Add snapshot matrix builder

[mahmudulh-tamim]

Location: utils/snapshot_matrix_builder/
The class from the module can be imported,

from snapshot_matrix_builder import Snapshot_matrix_builder

An object of the class can be instantiated the following way:

snapshot_matrix_builder_object = Snapshot_matrix_builder(snapshot_manager_object=sm_object, list_fields_paths=chosen_fields)

Here, sm_object is an instance of the Snapshot_manager class (from the module in utils/snapshot_manager/snapshot_manager.py) and chosen_fields is a list of fields(be that observable or non-observable) over which we are interested in applying DA algorithms. It should be noted that the matrices are created at the instantiation of the object. Moreover, it should be noted that two sets of matrices are created: one for training and another for testing. In each set we have two matrices of the same data: a 2D flattened numpy array and a 3D structured multilevel list. For extracting the fields from the encrypted 2D matrices, we store start index and end index of each field in a snap as a tuple in another list.