length = functions.check_length(args.test)
#Get model data
if args.model == None:
model = os.path.join(os.path.join(base_folder, "Data"),
"base_model_" + str(length))
else:
model = args.model
data = pickle.load(open(os.path.join(model, "data")))
#Tranform data
if algorithm == "CROSSED":
crossed_test = functions.convert_crossed(filename = args.test,
length = length, output = "test_crossed_" + args.output,
features = data['crossed_features'], positive = True,
folder = "", extra = data['extra'])
else:
if algorithm == "REC":
crossed_test = functions.convert_crossed(
filename = args.test, length = length,
output = "test_crossed_" + args.output,
features = data['rec_crossed_features'],
positive = True, folder = "", extra = data['extra'])
ric_test = functions.convert_ric(filename = args.test,
length = length, output = "test_ric_" + args.output,
folder = "")
#Apply algorithms
if algorithm == "CROSSED":
crossed_output = functions.run_crossed(filename=crossed_test,
negative = ric_negative_scores,
optim = "SEN")
#Do CRoSSeD
#Check if feature file was provided and if it's ok
if args.features != None:
features = functions.read_feature_file(args.features)
else:
features = functions.read_feature_file(os.path.join(
base_folder, "Data/crossed_features_" + str(length) +
".txt"))
#Convert input files to CRoSSeD format
crossed_positives = functions.convert_crossed(
filename=args.positive,
length=length,
output="train_crossed_positives.fasta",
features=features, positive=True, folder=unique,
extra=extra)
crossed_negatives = functions.convert_crossed(
filename=args.negative,
length=length,
output="train_crossed_negatives.fasta",
features=features, positive=False, folder=unique,
extra=extra)
#Train CRoSSeD model
model = functions.train_crossed(positive=crossed_positives,
negative=crossed_negatives, model="CRoSSeD_model",
unique=unique, base_folder=base_folder)
#Aply CRoSSeD to input data
crossed_positives_output = functions.run_crossed(
filename=crossed_positives, model=model,
