def testAddTaskSet(self):
ts = easy.TaskSet()
ts.add(self.task)
#
tsg = tasksetgraph.TaskSetGraph()
tsindex_id = tsg.add_taskset(ts)
dict_retrieve = tsg.digraph.node[tasksetgraph.get_tasksetid(
(self.task, ))]
ts_retrieve = dict_retrieve['taskset']
self.assertEqual(len(ts), len(ts_retrieve))
self.assertEqual(tuple(x.task_id for x in ts),
tuple(x.task_id for x in ts_retrieve))
# add connected tasks
taskset_align = easy.TaskSet()
aligner = rnaseq.Bowtie2()
for fq_name in ('foo.fq', 'bar.fq'):
Assets = aligner.Assets
assets = Assets(
Assets.Source(self.task.call.assets.target.indexfilepattern,
rnaseq.FASTQPossiblyGzipCompressed(fq_name),
None), None)
task_align = self.project.add_task(aligner, assets)
taskset_align.add(task_align)
tsalign_id = tsg.add_taskset(taskset_align)
lst = tsg.execution_list()
self.assertEqual(tsindex_id,
tasksetgraph.get_tasksetid(lst[0].taskset))
self.assertEqual(tsalign_id,
tasksetgraph.get_tasksetid(lst[1].taskset))
def testExecutionList(self):
tsg = tasksetgraph.TaskSetGraph()
# initial task/taskset is building a bowtie2 index
tsindex_id = tsg.add(self.task)
lst = tsg.execution_list()
self.assertEqual(1, len(lst))
# add a second task set that performs the alignment of several pairs of
# FASTQ files
n_fastq = 3
pairnames = self._create_fastq(n_fastq, self.wd2)
index_task = self.task
bowtie2 = rnaseq.Bowtie2()
Assets = bowtie2.Assets
# Class to model FASTQ files that are optionally compressed
FASTQ = rnaseq.FASTQPossiblyGzipCompressed
ts = easy.TaskSet()
# note that we are included the pair of FASTQ already aligned above
for read1_fn, read2_fn in pairnames:
task = self.project.add_task(
bowtie2,
Assets(
Assets.Source(
index_task.call.assets.target.indexfilepattern,
FASTQ(read1_fn), FASTQ(read2_fn))))
ts.add(task)
tsalign_id = tsg.add(ts)
lst = tsg.execution_list()
def testAddTask(self):
tsg = tasksetgraph.TaskSetGraph()
tsg.add(self.task)
dict_retrieve = tsg.digraph.node[tasksetgraph.get_tasksetid(
(self.task, ))]
ts_retrieve = dict_retrieve['taskset']
self.assertEqual(1, len(ts_retrieve))
self.assertEqual((self.task.task_id, ),
tuple(x.task_id for x in ts_retrieve))
def testAddTaskSetDuplicateTask(self):
ts = easy.TaskSet()
ts.add(self.task)
tsg = tasksetgraph.TaskSetGraph()
tsg.add_taskset(ts)
ts2 = easy.TaskSet()
ts2.add(self.task)
self.assertRaises(ValueError, tsg.add_taskset, ts2)
def testAddTaskDifferentProject(self):
project2 = easy.Project(rnaseq, self.wd2)
bowtie2index = rnaseq.Bowtie2Build()
reference_fn = PHAGEFASTA
Assets = bowtie2index.Assets
task2 = project2.add_task(
bowtie2index,
Assets(Assets.Source(rnaseq.FASTAFile(reference_fn))))
#
tsg = tasksetgraph.TaskSetGraph()
tsg.add(self.task)
self.assertRaises(ValueError, tsg.add, task2)
def testExecution(self):
tsg = tasksetgraph.TaskSetGraph()
tsg.add(self.task)
n_fastq = 3
pairnames = self._create_fastq(n_fastq, self.wd2)
index_task = self.task
bowtie2 = rnaseq.Bowtie2()
Assets = bowtie2.Assets
FASTQ = rnaseq.FASTQPossiblyGzipCompressed
ts = easy.TaskSet()
for read1_fn, read2_fn in pairnames:
task = self.project.add_task(
bowtie2,
Assets(
Assets.Source(
index_task.call.assets.target.indexfilepattern,
FASTQ(read1_fn), FASTQ(read2_fn))))
ts.add(task)
tsg.add(ts)
# no task mapper, check that an exception is raised
self.assertRaises(ValueError, tsg.execute)
# set the default task mapper to a multiprocessing based one
n_processes = 1
mpe = easy.MultiprocessingExecution(n_processes)
tsg.defaultmapper = mpe
tsg.execute()
# use a run-only-once filter (that is only try to execute
# tasks with the status "TO DO"
flt = lambda taskset: taskset.filter_on_status(easy._TASK_TODO)
tsg.defaulttasksetfilter = flt
lst = tsg.execution_list()
for elt in lst:
ts_f = tsg._filtered_taskset(elt)
# check that nothing is left to be exectuted
# (since everything has been run earlier)
self.assertEqual(0, len(ts_f))