[datareposrc] Add function to read flexible tensors
- If the tensor format of the caps of the JSON file is flexible,
Reads a sample using the flexible meta information (sample_offset, tensor_size, tensor_count)
of the JSON file and appends memory to GstBuffer as many as the number of flexible tensors.
- Add unit test
- Add checking if it is a flexible tensor after writing data to GstMemory
- Reference
* The start offset for reading is sample_offset(sample size).
* Save each flexible tensor stored in a sample to a gstbuffer according to each
tensor_size, tensor_size can get tensor_size field in JSON file.
* A shuffled index is mapped to an index of sample_offset field in JSON file.
* A shuffled index is also mapped to an index of tensor_count field in JSON file
* The index value is a number of cumulative tensors, so it is mapped to an index
of tensor_size field in JSON file
Signed-off-by: hyunil park <hyunil46.park@samsung.com>