def setup(self, inDataFile, inProbArrFile, archModelInFile):
"""Method to load in the data and process the target data"""
self.archModel = loadArchModel(archModelInFile)
self.archModel.setupParams()
ifs = open(inProbArrFile)
self.probArr = []
for line in ifs:
chunks = line.strip().split()
self.probArr.append([int(chunks[0]),
int(chunks[2])])
ifs.close()
self.probArr = array(self.probArr)
self.fDictList = []
ifs = gzip.open(inDataFile)
reader = FeatureDictReader(ifs)
for d in reader:
self.fDictList.append(d)
ifs.close()
self.classifier = PairMonteFeatDictClassifier(self.archModel,
self.fDictList,
self.probArr,
self.postEpochCallback)
self.classifier.setupModels()
def setup(self, inDataFile, inTargDataFile, archModelInFile):
"""Method to load in the data and process the target data"""
self.archModel = loadArchModel(archModelInFile)
self.archModel.setupParams()
ifs = open(inTargDataFile)
self.idxArr = []
self.targArr = []
for iRow, line in enumerate(ifs):
row = line.strip().split()
self.idxArr.append(int(row[0]))
self.targArr.append(float(row[self.targcolumn]))
ifs.close()
self.idxArr = array(self.idxArr)
self.targArr = array(self.targArr)
self.fDictList = []
ifs = gzip.open(inDataFile)
reader = FeatureDictReader(ifs)
for d in reader:
self.fDictList.append(d)
ifs.close()
self.classifier = MonteFeatDictClassifier(self.archModel,
self.fDictList, self.targArr,
self.idxArr,
self.postEpochCallback)
self.classifier.setupModels()
def main(self, argv):
"""Callable from Command line"""
if argv is None:
argv = sys.argv
usageStr = \
"""usage: %prog [options] archModelFile featDataFile idxArrFile outFile
"""
parser = OptionParser(usage = usageStr);
parser.add_option('-d', '--deduplicate', dest='deduplicate', action='store_true', default=False,
help='Deduplicate the idArr');
parser.add_option('--delimiter', dest='delimiter', default=' ');
(options, args) = parser.parse_args(argv[1:])
if len(args) == 4:
(archModelFile, featDataFile, idxArrFile, outFile) = args;
self.archModel = loadArchModel(archModelFile);
self.setup();
ifs = open(idxArrFile);
reader = csv.reader(ifs, quoting=csv.QUOTE_NONE, delimiter=options.delimiter);
seenIds = set([]);
idArr = [];
for row in reader:
row[0] = int(row[0]);
row[1:] = [float(x) for x in row[1:]];
if options.deduplicate:
if row[0] not in seenIds:
idArr.append(row);
seenIds.add(row[0]);
else:
idArr.append(row);
ifs.close();
ifs = gzip.open(featDataFile)
reader = FeatureDictReader(ifs);
fDictList = [];
for aDict in reader:
fDictList.append(aDict);
ifs.close();
progress = ProgressDots();
ofs = open(outFile, 'w');
for idChunk, predValChunk in self.predict(fDictList, idArr):
for iRow in range(len(idChunk)):
chunkTargData = [str(x) for x in idChunk[iRow]]
chunkTargData = ' '.join(chunkTargData)
print >>ofs, chunkTargData, predValChunk[iRow];
progress.Update();
ofs.close();
progress.PrintStatus();
else:
parser.print_help();
sys.exit(2);
def setup(self, inDataFile, inProbArrFile, archModelInFile):
"""Method to load in the data and process the target data"""
self.archModel = loadArchModel(archModelInFile);
self.archModel.setupParams();
ifs = open(inProbArrFile);
self.probArr = [];
for line in ifs:
chunks = line.strip().split()
self.probArr.append([int(chunks[0]), int(chunks[2])]);
ifs.close();
self.probArr = array(self.probArr);
self.fDictList = [];
ifs = gzip.open(inDataFile);
reader = FeatureDictReader(ifs);
for d in reader:
self.fDictList.append(d);
ifs.close();
self.classifier = PairMonteFeatDictClassifier(self.archModel, self.fDictList,
self.probArr, self.postEpochCallback)
self.classifier.setupModels();
def main(self, argv):
"""Callable from Command line"""
if argv is None:
argv = sys.argv
usageStr = \
"""usage: %prog [options] archModelFile featDataFile outFile
"""
parser = OptionParser(usage = usageStr);
(options, args) = parser.parse_args(argv[1:])
if len(args) == 3:
archModelFile = args[0]
featDataFile = args[1]
outFile = args[2]
#Read in the
self.archModel = loadArchModel(archModelFile);
self.setup();
# Then set up the reader and writer
ifs = gzip.open(featDataFile);
reader = FeatureDictReader(ifs);
ofs = open(outFile, 'w');
progress = ProgressDots();
for value in self.predict(reader):
print >> ofs, value;
progress.Update();
progress.PrintStatus();
ifs.close();
ofs.close();
else:
parser.print_help();
sys.exit(2);
def main(self, argv):
"""Callable from Command line"""
if argv is None:
argv = sys.argv
usageStr = \
"""usage: %prog [options] archModelFile featDataFile outFile
"""
parser = OptionParser(usage=usageStr)
(options, args) = parser.parse_args(argv[1:])
if len(args) == 3:
archModelFile = args[0]
featDataFile = args[1]
outFile = args[2]
#Read in the
self.archModel = loadArchModel(archModelFile)
self.setup()
# Then set up the reader and writer
ifs = gzip.open(featDataFile)
reader = FeatureDictReader(ifs)
ofs = open(outFile, 'w')
progress = ProgressDots()
for value in self.predict(reader):
print >> ofs, value
progress.Update()
progress.PrintStatus()
ifs.close()
ofs.close()
else:
parser.print_help()
sys.exit(2)
def loadArchModelFromFile(self, fileName):
"""Convenience to load up the archModel from a file"""
self.archModel = loadArchModel(fileName);
def loadArchModelFromFile(self, fileName):
"""Convenience to load up the archModel from a file"""
self.archModel = loadArchModel(fileName);