goflow.adapter.init_ivp

class goflow.adapter.init_ivp.morph(construct: networkx.classes.graph.Graph, mode: str = 'default', args: tuple = <factory>)

Bases: object

Basic Class for simulating flow network adaption dynamics.

‘Morph’ computes the long-term network development on the basis of flow and gradient descent. The class acts as a wrapper to ensure correct pre-/post-proccessing of the network’s data and output. Utilizing the modules ‘kirchhoff’ and ‘hailhydro’ as well as goflow.models.

Args:

construct (nx.Graph):

A network plexus, recommended to be ‘kirchhoff’ circuit or an initialized ‘hailhydro’ flow/flux. Handling for deviations only provided for standard models.

mode (string/model):

A descriptive string for selection internal model object.

args (tuple):

A parameter tuple for evaluation of the model given a plexus.

Returns:

type: morph

init_model_and_flow()

Set internal model and flow variables with constructor values.

Args:

None

Returns:

None

Raises:

Exception:

sys.exit(‘Terminate Program: No valid adaptation model provided.’)

link_model_flow()

Transfer model and flow parameters to internal variables of solver.

Args:

None

Returns:

None

class goflow.adapter.init_ivp.morph_dynamic(construct: networkx.classes.graph.Graph, mode: str = 'default', args: tuple = <factory>)

Bases: goflow.adapter.init_ivp.morph

Derived class definition from morph, for specific dynamic models.

Defines explicit ODE solvers and customizable forward integration solvers.

Args:

construct (nx.Graph):

A network plexus, recommended to be ‘kirchhoff’ circuit or an initialized ‘hailhydro’ flow/flux. Handling for deviations only provided for standard models.

mode (string/model):

A descriptive string for selection internal model object.

args (tuple):

A parameter tuple for evaluation of the model given a plexus.

Returns:

type: morph_dynamic

autoSolve(t_span, x0)

Default ODE solver/wrapper for given adaptation models.

Based on methods of scipy.integrate, hence solver options are customizable via the internal model.solver_options attribute and align with scipy standards.

Args:

t_span (tuple):

A tuple setting the linear range of values for explicit evaluation.

x0 (array):

An array of initial values to start the numeric integration of the dynamic system.

Returns:

Iterable: Same set of return values as scipy.integrate.solve_ivp (1.7.3)

nlogSolve(t_span, x0)

Custom ODE solver/wrapper for given adaptation models with logarithmic evaluation timescales.

Based on methods of scipy.integrate, hence solver options are customizable via the internal model.solver_options attribute and align with scipy standards.

Args:

t_span (tuple):

A tuple setting the logarithmic range of values for explicit evaluation.

x0 (array):

An array of initial values to start the numeric integration of the dynamic system.

Returns:

Iterable: Same set of return values as scipy.integrate.solve_ivp (1.7.3)

nsolve(ds_func, t_span, x0, **kwargs)

Custom ODE solver/wrapper for given adaptation models with linear evaluation timescales and custom update function. Based on methods of scipy.integrate.

Args:

ds_func (callable function):

The update function of the dynamic system, which is supposed to define dx/dt, according to scipy standards.

t_span (tuple):

A tuple setting the logarithmic range of values for explicit evaluation.

x0 (array):

An array of initial values to start the numeric integration of the dynamic system.

kwargs (dict):

Dictionary for setting solver options according to scipy.integrate.solve_ivp standards.

Returns:

Iterable: Same set of return values as scipy.integrate.solve_ivp (1.7.3)

nsolve_custom(ds_func, x0, **kwargs)

Custom ODE solver onthe basis of Foward-Euler, for given adaptation models with custom evaluation timescales and step width and custom update function. Acts as a wrapper for nsolve_fw_euler.

Args:

ds_func (callable function):

The update function of the dynamic system, which is supposed to

define dx/dt, according to scipy standards.

t_span (tuple):

A tuple setting the logarithmic range of values for explicit evaluation.

x0 (array):

An array of initial values to start the numeric integration of the dynamic system.

kwargs (dict):

A dictionary setting custom solver options such as number of evaluations ‘sample’, number of steps ‘num_steps’

Returns:

Iterable: Tuple of two arrays (t, x(t))

nsolve_fw_euler(ds_func, x0, **kwargs)

Custom ODE solver onthe basis of Foward-Euler, for given adaptation models with custom evaluation timescales and step width and custom update function.

Args:

ds_func (callable):

The update function of the dynamic system, which is supposed to define dx/dt, according to scipy standards.

t_span (tuple):

A tuple setting the logarithmic range of values for explicit evaluation.

x0 (array):

An array of initial values to start the numeric integration of the dynamic system.

kwargs (dict):

A dictionary setting custom solver options.

Returns:

Iterable: Tuple of two arrays (t, x(t))

set_integration_scale(Num_steps, sample)

Computes the sample rate for the given dynamic system.

Given the pre-set number of computing steps and number of evaluations, Adjusts/increases the number of integration cycles such that last evaluation coincicdes with last cycle.

Args:

Num_steps (int):

Number of integration steps to perform for a Forward-Euler scheme.

sample(int):

Number of desired function evaluations for output.

Returns:

Iterable: Num_steps, sample_rate

class goflow.adapter.init_ivp.proxy_solver(t_samples: list = <factory>, sol: list = <factory>)

Bases: object

A proxy solver class for custom numeric integration.

Args:

tsamples (array):

Setting the internal sample time points.

sol (array):

Time series of data to be stored internally.

Returns:

type: proxy_solver