mici.types module — Mici 0.2.0 documentation

mici.types.AdaptationStatisticFunction#

Function returning adaptation statistic given dictionary of transition statistics.

alias of Callable[[dict[str, bool | int | float | LogRepFloat | number]], float]

mici.types.AdapterState#

Dictionary defining current state of an mici.adapters.Adapter.

alias of dict[str, Any]

mici.types.ArrayFunction#

Function taking an array-like input and returning a array-like output.

mici.types.GradientFunction#

Function returning the gradient of a scalar-valued function.

May optionally also return scalar-like value of function.

mici.types.HessianFunction#

Function returning the Hessian matrix of a scalar-valued function.

May optionally also return gradient and value of function.

mici.types.JacobianFunction#

Function returning the Jacobian of a array (vector) valued function.

May optionally aslo return the array value of the function.

mici.types.MatrixHessianProduct#

Function returning the product between a matrix and Hessian of an array function.

Hessian here is the 3-dimensional array of second-order derivatives of an array-valued function.

mici.types.MatrixHessianProductFunction#

Function returning a matrix-Hessian product function for an array function.

Hessian here is the 3-dimensional array of second-order derivatives of an array-valued function.

May optionally also return the Jacobian matrix and array value of function.

mici.types.MatrixLike#

Matrix like objects.

mici.types.MatrixTressianProduct#

Function returning the product between a matrix and ‘Tressian’ of a scalar function.

Tressian here is the 3-dimensional array of third-order derivatives of a scalar function.

mici.types.MatrixTressianProductFunction#

Function returning a matrix-Tressian product function for a scalar function.

Tressian here is the 3-dimensional array of third-order derivatives of a scalar function.

May optionally also return the Hessian matrix, gradient and value of function.

mici.types.MetricLike#

Metric (positive definite matrix) like objects.

mici.types.PyTree#

Arbitrarily nested structure of Python dict, list and tuple types.

alias of dict[Any, PyTree] | list[PyTree] | tuple[PyTree] | T

mici.types.ReducerFunction#

Function combining per-chain control variables from adapters.

alias of Callable[[Collection[float]], float]

mici.types.ScalarFunction#

Function taking an array-like input and returning a scalar-like output.

mici.types.ScalarLike: TypeAlias = bool | int | float | mici.utils.LogRepFloat | numpy.number#

Scalar like objects.

mici.types.SystemStateMethod#

System method computing function of chain state.

mici.types.SystemStateWithAuxMethod#

System method computing function chain state with auxiliary return values.

mici.types.TerminationCriterion#

Function indicating whether to terminate trajectory tree expansion.

mici.types.TraceFunction#

Function returning a dictionary of chain variables to trace given chain state.

mici.types.TransitionStatistics#

Dictionary of statistics computed by mici.transitions.Transition.sample()

alias of dict[str, bool | int | float | LogRepFloat | number]

mici.types.VectorJacobianProduct#

Function returning the product between an array and Jacobian of a array function.

Jacobian here is the (d+1)-dimensional array of first-order derivatives of an d-dimensional array-valued function taking a 1-dimensional array as input.

mici.types.VectorJacobianProductFunction#

Function returning a vector-Jaccobian product function for an array function.

Jacobian here is the (d+1)-dimensional array of first-order derivatives of an d-dimensional array-valued function taking a 1-dimensional array as input.

May optionally also return the array value of function.

mici.types module#