MEHPForceEvaluator¶

class pylimer_tools_cpp.MEHPForceEvaluator(self: pylimer_tools_cpp.MEHPForceEvaluator)¶

Bases: pybind11_object

The base interface to change the way the force is evaluated during a MEHP run.

To create a custom force evaluator in Python, subclass this class and implement:

evaluate_force_set_gradient(): Compute the force and optionally its gradient
evaluate_stress_contribution(): Compute stress tensor contributions
prepare_for_evaluations(): Prepare for a batch of evaluations

Attributes Summary

`is_2d`
`network`

Methods Summary

`evaluate_force_set_gradient`(self, n, ...[, ...])	Evaluate the force and optionally compute its gradient.
`evaluate_stress_contribution`(self, ...)	Evaluate the stress-contribution for a single spring.
`prepare_for_evaluations`(self)	Prepare the evaluator for a series of evaluations.

Attributes Documentation

is_2d¶

network¶

Methods Documentation

evaluate_force_set_gradient(self: pylimer_tools_cpp.MEHPForceEvaluator, n: SupportsInt, spring_distances: Annotated[numpy.typing.ArrayLike, numpy.float64, '[m, 1]'], compute_gradient: bool = False) → tuple¶

Evaluate the force and optionally compute its gradient.

This is one of the primary methods to override when creating a custom force evaluator. Override this in Python by defining a method that returns a tuple of (force, gradient).

Parameters:

n – The dimensionality of the problem (the number of coordinates)
spring_distances – The sequential (x, y, z) spring distances as a vector
compute_gradient – If True, the gradient should be computed and returned

Returns:

A tuple (force, gradient) where: - force: The computed force value (float) - gradient: The gradient as a list (or None if compute_gradient was False)

Example implementation in Python:

def evaluate_force_set_gradient(self, n, spring_distances, compute_gradient):
    force = 0.0
    # ... compute force ...

    gradient = None
    if compute_gradient:
        gradient = [0.0] * n
        # ... compute gradient ...

    return (force, gradient)

evaluate_stress_contribution(self: pylimer_tools_cpp.MEHPForceEvaluator, spring_distances: SupportsFloat, i: SupportsInt, j: SupportsInt, spring_index: SupportsInt) → float¶

Evaluate the stress-contribution for a single spring.

Parameters:

spring_distances – The three coordinate differences for one spring (list of 3 floats)
i – The row index of the stress tensor
j – The column index of the stress tensor
spring_index – The index of the spring being evaluated

Returns:

The stress contribution value

prepare_for_evaluations(self: pylimer_tools_cpp.MEHPForceEvaluator) → None¶: Prepare the evaluator for a series of evaluations. This method is called before a batch of force evaluations, allowing for any necessary setup or precomputation.