Dataframes would be a better container for the metadata about a piece and about a group of pieces. This data structure would facilitate the assignment of metadata (which is currently laboriously element-wise) and combining of metadata about individual indexed_pieces to the metadata about their containing aggregated_piece object. This is related to issue #408 and unfortunately won't get addressed in time for VIS 3.0. Since the metadata is currently stored in a dictonary, and you can build a dataframe from a dictionary easily, this shouldn't be too difficult.
Dataframes would be a better container for the metadata about a piece and about a group of pieces. This data structure would facilitate the assignment of metadata (which is currently laboriously element-wise) and combining of metadata about individual indexed_pieces to the metadata about their containing aggregated_piece object. This is related to issue #408 and unfortunately won't get addressed in time for VIS 3.0. Since the metadata is currently stored in a dictonary, and you can build a dataframe from a dictionary easily, this shouldn't be too difficult.