def main(args):
if not args.absolutePaths:
args.labelDir = os.path.join(root, args.labelDir)
args.dataDir = os.path.join(root, args.dataDir)
args.destDir = os.path.join(root, args.destDir)
if not checkInputs(args):
return
corpus = Corpus(args.dataDir)
corpusLabel = CorpusLabel(args.labelDir, corpus=corpus)
corpusLabel.getEverything()
columnData = {}
for relativePath in corpusLabel.labels.keys():
columnData[relativePath] = pandas.Series(
corpusLabel.labels[relativePath]["label"])
corpus.addColumn("label", columnData)
corpus.copy(newRoot=args.destDir)
print "Done adding labels!"
def main(args):
if not args.absolutePaths:
args.labelDir = os.path.join(root, args.labelDir)
args.dataDir = os.path.join(root, args.dataDir)
args.destDir = os.path.join(root, args.destDir)
if not checkInputs(args):
return
corpus = Corpus(args.dataDir)
corpusLabel = CorpusLabel(args.labelDir, corpus=corpus)
corpusLabel.getEverything()
columnData = {}
for relativePath in corpusLabel.labels.keys():
columnData[relativePath] = pandas.Series(
corpusLabel.labels[relativePath]["label"])
corpus.addColumn("label", columnData)
corpus.copy(newRoot=args.destDir)
print "Done adding labels!"
