def depends_on(self, jobs):
for job in ppg.util.flatten_jobs(jobs):
if not hasattr(job, "is_prebuild") or not job.is_prebuild:
raise ppg.JobContractError(
"%s depended on a non-prebuild dependency %s - not supported"
% (self, job))
ppg.Job.depends_on(self, job)
return self
def align(self, aligner, genome, aligner_parameters, name=None):
from .lanes import AlignedSample
output_dir = (
Path("results")
/ "aligned"
/ ("%s_%s" % (aligner.name, aligner.version))
/ genome.name
/ self.name
)
output_dir.mkdir(parents=True, exist_ok=True)
output_filename = output_dir / (self.name + ".bam")
input_job = self.prepare_input()
index_job = genome.build_index(aligner)
alignment_job = aligner.align_job(
input_job.filenames[0],
input_job.filenames[1] if self.is_paired else None,
index_job.output_path
if hasattr(index_job, "output_path")
else index_job.filenames[0],
output_filename,
aligner_parameters if aligner_parameters else {},
)
alignment_job.depends_on(
input_job,
index_job,
# ppg.ParameterInvariant(output_filename, aligner_parameters), # that's the aligner's job.
)
for j in alignment_job.prerequisites:
if isinstance(j, ppg.ParameterInvariant):
break
else:
raise ppg.JobContractError(
"aligner (%s).align_job should have added a parameter invariant for aligner parameters"
% aligner
)
return AlignedSample(
f"{self.name if name is None else name}_{aligner.name}",
alignment_job,
genome,
self.is_paired,
self.vid,
output_dir,
aligner=aligner,
)
def add_annotator(self, anno):
if ppg.util.global_pipegraph.running:
raise ppg.JobContractError(
"Can not add_annotator in a running pipegraph"
" - the annotator structure get's fixed when a "
"pipegraph is run, you can't add to it in e.g. a "
"JobGeneratingJob")
cache_name = anno.get_cache_name()
forbidden_chars = "/", "?", "*"
if any((x in cache_name
for x in forbidden_chars)) or len(cache_name) > 60:
raise ValueError(
"annotator.column_names[0] not suitable as a cache_name (was %s), add cache_name property"
% repr(cache_name))
if not cache_name in self.ddf.annotators:
# if not hasattr(anno, "columns"): # handled by get_cache_name
# raise AttributeError("no columns property on annotator %s" % repr(anno))
self.ddf.annotators[cache_name] = anno
self.ddf.anno_jobs[cache_name] = self.get_anno_dependency_callback(
anno)
for c in self.ddf.children:
c += anno
return self.ddf.anno_jobs[cache_name]