<!-- filepath: /home/philip/Documents/elsciRL-Wiki/Documentation/Encoders/observable_objects_encoded.md --> # Observable Objects Encoder This encoder transforms a list of observable objects in the environment into a vector or tensor representation. It is useful for environments where the state is described by the presence or type of objects (e.g., pieces on a chessboard). ## Class: `ObjectEncoder` - Maps each object to a unique index. - Encodes the state as a tensor, where each entry represents the occurrence of an object type. - Supports both indexed and one-hot (flattened) representations. ### Example Usage ```python from elsciRL.encoders.observable_objects_encoded import ObjectEncoder local_objects = ['pawn', 'knight', 'bishop', 'rook', 'queen', 'king'] encoder = ObjectEncoder(local_objects) state = ['pawn', 'pawn', 'knight', 'empty', 'king'] encoded = encoder.encode(state) print(encoded.shape) # Output: (output_dim,) ``` ---