def createEvaluationfiles(self, dirCorpus, testPercentage, numberOfPartition, bibl_list): dirPartitions = self.getDirPartitionNames() for dirPartition in dirPartitions: (annotateDir, testDir, trainDir, modelDir, _) = self.getDirTestNames(dirPartition) testCorpus, trainCorpus = FormatEval.getShuffledCorpus(bibl_list, testPercentage) #print testCorpus #print trainCorpus # files used for training (100 - testPercentage % of the corpus) FormatEval.copy_files_for_eval(self.dirCorpus, trainDir, trainCorpus) # files used for evaluation keeping annotations (testPercentage % of the corpus) FormatEval.copy_files_for_eval(self.dirCorpus, testDir, testCorpus) # files used for evaluation, strip the annotations # they will be labeled by bilbo FormatEval.copy_files_for_eval(self.dirCorpus, annotateDir, testCorpus, 'bibl', strip=True)
