H5PYDataset - Different length sources with same length indices

[areiner222]

Is it possible to use the H5PYDataset with different length sources? I am using vector features and image data as input but the number of distinct images is much less than the number of vector features / targets. I thought possibly using an index reference with duplicate values like a one-to-many mapping might work but no luck. For the sake of clarity, a simple example would be 100 vector features / targets, 10 distinct images and then vector_fatures_index = numpy.arange(100) and image_features_index = numpy.repeat(numpy.arange(10), 10).