Skip to content

otf.system.linear_nonlinear

Helpers for systems with a separable linear and nonlinear part.

This module provides mixins and concrete System_* subclasses for ODEs that can be written as L(c) * state + N(c, state) where a parameter-dependent linear operator multiplies the state and a separate nonlinear function provides the remaining dynamics.

Classes:

Name Description
System_LinearNonlinear_ModelKnown

Concrete system: known true-model with separable linear/nonlinear parts.
System_LinearNonlinear_ModelUnknown

Concrete system: unknown true-model, only assimilated parts provided.

System_LinearNonlinear_ModelKnown 

System_LinearNonlinear_ModelKnown(
    mu: float,
    gs: jndarray,
    cs: jndarray,
    observed_mask: jndarray,
    linear_assimilated: Callable[[jndarray], jndarray],
    nonlinear_assimilated_ode: Callable[
        [jndarray, jndarray], jndarray
    ],
    linear_true: Callable[[jndarray], jndarray],
    nonlinear_true_ode: Callable[
        [jndarray, jndarray], jndarray
    ],
    complex_differentiation: bool = False,
    true_observed_mask: jndarray | None = None,
)

Bases: _AssimilatedLinearNonlinearMixin, _TrueLinearNonlinearMixin, System_ModelKnown

Concrete system: known true-model with separable linear/nonlinear parts.

The class constructs compatible assimilated and true ODEs from provided linear and nonlinear component callables.

Source code in src/otf/system/linear_nonlinear.py 
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
def __init__(
    self,
    mu: float,
    gs: jndarray,
    cs: jndarray,
    observed_mask: jndarray,
    linear_assimilated: Callable[[jndarray], jndarray],
    nonlinear_assimilated_ode: Callable[[jndarray, jndarray], jndarray],
    linear_true: Callable[[jndarray], jndarray],
    nonlinear_true_ode: Callable[[jndarray, jndarray], jndarray],
    complex_differentiation: bool = False,
    true_observed_mask: jndarray | None = None,
):
    self._set_assimilated_parts(
        linear_assimilated, nonlinear_assimilated_ode
    )
    assimilated_ode = self._define_ode(
        self.linear_assimilated, self.nonlinear_assimilated_ode
    )
    self._set_true_parts(linear_true, nonlinear_true_ode)
    true_ode = self._define_ode(linear_true, nonlinear_true_ode)
    super().__init__(
        mu,
        gs,
        cs,
        observed_mask,
        assimilated_ode,
        true_ode,
        complex_differentiation,
        true_observed_mask,
    )

System_LinearNonlinear_ModelUnknown 

System_LinearNonlinear_ModelUnknown(
    mu: float,
    cs: jndarray,
    observed_mask: jndarray,
    linear_assimilated: Callable[[jndarray], jndarray],
    nonlinear_assimilated_ode: Callable[
        [jndarray, jndarray], jndarray
    ],
    complex_differentiation: bool = False,
)

Bases: _AssimilatedLinearNonlinearMixin, System_ModelUnknown

Concrete system: unknown true-model, only assimilated parts provided.

Source code in src/otf/system/linear_nonlinear.py 
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
def __init__(
    self,
    mu: float,
    cs: jndarray,
    observed_mask: jndarray,
    linear_assimilated: Callable[[jndarray], jndarray],
    nonlinear_assimilated_ode: Callable[[jndarray, jndarray], jndarray],
    complex_differentiation: bool = False,
):
    self._set_assimilated_parts(
        linear_assimilated, nonlinear_assimilated_ode
    )
    assimilated_ode = self._define_ode(
        self.linear_assimilated, self.nonlinear_assimilated_ode
    )
    super().__init__(
        mu, cs, observed_mask, assimilated_ode, complex_differentiation
    )