strain_acc_list]
output_node(self.communicator, free_computing_nodes,
parameter_list, self.output_node_handler, self.report)
del writer
def run(self):
"""
11-16-05
11-19-05
use no_of_validations to multiply the setting(separate the one setting's validations
to different nodes)
the extra setting copy is for a non-validation real model fitting
--computing_handler()
--is_site_confirmed()
--get_no_of_mismatches_allowed()
--get_no_of_mismatches_for_consensus()
--is_good_consensus()
--get_no_of_mismatches_for_site()
"""
communicator = MPI.world.duplicate()
node_rank = communicator.rank
free_computing_nodes = range(1, communicator.size -
1) #exclude the last node
if node_rank == 0:
(conn, curs) = db_connect(self.hostname, self.dbname, self.schema)
unknown_data, known_data = self.get_data(curs, self.fname,
self.filter_type,
self.is_correct_type,
self.need_cal_hg_p_value)
known_data_pickle = cPickle.dumps(known_data, -1)
for node in free_computing_nodes: #send it to the computing_node
communicator.send(known_data_pickle, node, 0)
unknown_data_pickle = cPickle.dumps(unknown_data, -1)
for node in free_computing_nodes: #send it to the computing_node
communicator.send(unknown_data_pickle, node, 0)
elif node_rank in free_computing_nodes:
data, source, tag = communicator.receiveString(0, 0)
known_data = cPickle.loads(data) #take the data
"""
#11-19-05 shuffle data to check
index_ls = range(len(known_data))
random.shuffle(index_ls)
for i in range(len(index_ls)):
index_ls[i] = known_data[i]
known_data = index_ls
"""
data, source, tag = communicator.receiveString(0, 0)
unknown_data = cPickle.loads(data) #take the data
"""
#11-19-05 shuffle data to check
index_ls = range(len(unknown_data))
random.shuffle(index_ls)
for i in range(len(index_ls)):
index_ls[i] = unknown_data[i]
unknown_data = index_ls
"""
elif node_rank == communicator.size - 1:
writer = csv.writer(open(self.output_file, 'w'), delimiter='\t')
#write down the header
writer.writerow(['rpart_cp', 'loss_matrix', 'prior_prob', 'type', 'accuracy_avg','accuracy_std', 'no_of_predictions_avg',\
'no_of_predictions_std', 'no_of_genes_avg', 'no_of_genes_std'])
mpi_synchronize(communicator)
if node_rank == 0:
if self.type == 1:
setting_ls = self.form_setting_ls(self.rpart_cp_ls,
self.loss_matrix_ls,
self.prior_prob_ls,
self.no_of_validations)
elif self.type == 2:
#randomForest replaces rpart_cp_ls with mty_ls, others are ignored later
setting_ls = self.form_setting_ls(self.mty_ls,
self.loss_matrix_ls,
self.prior_prob_ls,
self.no_of_validations)
else:
sys.stderr.write("type %s not supported.\n" % self.type)
sys.exit(3)
self.input_node(communicator, setting_ls, free_computing_nodes,
self.report)
elif node_rank in free_computing_nodes:
parameter_list = [
unknown_data, known_data, self.training_perc,
self.no_of_validations, self.type, self.bit_string
] #03-17-06 add type, bit_string
computing_node(communicator,
parameter_list,
self.computing_handler,
report=self.report)
elif node_rank == communicator.size - 1:
setting2validation_stat = {}
setting2unknown_known_acc_ls = {}
parameter_list = [
writer, setting2validation_stat, setting2unknown_known_acc_ls,
self.no_of_validations
]
output_node(communicator, free_computing_nodes, parameter_list,
self.output_handler, self.report)
#cPickle.dump([setting2validation_stat, setting2unknown_known_acc_ls], open('/home/yuhuang/MpiRpartValidation.setting2result.pickle','w')) #11-23-05
del writer