def fextract(self): """Extract features""" self.load_char_dicts() for dirname, char_dict in (("eval", self.eval_char_dict), ("train", self.train_char_dict)): for char_code, char_list in char_dict.items(): output_dir = os.path.join(self.FEATURES_ROOT, dirname) if not os.path.exists(output_dir): os.makedirs(output_dir) sequence_set = [] for character in char_list: writing = character.get_writing() char_features = self.get_feature_vectors(writing) sequence_set.append(char_features) output_file = os.path.join(output_dir, str(char_code) + ".sset") self.print_verbose(output_file + " (%d chars)" % \ len(sequence_set)) sset = SequenceSet(sequence_set) sset.write(output_file)
def get_sequence_set(self, file_path): return SequenceSet.from_file(file_path)