def merge_results(self):
assert isinstance(self.root_problem,SeppProblem)
'''Generate single extended alignment'''
fullExtendedAlignment = ExtendedAlignment(self.root_problem.fragments.keys())
#self.root_problem.get_children()[0].jobs[get_placement_job_name(0)].get_attribute("full_extended_alignment_object")
for pp in self.root_problem.get_children():
for i in range(0,self.root_problem.fragment_chunks):
align_input = open(pp.jobs[get_placement_job_name(i)].full_extended_alignment_file,'rb')
extended_alignment = pickle.load(align_input)
align_input.close()
fullExtendedAlignment.merge_in(extended_alignment,convert_to_string=True)
self.results = fullExtendedAlignment
mergeinput = []
'''Append main tree to merge input'''
mergeinput.append("%s;" %(self.root_problem.subtree.compose_newick(labels = True)))
jsons = []
for pp in self.root_problem.get_children():
assert isinstance(pp,SeppProblem)
for i in range(0,self.root_problem.fragment_chunks):
if (pp.get_job_result_by_name(get_placement_job_name(i)) is None):
continue
'''Append subset trees and json locations to merge input'''
mergeinput.append("%s;\n%s" %(pp.subtree.compose_newick(labels = True),
pp.get_job_result_by_name(get_placement_job_name(i))))
mergeinput.append("")
mergeinput.append("")
meregeinputstring = "\n".join(mergeinput)
mergeJsonJob = self.get_merge_job(meregeinputstring)
mergeJsonJob.run()
def build_jobs(self):
assert isinstance(self.root_problem, SeppProblem)
for placement_problem in self.root_problem.get_children():
''' Create placer jobs'''
for i in range(0, self.root_problem.fragment_chunks):
if self.placer == "pplacer":
pj = PplacerJob()
pj.partial_setup_for_subproblem(placement_problem,
self.options.info_file, i)
elif self.placer == "epa":
raise ValueError("EPA Currently not supported")
# pj = EPAJob()
# pj.partial_setup_for_subproblem(
# placement_problem, self.molecule, i)
placement_problem.add_job(get_placement_job_name(i), pj)
'''For each alignment subproblem, ...'''
for alg_problem in placement_problem.children:
assert isinstance(alg_problem, SeppProblem)
''' create the build model job'''
bj = HMMBuildJob()
bj.setup_for_subproblem(alg_problem, molecule=self.molecule)
alg_problem.add_job(bj.job_type, bj)
''' create the search jobs'''
for fc_problem in alg_problem.get_children():
sj = HMMSearchJob()
sj.partial_setup_for_subproblem(fc_problem.fragments,
fc_problem, self.elim,
self.filters)
fc_problem.add_job(sj.job_type, sj)
''' create the align job'''
aj = HMMAlignJob()
fc_problem.add_job(aj.job_type, aj)
aj.partial_setup_for_subproblem(fc_problem,
molecule=self.molecule)
def merge_results(self):
assert isinstance(self.root_problem, SeppProblem)
'''Generate single extended alignment'''
fullExtendedAlignment = ExtendedAlignment(
self.root_problem.fragments.keys())
# self.root_problem.get_children()[0].jobs[get_placement_job_name(0)]\
# .get_attribute("full_extended_alignment_object")
for pp in self.root_problem.get_children():
for i in range(0, self.root_problem.fragment_chunks):
align_input = open(
pp.jobs[get_placement_job_name(i)]
.full_extended_alignment_file, 'rb')
extended_alignment = pickle.load(align_input)
align_input.close()
fullExtendedAlignment.merge_in(
extended_alignment, convert_to_string=True)
self.results = fullExtendedAlignment
mergeinput = []
'''Append main tree to merge input'''
mergeinput.append("%s;" % (
self.root_problem.subtree.compose_newick(labels=True)))
for pp in self.root_problem.get_children():
assert isinstance(pp, SeppProblem)
for i in range(0, self.root_problem.fragment_chunks):
if (pp.get_job_result_by_name(
get_placement_job_name(i)) is None):
continue
'''Append subset trees and json locations to merge input'''
mergeinput.append(
"%s;\n%s" % (
pp.subtree.compose_newick(labels=True),
pp.get_job_result_by_name(get_placement_job_name(i))))
mergeinput.append("")
mergeinput.append("")
meregeinputstring = "\n".join(mergeinput)
mergeJsonJob = self.get_merge_job(meregeinputstring)
mergeJsonJob.run()
def perform(self):
pp = self.placement_problem
fullExtendedAlignments = self.merge_subalignments()
for i in range(0, self.root_problem.fragment_chunks):
fullExtendedAlignment = fullExtendedAlignments[i]
# Split the backbone alignment and query sequences into
# separate files
queryExtendedAlignment = \
fullExtendedAlignment.get_fragments_readonly_alignment()
base_alignment = fullExtendedAlignment.\
get_base_readonly_alignment()
pj = pp.jobs[get_placement_job_name(i)]
if queryExtendedAlignment.is_empty():
pj.fake_run = True
if self.placer == "pplacer":
assert isinstance(pj, PplacerJob)
# Write out the extended alignments, split into query and
# full-length for pplacer
queryExtendedAlignment.write_to_path(
pj.extended_alignment_file)
base_alignment.write_to_path(pj.backbone_alignment_file)
elif self.placer == "epa":
# assert isinstance(pj, EPAJob)
raise ValueError("EPA Currently not supported")
# Write out the extended alignments in phylip for EPA
# fullExtendedAlignment.write_to_path(
# pj.extended_alignment_file, schema="PHYLIP")
# keep the extended alignment on everything
# pj.set_attribute("full_extended_alignment_object",
# fullExtendedAlignment)
# TODO: Removed this, as it can cause unexpected lockups
output = open(pj.full_extended_alignment_file, 'wb')
pickle.dump(fullExtendedAlignment, output)
output.close()
# Enqueue the placement job
JobPool().enqueue_job(pj)
def perform(self):
pp = self.placement_problem
fullExtendedAlignments = self.merge_subalignments()
for i in range(0, self.root_problem.fragment_chunks):
fullExtendedAlignment = fullExtendedAlignments[i]
# Split the backbone alignment and query sequences into
# separate files
queryExtendedAlignment = \
fullExtendedAlignment.get_fragments_readonly_alignment()
baseAlignment = fullExtendedAlignment.get_base_readonly_alignment()
pj = pp.jobs[get_placement_job_name(i)]
if (queryExtendedAlignment.is_empty()):
pj.fake_run = True
if self.placer == "pplacer":
assert isinstance(pj, PplacerJob)
# Write out the extended alignments, split into query and
# full-length for pplacer
queryExtendedAlignment.write_to_path(
pj.extended_alignment_file)
baseAlignment.write_to_path(pj.backbone_alignment_file)
elif self.placer == "epa":
# assert isinstance(pj, EPAJob)
raise ValueError("EPA Currently not supported")
# Write out the extended alignments in phylip for EPA
fullExtendedAlignment.write_to_path(
pj.extended_alignment_file, schema="PHYLIP")
# keep the extended alignment on everything
# pj.set_attribute("full_extended_alignment_object",
# fullExtendedAlignment)
# TODO: Removed this, as it can cause unexpected lockups
output = open(pj.full_extended_alignment_file, 'wb')
pickle.dump(fullExtendedAlignment, output)
output.close()
# Enqueue the placement job
JobPool().enqueue_job(pj)
def build_jobs(self):
assert isinstance(self.root_problem, SeppProblem)
for placement_problem in self.root_problem.get_children():
''' Create placer jobs'''
for i in range(0, self.root_problem.fragment_chunks):
if self.placer == "pplacer":
pj = PplacerJob()
pj.partial_setup_for_subproblem(
placement_problem, self.options.info_file, i)
elif self.placer == "epa":
raise ValueError("EPA Currently not supported")
# pj = EPAJob()
# pj.partial_setup_for_subproblem(
# placement_problem, self.molecule, i)
placement_problem.add_job(get_placement_job_name(i), pj)
'''For each alignment subproblem, ...'''
for alg_problem in placement_problem.children:
assert isinstance(alg_problem, SeppProblem)
''' create the build model job'''
bj = HMMBuildJob()
bj.setup_for_subproblem(alg_problem, molecule=self.molecule)
alg_problem.add_job(bj.job_type, bj)
''' create the search jobs'''
for fc_problem in alg_problem.get_children():
sj = HMMSearchJob()
sj.partial_setup_for_subproblem(
fc_problem.fragments, fc_problem, self.elim,
self.filters)
fc_problem.add_job(sj.job_type, sj)
''' create the align job'''
aj = HMMAlignJob()
fc_problem.add_job(aj.job_type, aj)
aj.partial_setup_for_subproblem(
fc_problem, molecule=self.molecule)