def main(args=None):
if args is None:
args = sys_argv[1:]
parser = ArgumentParser(
description='Generate a phylogeny from an alignment.')
parser.add_argument('ALIGNMENT', type=PathType)
parser.add_argument('OUTPUT', type=PathType)
ns = parser.parse_args(args)
msa = load_stockholm(ns.ALIGNMENT)
try:
refidx = reference_index(msa, is_refseq)
except IndexError:
raise RuntimeError('No reference sequence found!')
labels = MSAVectorizer(AminoEncoder).fit(msa).get_feature_names()
seqrecords = [r for i, r in enumerate(msa) if not i == refidx]
tree, alignment = Phylo()(seqrecords)
PhyloGzFile.write(ns.OUTPUT, tree, alignment, labels)
return 0
def main(args=None):
if args is None:
args = sys_argv[1:]
parser = ArgumentParser(description='Generate a phylogeny from an alignment.')
parser.add_argument('ALIGNMENT', type=PathType)
parser.add_argument('OUTPUT', type=PathType)
ns = parser.parse_args(args)
msa = load_stockholm(ns.ALIGNMENT)
try:
refidx = reference_index(msa, is_refseq)
except IndexError:
raise RuntimeError('No reference sequence found!')
labels = SiteVectorizer(AminoEncoder).fit(msa).get_feature_names()
seqrecords = [r for i, r in enumerate(msa) if not i == refidx]
tree, alignment = Phylo()(seqrecords)
PhyloGzFile.write(ns.OUTPUT, tree, alignment, labels)
return 0
def main(args=None):
if args is None:
args = sys.argv[1:]
np.seterr(all='raise')
parser, ns, args = init_args(description='Predict label for unlabeled sequences', args=args)
parser = hmmer_args(parser)
parser.add_argument('MODEL', type=PathType)
parser.add_argument('SEQUENCES', type=PathType)
ARGS = parse_args(parser, args, namespace=ns)
with gzip_open(ARGS.MODEL, 'rb') as fh:
try:
model = pickle_load(fh)
if model[0] != MODEL_VERSION:
raise ImportError('incompatible model version')
ARGS.ENCODER, ARGS.LABEL, hmm, extractor, clf = model[1:]
except ImportError:
msg = 'your model is not of the appropriate version, please re-learn your model'
raise RuntimeError(msg)
# create a temporary file wherein space characters have been removed
with open(ARGS.SEQUENCES) as seq_fh:
def seqrecords():
is_dna = ARGS.ENCODER == DNAEncoder
seq_fmt = seqfile_format(ARGS.SEQUENCES)
source = Verifier(SeqIO.parse(seq_fh, seq_fmt), DNAAlphabet)
try:
for record in source:
yield record if is_dna else translate(record)
except VerifyError:
if is_dna:
msg = (
"your model specifies a DNA encoding "
"which is incompatible with protein sequences"
)
raise RuntimeError(msg)
source.set_alphabet(AminoAlphabet)
for record in source:
yield record
try:
fd, tmphmm = mkstemp(); close(fd)
with open(tmphmm, 'wb') as hmm_fh:
hmm_fh.write(hmm)
# explicitly gc hmm
hmm = None
tmpaln = generate_alignment_(seqrecords(), tmphmm, ARGS)
alignment = load_stockholm(tmpaln, trim=True)
finally:
if exists(tmphmm):
remove(tmphmm)
if exists(tmpaln):
remove(tmpaln)
X = extractor.transform(alignment)
y = clf.predict(X)
feature_names = extractor.get_feature_names()
support = clf.named_steps['mrmr'].support_
labels = ['"{0:s}"'.format(feature_names[i]) for i, s in enumerate(support) if s]
emptys = [' ' * (len(label) + 2) for label in labels]
idlen = max(len(r.id) for r in alignment) + 3
print('{{\n "label": "{0:s}",\n "predictions": ['.format(ARGS.LABEL), file=ARGS.OUTPUT)
for i, r in enumerate(alignment):
if i > 0:
print(',')
features = ['[ ']
for j, x in enumerate(X[i, support]):
if x:
features.append(labels[j])
features.append(', ')
else:
features.append(emptys[j])
features.append(' ]')
# replace the last comma with a space
idx = None
for k, f in enumerate(features):
if f == ', ':
idx = k
if idx is None:
features[0] = features[0].rstrip()
features[-1] = features[-1].lstrip()
else:
features[idx] = ''
features_ = ''.join(features)
print(
' {{{{ "id": {{0:<{0:d}s}} "value": {{1: d}}, "features": {{2:s}} }}}}'.format(
idlen).format('"{0:s}",'.format(r.id), y[i], features_),
file=ARGS.OUTPUT, end='')
print('\n ]\n}', file=ARGS.OUTPUT)
finalize_args(ARGS)
return 0
def main(args=None):
if args is None:
args = sys.argv[1:]
np.seterr(all='raise')
parser, ns, args = init_args(description='Predict label for unlabeled sequences', args=args)
parser = hmmer_args(parser)
parser.add_argument('MODEL', type=PathType)
parser.add_argument('SEQUENCES', type=PathType)
ARGS = parse_args(parser, args, namespace=ns)
with gzip_open(ARGS.MODEL, 'rb') as fh:
try:
model = pickle_load(fh)
if model[0] != 4:
raise ImportError('incompatible model version')
ARGS.ENCODER, ARGS.LABEL, hmm, extractor, clf = model[1:]
except ImportError:
msg = 'your model is not of the appropriate version, please re-learn your model'
raise RuntimeError(msg)
# create a temporary file wherein space characters have been removed
with open(ARGS.SEQUENCES) as seq_fh:
def seqrecords():
is_dna = ARGS.ENCODER == DNAEncoder
seq_fmt = seqfile_format(ARGS.SEQUENCES)
source = Verifier(SeqIO.parse(seq_fh, seq_fmt), DNAAlphabet)
try:
for record in source:
yield record if is_dna else translate(record)
except VerifyError:
if is_dna:
msg = (
"your model specifies a DNA encoding "
"which is incompatible with protein sequences"
)
raise RuntimeError(msg)
source.set_alphabet(AminoAlphabet)
for record in source:
yield record
try:
fd, tmphmm = mkstemp(); close(fd)
with open(tmphmm, 'wb') as hmm_fh:
hmm_fh.write(hmm)
# explicitly gc hmm
hmm = None
tmpaln = generate_alignment_(seqrecords(), tmphmm, ARGS)
alignment = load_stockholm(tmpaln, trim=True)
finally:
if exists(tmphmm):
remove(tmphmm)
if exists(tmpaln):
remove(tmpaln)
X = extractor.transform(alignment)
y = clf.predict(X)
feature_names = extractor.get_feature_names()
support = clf.named_steps['mrmr'].support_
labels = ['"{0:s}"'.format(feature_names[i]) for i, s in enumerate(support) if s]
emptys = [' ' * (len(label) + 2) for label in labels]
idlen = max(len(r.id) for r in alignment) + 3
print('{{\n "label": "{0:s}",\n "predictions": ['.format(ARGS.LABEL), file=ARGS.OUTPUT)
for i, r in enumerate(alignment):
if i > 0:
print(',')
features = ['[ ']
for j, x in enumerate(X[i, support]):
if x:
features.append(labels[j])
features.append(', ')
else:
features.append(emptys[j])
features.append(' ]')
# replace the last comma with a space
idx = None
for k, f in enumerate(features):
if f == ', ':
idx = k
if idx is None:
features[0] = features[0].rstrip()
features[-1] = features[-1].lstrip()
else:
features[idx] = ''
features_ = ''.join(features)
print(
' {{{{ "id": {{0:<{0:d}s}} "value": {{1: d}}, "features": {{2:s}} }}}}'.format(
idlen).format('"{0:s}",'.format(r.id), y[i], features_),
file=ARGS.OUTPUT, end='')
print('\n ]\n}', file=ARGS.OUTPUT)
finalize_args(ARGS)
return 0
