concepts.dsl.executors.tensor_value_executor.compose_bvdict

compose_bvdict(input_dict, state=None)

Compose a bounded variable dict from raw inputs.

Parameters:

• input_dict (None | Sequence[Variable] | Dict[str | Variable, str | int | slice | bool | float | Tensor | TensorValue | ObjectConstant | StateObjectReference] | Dict[str, Dict[str, StateObjectReference | slice | TensorValue]]) –

  the input dict. There are three types of inputs:

  1. A sequence of concepts.dsl.dsl_types.Variable instances, which represents a set of variables with no values.

  2. A dictionary mapping from concepts.dsl.dsl_types.Variable instances to the actual value.

  3. A dictionary mapping from strings (the name of the variables) to values.

  Acceptable values are:

  1. A str, which represents a reference to an object in the state (so the state must be object-named).

  2. An integer, which represents a reference to an object in the state (so the state must be object-named).

  3. A QINDEX (a.k.a., slice(None)), which represents all objects in the state of a given type (so the state must be object-named).

  4. A concepts.dsl.tensor_value.TensorValue instance, which represents an actual value.

  5. A StateObjectReference instance, which represents a reference to an object in the state (so the state must be object-named).

  6. A bool, int, float, or torch.Tensor instance, which represents an actual value. They will be converted to a concepts.dsl.tensor_value.TensorValue instance.

• state (TensorState | None) – the state.

Returns:

a dictionary mapping from strings (the typename) to a dictionary mapping from strings (the name of the variables) to values.

Return type:

Dict[str, Dict[str, StateObjectReference | slice | TensorValue]]