def test_search_recover(self):
# first job searching using hmmsearch
gene_name = "abc"
c_gene_abc = CoreGene(self.model_location, gene_name,
self.profile_factory)
report = search_genes([c_gene_abc], self.cfg)
expected_hit = [
Hit(c_gene_abc, "ESCO030p01_000260", 706, "ESCO030p01", 26,
float(1.000e-200), float(660.800), float(1.000), float(0.714),
160, 663)
]
# second job using recover
# disable hmmer to be sure that test use the recover inner function
self.cfg.hmmer = lambda: "hmmer_disable"
# and create a new dir for the second job
previous_job_path = self.cfg.working_dir()
self.cfg.previous_run = lambda: previous_job_path
self.cfg.out_dir = lambda: os.path.join(self.tmp_dir, 'job_2')
os.mkdir(self.cfg.out_dir())
# rerun with previous run
# but we have to reset the profile attached to the gene gene._profile._report
self.profile_factory = ProfileFactory(self.cfg)
c_gene_abc = CoreGene(self.model_location, gene_name,
self.profile_factory)
report = search_genes([c_gene_abc], self.cfg)
self.assertEqual(len(report), 1)
self.assertEqual(expected_hit[0], report[0].hits[0])
def test_search(self):
gene_name = "abc"
c_gene_abc = CoreGene(self.model_location, gene_name,
self.profile_factory)
report = search_genes([c_gene_abc], self.cfg)
expected_hit = [
Hit(c_gene_abc, "ESCO030p01_000260", 706, "ESCO030p01", 26,
float(1.000e-200), float(660.800), float(1.000), float(0.714),
160, 663)
]
self.assertEqual(len(report), 1)
self.assertEqual(expected_hit[0], report[0].hits[0])
def search_systems(config, model_bank, gene_bank, profile_factory, logger):
"""
Do the job, this function is the orchestrator of all the macsyfinder mechanics
at the end several files are produced containing the results
- macsyfinder.conf: The set of variables used to runt this job
- macsyfinder.systems: The list of the potential systems
- macsyfinder.rejected_cluster: The list of all clusters and clustrs combination
which has been rejected and the reason
- macsyfinder.log: the copy of the standard output
:param config: The MacSyFinder Configuration
:type config: :class:`macsypy.config.Config` object
:param model_bank: The bank populated with the available models
:type model_bank: :class:`macsypy.model.ModelBank` object
:param gene_bank: the bank containing all genes
:type gene_bank: :class:`macsypy.gene.GeneBank` object
:param profile_factory: The profile factory
:type profile_factory: :class:`macsypy.gene.ProfileFactory`
:param logger: The logger use to display information to the user.
It must be initialized. see :func:`macsypy.init_logger`
:type logger: :class:`colorlog.Logger` object
:return: the systems and rejected clusters found
:rtype: ([:class:`macsypy.system.System`, ...], [:class:`macsypy.cluster.RejectedCluster`, ...])
"""
working_dir = config.working_dir()
config.save(path_or_buf=os.path.join(working_dir, config.cfg_name))
registry = ModelRegistry()
models_loc_available = scan_models_dir(
config.models_dir(),
profile_suffix=config.profile_suffix(),
relative_path=config.relative_path())
for model_loc in models_loc_available:
registry.add(model_loc)
# build indexes
idx = Indexes(config)
idx.build(force=config.idx)
# create models
parser = DefinitionParser(config, model_bank, gene_bank, registry,
profile_factory)
try:
models_def_to_detect = get_def_to_detect(config.models(), registry)
except KeyError as err:
sys.exit(f"macsyfinder: {err}")
parser.parse(models_def_to_detect)
logger.info(
f"MacSyFinder's results will be stored in working_dir{working_dir}")
logger.info(f"Analysis launched on {config.sequence_db()} for model(s):")
for m in models_def_to_detect:
logger.info(f"\t- {m.fqn}")
models_to_detect = [
model_bank[model_loc.fqn] for model_loc in models_def_to_detect
]
all_genes = []
for model in models_to_detect:
genes = model.mandatory_genes + model.accessory_genes + model.neutral_genes + model.forbidden_genes
# Exchangeable (formerly homologs/analogs) are also added because they can "replace" an important gene...
ex_genes = []
for g in genes:
ex_genes += g.exchangeables
all_genes += (genes + ex_genes)
#############################################
# this part of code is executed in parallel
#############################################
try:
all_reports = search_genes(all_genes, config)
except Exception as err:
sys.exit(str(err))
#############################################
# end of parallel code
#############################################
all_hits = [
hit for subl in [report.hits for report in all_reports] for hit in subl
]
if len(all_hits) > 0:
# It's important to keep this sorting to have in last all_hits version
# the hits with the same replicon_name and position sorted by score
# the best score in first
hits_by_replicon = {}
for hit in all_hits:
if hit.replicon_name in hits_by_replicon:
hits_by_replicon[hit.replicon_name].append(hit)
else:
hits_by_replicon[hit.replicon_name] = [hit]
for rep_name in hits_by_replicon:
hits_by_replicon[rep_name] = get_best_hits(
hits_by_replicon[rep_name], key='score')
hits_by_replicon[rep_name].sort(key=attrgetter('position'))
models_to_detect = sorted(models_to_detect, key=attrgetter('name'))
db_type = config.db_type()
if db_type in ('ordered_replicon', 'gembase'):
systems, rejected_clusters = _search_in_ordered_replicon(
hits_by_replicon, models_to_detect, config, logger)
return systems, rejected_clusters
elif db_type == "unordered":
likely_systems, rejected_hits = _search_in_unordered_replicon(
hits_by_replicon, models_to_detect, logger)
return likely_systems, rejected_hits
else:
assert False, f"dbtype have an invalid value {db_type}"
else:
# No hits detected
return [], []