Scalable models¶

Scalability study - API¶

This API facilitates the use of the gemseo.problems.scalable.data_driven.study package implementing classes to benchmark MDO formulations based on scalable disciplines.

ScalabilityStudy class implements the concept of scalability study:

By instantiating a ScalabilityStudy, the user defines the MDO problem in terms of design parameters, objective function and constraints.
For each discipline, the user adds a dataset stored in a Dataset and select a type of ScalableModel to build the ScalableDiscipline associated with this discipline.
The user adds different optimization strategies, defined in terms of both optimization algorithms and MDO formulation.
The user adds different scaling strategies, in terms of sizes of design parameters, coupling variables and equality and inequality constraints. The user can also define a scaling strategies according to particular parameters rather than groups of parameters.
Lastly, the user executes the ScalabilityStudy and the results are written in several files and stored into directories in a hierarchical way, where names depends on both MDO formulation, scaling strategy and replications when it is necessary. Different kinds of files are stored: optimization graphs, dependency matrix plots and of course, scalability results by means of a dedicated class: ScalabilityResult.

Functions:

`create_scalability_study`(objective, ...[, ...])	This method creates a `ScalabilityStudy`.
`plot_scalability_results`(study_directory)	This method plots the set of `ScalabilityResult` generated by a `ScalabilityStudy` and located in the directory created by this study.

gemseo.problems.scalable.data_driven.api.create_scalability_study(objective, design_variables, directory='study', prefix='', eq_constraints=None, ineq_constraints=None, maximize_objective=False, fill_factor=0.7, active_probability=0.1, feasibility_level=0.8, start_at_equilibrium=True, early_stopping=True, coupling_variables=None)[source]

This method creates a ScalabilityStudy. It requires two mandatory arguments:

the 'objective' name,
the list of 'design_variables' names.

Concerning output files, we can specify:

the directory which is 'study' by default,
the prefix of output file names (default: no prefix).

Regarding optimization parametrization, we can specify:

the list of equality constraints names (eq_constraints),
the list of inequality constraints names (ineq_constraints),
the choice of maximizing the objective function (maximize_objective).

By default, the objective function is minimized and the MDO problem is unconstrained.

Last but not least, with regard to the scalability methodology, we can overwrite:

the default fill factor of the input-output dependency matrix ineq_constraints,
the probability to set the inequality constraints as active at initial step of the optimization active_probability,
the offset of satisfaction for inequality constraints feasibility_level,
the use of a preliminary MDA to start at equilibrium start_at_equilibrium,
the post-processing of the optimization database to get results earlier than final step early_stopping.

Parameters

objective (str) – name of the objective
design_variables (list(str)) – names of the design variables
directory (str) –
working directory of the study. Default: ‘study’.

By default it is set to study.
prefix (str) –
prefix for the output filenames. Default: ‘’.

By default it is set to .
eq_constraints (list(str)) –
names of the equality constraints. Default: None.

By default it is set to None.
ineq_constraints (list(str)) –
names of the inequality constraints Default: None.

By default it is set to None.
maximize_objective (bool) –
maximizing objective. Default: False.

By default it is set to False.
fill_factor (float) –
default fill factor of the input-output dependency matrix. Default: 0.7.

By default it is set to 0.7.
active_probability (float) –
probability to set the inequality constraints as active at initial step of the optimization. Default: 0.1

By default it is set to 0.1.
feasibility_level (float) –
offset of satisfaction for inequality constraints. Default: 0.8.

By default it is set to 0.8.
start_at_equilibrium (bool) –
start at equilibrium using a preliminary MDA. Default: True.

By default it is set to True.
early_stopping (bool) –
post-process the optimization database to get results earlier than final step.

By default it is set to True.

gemseo.problems.scalable.data_driven.api.plot_scalability_results(study_directory)[source]

This method plots the set of ScalabilityResult generated by a ScalabilityStudy and located in the directory created by this study.

Parameters: study_directory (str) – directory of the scalability study.

“

Scalable MDO problem¶

This module implements the concept of scalable problem by means of the ScalableProblem class.

Given

a MDO scenario based on a set of sampled disciplines with a particular problem dimension,
a new problem dimension (= number of inputs and outputs),

a scalable problem:

makes each discipline scalable based on the new problem dimension,
creates the corresponding MDO scenario.

Then, this MDO scenario can be executed and post-processed.

We can repeat this tasks for different sizes of variables and compare the scalability, which is the dependence of the scenario results on the problem dimension.

Classes:

ScalableProblem(datasets, design_variables, ...)

Scalable problem.

class gemseo.problems.scalable.data_driven.problem.ScalableProblem(datasets, design_variables, objective_function, eq_constraints=None, ineq_constraints=None, maximize_objective=False, sizes=None, **parameters)[source]

Scalable problem.

Constructor.

Parameters

datasets (list(Dataset)) – disciplinary datasets.
design_variables (list(str)) – list of design variable names
objective_function (str) – objective function
eq_constraints (list(str)) –
equality constraints. Default: None.

By default it is set to None.
eq_constraints –
inequality constraints. Default: None.

By default it is set to None.
maximize_objective (bool) –
maximize objective. Default: False.

By default it is set to False.
sizes (dict) –
sizes of input and output variables. If None, use the original sizes. Default: None.

By default it is set to None.
parameters – optional parameters for the scalable model.

Methods:

`create_scenario`([formulation, ...])	Create MDO scenario from the scalable disciplines.
`exec_time`([do_sum])	Get total execution time per discipline.
`plot_1d_interpolations`([save, show, step, ...])	Plot 1d interpolations.
`plot_coupling_graph`()	Plot a coupling graph.
`plot_dependencies`([save, show, directory])	Plot dependency matrices.
`plot_n2_chart`([save, show])	Plot a N2 chart.

Attributes:

`is_feasible`	Get the feasibility property of the scenario.
`n_calls`	Get number of disciplinary calls per discipline.
`n_calls_linearize`	Get number of disciplinary calls per discipline.
`n_calls_linearize_top_level`	Get number of top level disciplinary calls per discipline.
`n_calls_top_level`	Get number of top level disciplinary calls per discipline.
`status`	Get the status of the scenario.

create_scenario(formulation='DisciplinaryOpt', scenario_type='MDO', start_at_equilibrium=False, active_probability=0.1, feasibility_level=0.5, **options)[source]

Create MDO scenario from the scalable disciplines.

Parameters

formulation (str) –
MDO formulation. Default: ‘DisciplinaryOpt’.

By default it is set to DisciplinaryOpt.
scenario_type (str) –
type of scenario (‘MDO’ or ‘DOE’). Default: ‘MDO’.

By default it is set to MDO.
start_at_equilibrium (bool) –
start at equilibrium using a preliminary MDA. Default: True.

By default it is set to False.
active_probability (float) –
probability to set the inequality constraints as active at initial step of the optimization. Default: 0.1.

By default it is set to 0.1.
feasibility_level (float) –
offset of satisfaction for inequality constraints. Default: 0.5.

By default it is set to 0.5.
options – formulation options.

exec_time(do_sum=True)[source]

Get total execution time per discipline.

Parameters

do_sum (bool) –

sum over disciplines (default: True)

By default it is set to True.

Returns

execution time

Return type

list(float) or float

property is_feasible: Get the feasibility property of the scenario.

property n_calls

Get number of disciplinary calls per discipline.

Returns: number of disciplinary calls per discipline
Return type: list(int) or int

property n_calls_linearize

Get number of disciplinary calls per discipline.

Returns: number of disciplinary calls per discipline
Return type: list(int) or int

property n_calls_linearize_top_level

Get number of top level disciplinary calls per discipline.

Returns: number of top level disciplinary calls per discipline
Return type: list(int) or int

property n_calls_top_level

Get number of top level disciplinary calls per discipline.

Returns: number of top level disciplinary calls per discipline
Return type: list(int) or int

plot_1d_interpolations(save=True, show=False, step=0.01, varnames=None, directory='.', png=False)[source]

Plot 1d interpolations.

Parameters

save (bool) –
save plot. Default: True.

By default it is set to True.
show (bool) –
show plot. Default: False.

By default it is set to False.
step (bool) –
Step to evaluate the 1d interpolation function Default: 0.01.

By default it is set to 0.01.
varnames (list(str)) –
names of the variable to plot; if None, all variables are plotted. Default: None.

By default it is set to None.
directory (str) –
directory path. Default: ‘.’.

By default it is set to ..
png (bool) –
if True, the file format is PNG. Otherwise, use PDF. Default: False.

By default it is set to False.

plot_coupling_graph()[source]

Plot a coupling graph.

plot_dependencies(save=True, show=False, directory='.')[source]

Plot dependency matrices.

Parameters

save (bool) –
save plot (default: True)

By default it is set to True.
show (bool) –
show plot (default: False)

By default it is set to False.
directory (str) –
directory path (default: ‘.’)

By default it is set to ..

plot_n2_chart(save=True, show=False)[source]

Plot a N2 chart.

Parameters

save (bool) –
save plot. Default: True.

By default it is set to True.
show (bool) –
show plot. Default: False.

By default it is set to False.

property status: Get the status of the scenario.

Scalable discipline¶

The discipline implements the concept of scalable discipline. This is a particular discipline built from a input-output learning dataset associated with a function and generalizing its behavior to a new user-defined problem dimension, that is to say new user-defined input and output dimensions.

Alone or in interaction with other objects of the same type, a scalable discipline can be used to compare the efficiency of an algorithm applying to disciplines with respect to the problem dimension, e.g. optimization algorithm, surrogate model, MDO formulation, MDA, …

The ScalableDiscipline class implements this concept. It inherits from the MDODiscipline class in such a way that it can easily be used in a Scenario. It is composed of a ScalableModel.

The user only needs to provide:

the name of a class overloading ScalableModel,
a dataset as an Dataset
variables sizes as a dictionary whose keys are the names of inputs and outputs and values are their new sizes. If a variable is missing, its original size is considered.

The ScalableModel parameters can also be filled in, otherwise the model uses default values.

Classes:

ScalableDiscipline(name, data[, sizes])

Scalable discipline.

class gemseo.problems.scalable.data_driven.discipline.ScalableDiscipline(name, data, sizes=None, **parameters)[source]

Scalable discipline.

input_grammar

The input grammar.

Type: AbstractGrammar

output_grammar

The output grammar.

Type: AbstractGrammar

grammar_type

The type of grammar to be used for inputs and outputs declaration.

Type: str

comp_dir

The path to the directory of the discipline module file if any.

Type: str

data_processor

A tool to pre- and post-process discipline data.

Type: DataProcessor

re_exec_policy

The policy to re-execute the same discipline.

Type: str

residual_variables

The output variables to be considered as residuals; they shall be equal to zero.

Type: List[str]

jac

The Jacobians of the outputs wrt inputs of the form {output: {input: matrix}}.

Type: Dict[str, Dict[str, ndarray]]

exec_for_lin

Whether the last execution was due to a linearization.

Type: bool

name

The name of the discipline.

Type: str

cache

The cache containing one or several executions of the discipline according to the cache policy.

Type: AbstractCache

local_data

The last input and output data.

Type: Dict[str, Any]

Constructor.

Parameters

name (str) – scalable model class name.
data (Dataset) – learning dataset.
sizes (dict) –
sizes of input and output variables. If None, use the original sizes. Default: None.

By default it is set to None.
parameters – model parameters
name – The name of the discipline. If None, use the class name.

Methods:

`activate_time_stamps`()	Activate the time stamps.
`add_differentiated_inputs`([inputs])	Add inputs against which to differentiate the outputs.
`add_differentiated_outputs`([outputs])	Add outputs to be differentiated.
`add_status_observer`(obs)	Add an observer for the status.
`auto_get_grammar_file`([is_input, name, comp_dir])	Use a naming convention to associate a grammar file to a discipline.
`check_input_data`(input_data[, raise_exception])	Check the input data validity.
`check_jacobian`([input_data, derr_approx, ...])	Check if the analytical Jacobian is correct with respect to a reference one.
`check_output_data`([raise_exception])	Check the output data validity.
`deactivate_time_stamps`()	Deactivate the time stamps.
`deserialize`(in_file)	Deserialize a discipline from a file.
`execute`([input_data])	Execute the discipline.
`get_all_inputs`()	Return the local input data as a list.
`get_all_outputs`()	Return the local output data as a list.
`get_attributes_to_serialize`()	Define the names of the attributes to be serialized.
`get_data_list_from_dict`(keys, data_dict)	Filter the dict from a list of keys or a single key.
`get_expected_dataflow`()	Return the expected data exchange sequence.
`get_expected_workflow`()	Return the expected execution sequence.
`get_input_data`()	Return the local input data as a dictionary.
`get_input_data_names`()	Return the names of the input variables.
`get_input_output_data_names`()	Return the names of the input and output variables.
`get_inputs_asarray`()	Return the local output data as a large NumPy array.
`get_inputs_by_name`(data_names)	Return the local data associated with input variables.
`get_local_data_by_name`(data_names)	Return the local data of the discipline associated with variables names.
`get_output_data`()	Return the local output data as a dictionary.
`get_output_data_names`()	Return the names of the output variables.
`get_outputs_asarray`()	Return the local input data as a large NumPy array.
`get_outputs_by_name`(data_names)	Return the local data associated with output variables.
`get_sub_disciplines`()	Return the sub-disciplines if any.
`initialize_grammars`(data)	Initialize input and output grammars from data names.
`is_all_inputs_existing`(data_names)	Test if several variables are discipline inputs.
`is_all_outputs_existing`(data_names)	Test if several variables are discipline outputs.
`is_input_existing`(data_name)	Test if a variable is a discipline input.
`is_output_existing`(data_name)	Test if a variable is a discipline output.
`is_scenario`()	Whether the discipline is a scenario.
`linearize`([input_data, force_all, force_no_exec])	Execute the linearized version of the code.
`notify_status_observers`()	Notify all status observers that the status has changed.
`remove_status_observer`(obs)	Remove an observer for the status.
`reset_statuses_for_run`()	Set all the statuses to `PENDING`.
`serialize`(out_file)	Serialize the discipline and store it in a file.
`set_cache_policy`([cache_type, ...])	Set the type of cache to use and the tolerance level.
`set_disciplines_statuses`(status)	Set the sub-disciplines statuses.
`set_jacobian_approximation`([...])	Set the Jacobian approximation method.
`set_optimal_fd_step`([outputs, inputs, ...])	Compute the optimal finite-difference step.
`store_local_data`(**kwargs)	Store discipline data in local data.

Attributes:

`cache_tol`	The cache input tolerance.
`default_inputs`	The default inputs.
`exec_time`	The cumulated execution time of the discipline.
`grammar_type`	The grammar type.
`linearization_mode`	The linearization mode among `LINEARIZE_MODE_LIST`.
`n_calls`	The number of times the discipline was executed.
`n_calls_linearize`	The number of times the discipline was linearized.
`status`	The status of the discipline.

classmethod activate_time_stamps()

Activate the time stamps.

For storing start and end times of execution and linearizations.

Return type: None

add_differentiated_inputs(inputs=None)

Add inputs against which to differentiate the outputs.

This method updates _differentiated_inputs with inputs.

Parameters

inputs (Optional[Iterable[str]]) –

The input variables against which to differentiate the outputs. If None, all the inputs of the discipline are used.

By default it is set to None.

Raises

ValueError – When the inputs wrt which differentiate the discipline are not inputs of the latter.

Return type

None

add_differentiated_outputs(outputs=None)

Add outputs to be differentiated.

This method updates _differentiated_outputs with outputs.

Parameters

outputs (Optional[Iterable[str]]) –

The output variables to be differentiated. If None, all the outputs of the discipline are used.

By default it is set to None.

Raises

ValueError – When the outputs to differentiate are not discipline outputs.

Return type

None

add_status_observer(obs)

Add an observer for the status.

Add an observer for the status to be notified when self changes of status.

Parameters: obs (Any) – The observer to add.
Return type: None

auto_get_grammar_file(is_input=True, name=None, comp_dir=None)

Use a naming convention to associate a grammar file to a discipline.

This method searches in a directory for either an input grammar file named name + "_input.json" or an output grammar file named``name + “_output.json”``.

Parameters

is_input (bool) –
If True, autodetect the input grammar file; otherwise, autodetect the output grammar file.

By default it is set to True.
name (Optional[str]) –
The name to be searched in the file names. If None, use the name name of the discipline.

By default it is set to None.
comp_dir (Optional[Union[str, pathlib.Path]]) –
The directory in which to search the grammar file. If None, use comp_dir.

By default it is set to None.

Returns

The grammar file path.

Return type

pathlib.Path

property cache_tol

The cache input tolerance.

This is the tolerance for equality of the inputs in the cache. If norm(stored_input_data-input_data) <= cache_tol * norm(stored_input_data), the cached data for stored_input_data is returned when calling self.execute(input_data).

check_input_data(input_data, raise_exception=True)

Check the input data validity.

Parameters

input_data (Dict[str, Any]) – The input data needed to execute the discipline according to the discipline input grammar.
raise_exception (bool) –

By default it is set to True.

Return type

None

check_jacobian(input_data=None, derr_approx='finite_differences', step=1e-07, threshold=1e-08, linearization_mode='auto', inputs=None, outputs=None, parallel=False, n_processes=2, use_threading=False, wait_time_between_fork=0, auto_set_step=False, plot_result=False, file_path='jacobian_errors.pdf', show=False, figsize_x=10, figsize_y=10, reference_jacobian_path=None, save_reference_jacobian=False, indices=None)

Check if the analytical Jacobian is correct with respect to a reference one.

If reference_jacobian_path is not None and save_reference_jacobian is True, compute the reference Jacobian with the approximation method and save it in reference_jacobian_path.

If reference_jacobian_path is not None and save_reference_jacobian is False, do not compute the reference Jacobian but read it from reference_jacobian_path.

If reference_jacobian_path is None, compute the reference Jacobian without saving it.

Parameters

input_data –
The input data needed to execute the discipline according to the discipline input grammar. If None, use the default_inputs.

By default it is set to None.
derr_approx –
The approximation method, either “complex_step” or “finite_differences”.

By default it is set to finite_differences.
threshold –
The acceptance threshold for the Jacobian error.

By default it is set to 1e-08.
linearization_mode –
the mode of linearization: direct, adjoint or automated switch depending on dimensions of inputs and outputs (Default value = ‘auto’)

By default it is set to auto.
inputs –
The names of the inputs wrt which to differentiate the outputs.

By default it is set to None.
outputs –
The names of the outputs to be differentiated.

By default it is set to None.
step –
The differentiation step.

By default it is set to 1e-07.
parallel –
Whether to differentiate the discipline in parallel.

By default it is set to False.
n_processes –
The maximum number of processors on which to run.

By default it is set to 2.
use_threading –
Whether to use threads instead of processes to parallelize the execution; multiprocessing will copy (serialize) all the disciplines, while threading will share all the memory This is important to note if you want to execute the same discipline multiple times, you shall use multiprocessing.

By default it is set to False.
wait_time_between_fork –
The time waited between two forks of the process / thread.

By default it is set to 0.
auto_set_step –
Whether to compute the optimal step for a forward first order finite differences gradient approximation.

By default it is set to False.
plot_result –
Whether to plot the result of the validation (computed vs approximated Jacobians).

By default it is set to False.
file_path –
The path to the output file if plot_result is True.

By default it is set to jacobian_errors.pdf.
show –
Whether to open the figure.

By default it is set to False.
figsize_x –
The x-size of the figure in inches.

By default it is set to 10.
figsize_y –
The y-size of the figure in inches.

By default it is set to 10.
reference_jacobian_path –
The path of the reference Jacobian file.

By default it is set to None.
save_reference_jacobian –
Whether to save the reference Jacobian.

By default it is set to False.
indices –
The indices of the inputs and outputs for the different sub-Jacobian matrices, formatted as {variable_name: variable_components} where variable_components can be either an integer, e.g. 2 a sequence of integers, e.g. [0, 3], a slice, e.g. slice(0,3), the ellipsis symbol (…) or None, which is the same as ellipsis. If a variable name is missing, consider all its components. If None, consider all the components of all the inputs and outputs.

By default it is set to None.

Returns

Whether the analytical Jacobian is correct with respect to the reference one.

check_output_data(raise_exception=True)

Check the output data validity.

Parameters

raise_exception (bool) –

Whether to raise an exception when the data is invalid.

By default it is set to True.

Return type

None

classmethod deactivate_time_stamps()

Deactivate the time stamps.

For storing start and end times of execution and linearizations.

Return type: None

property default_inputs

The default inputs.

Raises: TypeError – When the default inputs are not passed as a dictionary.

static deserialize(in_file)

Deserialize a discipline from a file.

Parameters: in_file (Union[str, pathlib.Path]) – The path to the file containing the discipline.
Returns: The discipline instance.
Return type: gemseo.core.discipline.MDODiscipline

property exec_time: The cumulated execution time of the discipline.

Note

This property is multiprocessing safe.

execute(input_data=None)

Execute the discipline.

This method executes the discipline:

Adds the default inputs to the input_data if some inputs are not defined in input_data but exist in _default_inputs.
Checks whether the last execution of the discipline was called with identical inputs, ie. cached in cache; if so, directly returns self.cache.get_output_cache(inputs).
Caches the inputs.
Checks the input data against input_grammar.
If data_processor is not None, runs the preprocessor.
Updates the status to RUNNING.
Calls the _run() method, that shall be defined.
If data_processor is not None, runs the postprocessor.
Checks the output data.
Caches the outputs.
Updates the status to DONE or FAILED.
Updates summed execution time.

Parameters

input_data (Optional[Dict[str, Any]]) –

The input data needed to execute the discipline according to the discipline input grammar. If None, use the default_inputs.

By default it is set to None.

Returns

The discipline local data after execution.

Return type

Dict[str, Any]

get_all_inputs()

Return the local input data as a list.

The order is given by get_input_data_names().

Returns: The local input data.
Return type: List[Any]

get_all_outputs()

Return the local output data as a list.

The order is given by get_output_data_names().

Returns: The local output data.
Return type: List[Any]

get_attributes_to_serialize()

Define the names of the attributes to be serialized.

Shall be overloaded by disciplines

Returns: The names of the attributes to be serialized.

static get_data_list_from_dict(keys, data_dict)

Filter the dict from a list of keys or a single key.

If keys is a string, then the method return the value associated to the key. If keys is a list of strings, then the method returns a generator of value corresponding to the keys which can be iterated.

Parameters

keys (Union[str, Iterable]) – One or several names.
data_dict (Dict[str, Any]) – The mapping from which to get the data.

Returns

Either a data or a generator of data.

Return type

Union[Any, Generator[Any]]

get_expected_dataflow()

Return the expected data exchange sequence.

This method is used for the XDSM representation.

The default expected data exchange sequence is an empty list.

See also

MDOFormulation.get_expected_dataflow

Returns: The data exchange arcs.
Return type: List[Tuple[gemseo.core.discipline.MDODiscipline, gemseo.core.discipline.MDODiscipline, List[str]]]

get_expected_workflow()

Return the expected execution sequence.

This method is used for the XDSM representation.

The default expected execution sequence is the execution of the discipline itself.

See also

MDOFormulation.get_expected_workflow

Returns: The expected execution sequence.
Return type: SerialExecSequence

get_input_data()

Return the local input data as a dictionary.

Returns: The local input data.
Return type: Dict[str, Any]

get_input_data_names()

Return the names of the input variables.

Returns: The names of the input variables.
Return type: List[str]

get_input_output_data_names()

Return the names of the input and output variables.

Returns: The name of the input and output variables.
Return type: List[str]

get_inputs_asarray()

Return the local output data as a large NumPy array.

The order is the one of get_all_outputs().

Returns: The local output data.
Return type: numpy.ndarray

get_inputs_by_name(data_names)

Return the local data associated with input variables.

Parameters: data_names (Iterable[str]) – The names of the input variables.
Returns: The local data for the given input variables.
Raises: ValueError – When a variable is not an input of the discipline.
Return type: List[Any]

get_local_data_by_name(data_names)

Return the local data of the discipline associated with variables names.

Parameters: data_names (Iterable[str]) – The names of the variables.
Returns: The local data associated with the variables names.
Raises: ValueError – When a name is not not a discipline input name.
Return type: Generator[Any]

get_output_data()

Return the local output data as a dictionary.

Returns: The local output data.
Return type: Dict[str, Any]

get_output_data_names()

Return the names of the output variables.

Returns: The names of the output variables.
Return type: List[str]

get_outputs_asarray()

Return the local input data as a large NumPy array.

The order is the one of get_all_inputs().

Returns: The local input data.
Return type: numpy.ndarray

get_outputs_by_name(data_names)

Return the local data associated with output variables.

Parameters: data_names (Iterable[str]) – The names of the output variables.
Returns: The local data for the given output variables.
Raises: ValueError – When a variable is not an output of the discipline.
Return type: List[Any]

get_sub_disciplines()

Return the sub-disciplines if any.

Returns: The sub-disciplines.
Return type: List[gemseo.core.discipline.MDODiscipline]

property grammar_type: The grammar type.

initialize_grammars(data)[source]

Initialize input and output grammars from data names.

Parameters: data (Dataset) – learning dataset.

is_all_inputs_existing(data_names)

Test if several variables are discipline inputs.

Parameters: data_names (Iterable[str]) – The names of the variables.
Returns: Whether all the variables are discipline inputs.
Return type: bool

is_all_outputs_existing(data_names)

Test if several variables are discipline outputs.

Parameters: data_names (Iterable[str]) – The names of the variables.
Returns: Whether all the variables are discipline outputs.
Return type: bool

is_input_existing(data_name)

Test if a variable is a discipline input.

Parameters: data_name (str) – The name of the variable.
Returns: Whether the variable is a discipline input.
Return type: bool

is_output_existing(data_name)

Test if a variable is a discipline output.

Parameters: data_name (str) – The name of the variable.
Returns: Whether the variable is a discipline output.
Return type: bool

static is_scenario()

Whether the discipline is a scenario.

Return type: bool

property linearization_mode

The linearization mode among LINEARIZE_MODE_LIST.

Raises: ValueError – When the linearization mode is unknown.

linearize(input_data=None, force_all=False, force_no_exec=False)

Execute the linearized version of the code.

Parameters

input_data (Optional[Dict[str, Any]]) –
The input data needed to linearize the discipline according to the discipline input grammar. If None, use the default_inputs.

By default it is set to None.
force_all (bool) –
If False, _differentiated_inputs and differentiated_output are used to filter the differentiated variables. otherwise, all outputs are differentiated wrt all inputs.

By default it is set to False.
force_no_exec (bool) –
If True, the discipline is not re executed, cache is loaded anyway.

By default it is set to False.

Returns

The Jacobian of the discipline.

Return type

Dict[str, Dict[str, numpy.ndarray]]

property n_calls: The number of times the discipline was executed.

Note

This property is multiprocessing safe.

property n_calls_linearize: The number of times the discipline was linearized.

Note

This property is multiprocessing safe.

notify_status_observers()

Notify all status observers that the status has changed.

Return type: None

remove_status_observer(obs)

Remove an observer for the status.

Parameters: obs (Any) – The observer to remove.
Return type: None

reset_statuses_for_run()

Set all the statuses to PENDING.

Raises: ValueError – When the discipline cannot be run because of its status.
Return type: None

serialize(out_file)

Serialize the discipline and store it in a file.

Parameters: out_file (Union[str, pathlib.Path]) – The path to the file to store the discipline.
Return type: None

set_cache_policy(cache_type='SimpleCache', cache_tolerance=0.0, cache_hdf_file=None, cache_hdf_node_name=None, is_memory_shared=True)

Set the type of cache to use and the tolerance level.

This method defines when the output data have to be cached according to the distance between the corresponding input data and the input data already cached for which output data are also cached.

The cache can be either a SimpleCache recording the last execution or a cache storing all executions, e.g. MemoryFullCache and HDF5Cache. Caching data can be either in-memory, e.g. SimpleCache and MemoryFullCache, or on the disk, e.g. HDF5Cache.

The attribute CacheFactory.caches provides the available caches types.

Parameters

cache_type (str) –
The type of cache.

By default it is set to SimpleCache.
cache_tolerance (float) –
The maximum relative norm of the difference between two input arrays to consider that two input arrays are equal.

By default it is set to 0.0.
cache_hdf_file (Optional[Union[str, pathlib.Path]]) –
The path to the HDF file to store the data; this argument is mandatory when the HDF5Cache policy is used.

By default it is set to None.
cache_hdf_node_name (Optional[str]) –
The name of the HDF file node to store the discipline data. If None, name is used.

By default it is set to None.
is_memory_shared (bool) –
Whether to store the data with a shared memory dictionary, which makes the cache compatible with multiprocessing.

By default it is set to True.

Return type

None

set_disciplines_statuses(status)

Set the sub-disciplines statuses.

To be implemented in subclasses.

Parameters: status (str) – The status.
Return type: None

set_jacobian_approximation(jac_approx_type='finite_differences', jax_approx_step=1e-07, jac_approx_n_processes=1, jac_approx_use_threading=False, jac_approx_wait_time=0)

Set the Jacobian approximation method.

Sets the linearization mode to approx_method, sets the parameters of the approximation for further use when calling linearize().

Parameters

jac_approx_type (str) –
The approximation method, either “complex_step” or “finite_differences”.

By default it is set to finite_differences.
jax_approx_step (float) –
The differentiation step.

By default it is set to 1e-07.
jac_approx_n_processes (int) –
The maximum number of processors on which to run.

By default it is set to 1.
jac_approx_use_threading (bool) –
Whether to use threads instead of processes to parallelize the execution; multiprocessing will copy (serialize) all the disciplines, while threading will share all the memory This is important to note if you want to execute the same discipline multiple times, you shall use multiprocessing.

By default it is set to False.
jac_approx_wait_time (float) –
The time waited between two forks of the process / thread.

By default it is set to 0.

Return type

None

set_optimal_fd_step(outputs=None, inputs=None, force_all=False, print_errors=False, numerical_error=2.220446049250313e-16)

Compute the optimal finite-difference step.

Compute the optimal step for a forward first order finite differences gradient approximation. Requires a first evaluation of the perturbed functions values. The optimal step is reached when the truncation error (cut in the Taylor development), and the numerical cancellation errors (roundoff when doing f(x+step)-f(x))

are approximately equal.

Warning

This calls the discipline execution twice per input variables.

See also

https://en.wikipedia.org/wiki/Numerical_differentiation and “Numerical Algorithms and Digital Representation”, Knut Morken , Chapter 11, “Numerical Differenciation”

Parameters

inputs –
The inputs wrt which the outputs are linearized. If None, use the _differentiated_inputs.

By default it is set to None.
outputs –
The outputs to be linearized. If None, use the _differentiated_outputs.

By default it is set to None.
force_all –
Whether to consider all the inputs and outputs of the discipline;

By default it is set to False.
print_errors –
Whether to display the estimated errors.

By default it is set to False.
numerical_error –
The numerical error associated to the calculation of f. By default, this is the machine epsilon (appx 1e-16), but can be higher when the calculation of f requires a numerical resolution.

By default it is set to 2.220446049250313e-16.

Returns

The estimated errors of truncation and cancellation error.

Raises

ValueError – When the Jacobian approximation method has not been set.

property status: The status of the discipline.

store_local_data(**kwargs)

Store discipline data in local data.

Parameters

kwargs – The data to be stored in local_data.
**kwargs (Any) –

Return type

None

Scalable model factory¶

This module contains the ScalableModelFactory which is a factory to create a ScalableModel from its class name by means of the ScalableModelFactory.create() method. It is also possible to get a list of available scalable models (see ScalableModelFactory.scalable_models method) and to check is a type of scalable model is available (see ScalableModelFactory.is_available() method)

Classes:

ScalableModelFactory()

This factory instantiates a class:.ScalableModel from its class name.

class gemseo.problems.scalable.data_driven.factory.ScalableModelFactory[source]

This factory instantiates a class:.ScalableModel from its class name.

The class can be internal to GEMSEO or located in an external module whose path is provided to the constructor.

Initializes the factory: scans the directories to search for subclasses of ScalableModel.

Searches in “GEMSEO_PATH” and gemseo.caches

Methods:

`create`(model_name, data[, sizes])	Create a scalable model.
`is_available`(model_name)	Checks the availability of a scalable model.

Attributes:

scalable_models

Lists the available classes for scalable models.

create(model_name, data, sizes=None, **parameters)[source]

Create a scalable model.

Parameters

model_name (str) – name of the scalable model (its classname)
data (Dataset) – learning dataset.
sizes (dict) –
sizes of input and output variables. If None, use the original sizes. Default: None.

By default it is set to None.
parameters – model parameters

Returns

model_name scalable model

is_available(model_name)[source]

Checks the availability of a scalable model.

Parameters: model_name (str) – model_name of the scalable model.
Returns: True if the scalable model is available.
Return type: bool

property scalable_models

Lists the available classes for scalable models.

Returns: the list of classes names.
Return type: list(str)

Scalable model¶

This module implements the abstract concept of scalable model which is used by scalable disciplines. A scalable model is built from a input-output learning dataset associated with a function and generalizing its behavior to a new user-defined problem dimension, that is to say new user-defined input and output dimensions.

The concept of scalable model is implemented through ScalableModel, an abstract class which is instantiated from:

data provided as a Dataset
variables sizes provided as a dictionary whose keys are the names of inputs and outputs and values are their new sizes. If a variable is missing, its original size is considered.

Scalable model parameters can also be filled in. Otherwise the model uses default values.

See also

The ScalableDiagonalModel class overloads ScalableModel.

Classes:

ScalableModel(data[, sizes])

Scalable model.

class gemseo.problems.scalable.data_driven.model.ScalableModel(data, sizes=None, **parameters)[source]

Scalable model.

Constructor.

Parameters

data (Dataset) – learning dataset.
sizes (dict) –
sizes of input and output variables. If None, use the original sizes. Default: None.

By default it is set to None.
parameters – model parameters

Methods:

`build_model`()	Build model with original sizes for input and output variables.
`compute_bounds`()	Compute lower and upper bounds of both input and output variables.
`normalize_data`()	Normalize dataset from lower and upper bounds.
`scalable_derivatives`([input_value])	Evaluate the scalable derivatives.
`scalable_function`([input_value])	Evaluate the scalable function.

Attributes:

`inputs_names`	Inputs names.
`original_sizes`	Original sizes of variables.
`outputs_names`	Outputs names.

build_model()[source]

Build model with original sizes for input and output variables.

compute_bounds()[source]

Compute lower and upper bounds of both input and output variables.

Returns: lower bounds, upper bounds.
Return type: dict, dict

property inputs_names

Inputs names.

Returns: names of the inputs.
Return type: list(str)

normalize_data()[source]

Normalize dataset from lower and upper bounds.

property original_sizes

Original sizes of variables.

Returns: original sizes of variables.
Return type: dict

property outputs_names

Outputs names.

Returns: names of the outputs.
Return type: list(str)

scalable_derivatives(input_value=None)[source]

Evaluate the scalable derivatives.

Parameters

input_value (dict) –

input values. If None, use default inputs. Default: None

By default it is set to None.

Returns

evaluation of the scalable derivatives.

Return type

dict

scalable_function(input_value=None)[source]

Evaluate the scalable function.

Parameters

input_value (dict) –

input values. If None, use default inputs. Default: None.

By default it is set to None.

Returns

evaluation of the scalable function.

Return type

dict

Scalable diagonal model¶

This module implements the concept of scalable diagonal model, which is a particular scalable model built from an input-output dataset relying on a diagonal design of experiments (DOE) where inputs vary proportionally from their lower bounds to their upper bounds, following the diagonal of the input space.

So for every output, the dataset catches its evolution with respect to this proportion, which makes it a monodimensional behavior. Then, for a new user-defined problem dimension, the scalable model extrapolates this monodimensional behavior to the different input directions.

The concept of scalable diagonal model is implemented through the ScalableDiagonalModel class which is composed of a ScalableDiagonalApproximation. With regard to the diagonal DOE, GEMSEO proposes the DiagonalDOE class.

Classes:

`ScalableDiagonalApproximation`(sizes, ...[, seed])	Methodology that captures the trends of a physical problem, and extends it into a problem that has scalable input and outputs dimensions The original and the resulting scalable problem have the same interface:
`ScalableDiagonalModel`(data[, sizes, ...])	Scalable diagonal model.

Functions:

`choice`(a[, size, replace, p])	Generates a random sample from a given 1-D array
`npseed`	seed(self, seed=None)
`rand`(d0, d1, ..., dn)	Random values in a given shape.
`randint`(low[, high, size, dtype])	Return random integers from low (inclusive) to high (exclusive).

class gemseo.problems.scalable.data_driven.diagonal.ScalableDiagonalApproximation(sizes, output_dependency, io_dependency, seed=0)[source]

Methodology that captures the trends of a physical problem, and extends it into a problem that has scalable input and outputs dimensions The original and the resulting scalable problem have the same interface:

all inputs and outputs have the same names; only their dimensions vary.

Constructor:

Parameters

sizes (dict) – sizes of both input and output variables.
output_dependency (dict) – dependency between old and new outputs.
io_dependency (dict) – dependency between new inputs and new outputs.

Methods:

`build_scalable_function`(function_name, ...)	Build interpolation interpolation from a 1D input and output function.
`get_scalable_derivative`(output_function)	Retrieve the (scalable) gradient of the scalable function generated from the original discipline.
`get_scalable_function`(output_function)	Retrieve the scalable function generated from the original discipline.
`scale_samples`(samples)	Scale samples of array into [0, 1]

build_scalable_function(function_name, dataset, input_names, degree=3)[source]

Build interpolation interpolation from a 1D input and output function. Add the model to the local dictionary.

Parameters

function_name (str) – name of the output function
dataset (Dataset) – the input-output dataset
input_names (list(str)) – names of the input variables
degree (int) –
degree of interpolation (Default value = 3)

By default it is set to 3.

get_scalable_derivative(output_function)[source]

Retrieve the (scalable) gradient of the scalable function generated from the original discipline.

Parameters: output_function (str) – name of the output function

get_scalable_function(output_function)[source]

Retrieve the scalable function generated from the original discipline.

Parameters: output_function (str) – name of the output function

static scale_samples(samples)[source]

Scale samples of array into [0, 1]

Parameters: samples (list(array)) – samples of multivariate array
Returns: samples of multivariate array
Return type: array

class gemseo.problems.scalable.data_driven.diagonal.ScalableDiagonalModel(data, sizes=None, fill_factor=- 1, comp_dep=None, inpt_dep=None, force_input_dependency=False, allow_unused_inputs=True, seed=1, group_dep=None)[source]

Scalable diagonal model.

Constructor.

Parameters

data (Dataset) – learning dataset.
sizes (dict) –
sizes of input and output variables. If None, use the original sizes. Default: None.

By default it is set to None.
fill_factor –
degree of sparsity of the dependency matrix. Default: -1.

By default it is set to -1.
comp_dep –
matrix that establishes the selection of a single original component for each scalable component

By default it is set to None.
inpt_dep –
dependency matrix that establishes the dependency of outputs wrt inputs

By default it is set to None.
force_input_dependency (bool) –
for any output, force dependency with at least on input.

By default it is set to False.
allow_unused_inputs (bool) –
possibility to have an input with no dependence with any output

By default it is set to True.
seed (int) –
seed

By default it is set to 1.
group_dep (dict(list(str))) –
dependency between inputs and outputs

By default it is set to None.

Methods:

`build_model`()	Build model with original sizes for input and output variables.
`compute_bounds`()	Compute lower and upper bounds of both input and output variables.
`generate_random_dependency`()	Generates a random dependency structure for use in scalable discipline.
`normalize_data`()	Normalize dataset from lower and upper bounds.
`plot_1d_interpolations`([save, show, step, ...])	This methods plots the scaled 1D interpolations, a.k.a.
`plot_dependency`([add_levels, save, show, ...])	This method plots the dependency matrix of a discipline in the form of a chessboard, where rows represent inputs, columns represent output and gray scale represent the dependency level between inputs and outputs.
`scalable_derivatives`([input_value])	Evaluate the scalable derivatives.
`scalable_function`([input_value])	Evaluate the scalable functions.

Attributes:

`inputs_names`	Inputs names.
`original_sizes`	Original sizes of variables.
`outputs_names`	Outputs names.

build_model()[source]

Build model with original sizes for input and output variables.

Returns: scalable approximation.
Return type: ScalableDiagonalApproximation

compute_bounds()

Compute lower and upper bounds of both input and output variables.

Returns: lower bounds, upper bounds.
Return type: dict, dict

generate_random_dependency()[source]

Generates a random dependency structure for use in scalable discipline.

Returns: output component dependency and input-output dependency
Return type: dict(int), dict(dict(array))

property inputs_names

Inputs names.

Returns: names of the inputs.
Return type: list(str)

normalize_data()

Normalize dataset from lower and upper bounds.

property original_sizes

Original sizes of variables.

Returns: original sizes of variables.
Return type: dict

property outputs_names

Outputs names.

Returns: names of the outputs.
Return type: list(str)

plot_1d_interpolations(save=False, show=False, step=0.01, varnames=None, directory='.', png=False)[source]

This methods plots the scaled 1D interpolations, a.k.a. basis functions.

A basis function is a monodimensional function interpolating the samples of a given output component over the input sampling line \(t\in[0,1]\mapsto \\underline{x}+t(\overline{x}-\\underline{x})\).

There are as many basis functions as there are output components from the discipline. Thus, for a discipline with a single output in dimension 1, there is 1 basis function. For a discipline with a single output in dimension 2, there are 2 basis functions. For a discipline with an output in dimension 2 and an output in dimension 13, there are 15 basis functions. And so on. This method allows to plot the basis functions associated with all outputs or only part of them, either on screen (show=True), in a file (save=True) or both. We can also specify the discretization step whose default value is 0.01.

Parameters

save (bool) –
if True, export the plot as a PDF file (Default value = False)

By default it is set to False.
show (bool) –
if True, display the plot (Default value = False)

By default it is set to False.
step (bool) –
Step to evaluate the 1d interpolation function (Default value = 0.01)

By default it is set to 0.01.
varnames (list(str)) –
names of the variable to plot; if None, all variables are plotted (Default value = None)

By default it is set to None.
directory (str) –
directory path. Default: ‘.’.

By default it is set to ..
png (bool) –
if True, the file format is PNG. Otherwise, use PDF. Default: False.

By default it is set to False.

plot_dependency(add_levels=True, save=True, show=False, directory='.', png=False)[source]

This method plots the dependency matrix of a discipline in the form of a chessboard, where rows represent inputs, columns represent output and gray scale represent the dependency level between inputs and outputs.

Parameters

add_levels (bool) –
add values of dependency levels in percentage. Default: True.

By default it is set to True.
save (bool) –
if True, export the plot into a file. Default: True.

By default it is set to True.
show (bool) –
if True, display the plot. Default: False.

By default it is set to False.
directory (str) –
directory path. Default: ‘.’.

By default it is set to ..
png (bool) –
if True, the file format is PNG. Otherwise, use PDF. Default: False.

By default it is set to False.

scalable_derivatives(input_value=None)[source]

Evaluate the scalable derivatives.

Parameters

input_value (dict) –

input values. If None, use default inputs.

By default it is set to None.

Returns

evaluation of the scalable derivatives.

Return type

dict

scalable_function(input_value=None)[source]

Evaluate the scalable functions.

Parameters

input_value (dict) –

input values. If None, use default inputs.

By default it is set to None.

Returns

evaluation of the scalable functions.

Return type

dict

gemseo.problems.scalable.data_driven.diagonal.choice(a, size=None, replace=True, p=None)

Generates a random sample from a given 1-D array

New in version 1.7.0.

Note

New code should use the choice method of a default_rng() instance instead; please see the random-quick-start.

Parameters

a (1-D array-like or int) – If an ndarray, a random sample is generated from its elements. If an int, the random sample is generated as if a were np.arange(a)
size (int or tuple of ints, optional) – Output shape. If the given shape is, e.g., (m, n, k), then m * n * k samples are drawn. Default is None, in which case a single value is returned.
replace (boolean, optional) – Whether the sample is with or without replacement
p (1-D array-like, optional) – The probabilities associated with each entry in a. If not given the sample assumes a uniform distribution over all entries in a.

Returns

samples – The generated random samples

Return type

single item or ndarray

Raises

ValueError – If a is an int and less than zero, if a or p are not 1-dimensional, if a is an array-like of size 0, if p is not a vector of probabilities, if a and p have different lengths, or if replace=False and the sample size is greater than the population size

See also

randint, shuffle, permutation

Generator.choice: which should be used in new code

Notes

Sampling random rows from a 2-D array is not possible with this function, but is possible with Generator.choice through its axis keyword.

Examples

Generate a uniform random sample from np.arange(5) of size 3:

>>> np.random.choice(5, 3)
array([0, 3, 4]) # random
>>> #This is equivalent to np.random.randint(0,5,3)

Generate a non-uniform random sample from np.arange(5) of size 3:

>>> np.random.choice(5, 3, p=[0.1, 0, 0.3, 0.6, 0])
array([3, 3, 0]) # random

Generate a uniform random sample from np.arange(5) of size 3 without replacement:

>>> np.random.choice(5, 3, replace=False)
array([3,1,0]) # random
>>> #This is equivalent to np.random.permutation(np.arange(5))[:3]

Generate a non-uniform random sample from np.arange(5) of size 3 without replacement:

>>> np.random.choice(5, 3, replace=False, p=[0.1, 0, 0.3, 0.6, 0])
array([2, 3, 0]) # random

Any of the above can be repeated with an arbitrary array-like instead of just integers. For instance:

>>> aa_milne_arr = ['pooh', 'rabbit', 'piglet', 'Christopher']
>>> np.random.choice(aa_milne_arr, 5, p=[0.5, 0.1, 0.1, 0.3])
array(['pooh', 'pooh', 'pooh', 'Christopher', 'piglet'], # random
      dtype='<U11')

gemseo.problems.scalable.data_driven.diagonal.npseed()

seed(self, seed=None)

Reseed a legacy MT19937 BitGenerator

Notes

This is a convenience, legacy function.

The best practice is to not reseed a BitGenerator, rather to recreate a new one. This method is here for legacy reasons. This example demonstrates best practice.

>>> from numpy.random import MT19937
>>> from numpy.random import RandomState, SeedSequence
>>> rs = RandomState(MT19937(SeedSequence(123456789)))
# Later, you want to restart the stream
>>> rs = RandomState(MT19937(SeedSequence(987654321)))

gemseo.problems.scalable.data_driven.diagonal.rand(d0, d1, ..., dn)

Random values in a given shape.

Note

This is a convenience function for users porting code from Matlab, and wraps random_sample. That function takes a tuple to specify the size of the output, which is consistent with other NumPy functions like numpy.zeros and numpy.ones.

Create an array of the given shape and populate it with random samples from a uniform distribution over [0, 1).

Parameters

d0 (int, optional) – The dimensions of the returned array, must be non-negative. If no argument is given a single Python float is returned.
d1 (int, optional) – The dimensions of the returned array, must be non-negative. If no argument is given a single Python float is returned.
... (int, optional) – The dimensions of the returned array, must be non-negative. If no argument is given a single Python float is returned.
dn (int, optional) – The dimensions of the returned array, must be non-negative. If no argument is given a single Python float is returned.

Returns

out – Random values.

Return type

ndarray, shape (d0, d1, ..., dn)

See also

random

Examples

>>> np.random.rand(3,2)
array([[ 0.14022471,  0.96360618],  #random
       [ 0.37601032,  0.25528411],  #random
       [ 0.49313049,  0.94909878]]) #random

gemseo.problems.scalable.data_driven.diagonal.randint(low, high=None, size=None, dtype=int)

Return random integers from low (inclusive) to high (exclusive).

Return random integers from the “discrete uniform” distribution of the specified dtype in the “half-open” interval [low, high). If high is None (the default), then results are from [0, low).

Note

New code should use the integers method of a default_rng() instance instead; please see the random-quick-start.

Parameters

low (int or array-like of ints) – Lowest (signed) integers to be drawn from the distribution (unless high=None, in which case this parameter is one above the highest such integer).
high (int or array-like of ints, optional) – If provided, one above the largest (signed) integer to be drawn from the distribution (see above for behavior if high=None). If array-like, must contain integer values
size (int or tuple of ints, optional) – Output shape. If the given shape is, e.g., (m, n, k), then m * n * k samples are drawn. Default is None, in which case a single value is returned.
dtype (dtype, optional) –
Desired dtype of the result. Byteorder must be native. The default value is int.

New in version 1.11.0.

Returns

out – size-shaped array of random integers from the appropriate distribution, or a single such random int if size not provided.

Return type

int or ndarray of ints

See also

random_integers: similar to randint, only for the closed interval [low, high], and 1 is the lowest value if high is omitted.
Generator.integers: which should be used for new code.

Examples

>>> np.random.randint(2, size=10)
array([1, 0, 0, 0, 1, 1, 0, 0, 1, 0]) # random
>>> np.random.randint(1, size=10)
array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0])

Generate a 2 x 4 array of ints between 0 and 4, inclusive:

>>> np.random.randint(5, size=(2, 4))
array([[4, 0, 2, 1], # random
       [3, 2, 2, 0]])

Generate a 1 x 3 array with 3 different upper bounds

>>> np.random.randint(1, [3, 5, 10])
array([2, 2, 9]) # random

Generate a 1 by 3 array with 3 different lower bounds

>>> np.random.randint([1, 5, 7], 10)
array([9, 8, 7]) # random

Generate a 2 by 4 array using broadcasting with dtype of uint8

>>> np.random.randint([1, 3, 5, 7], [[10], [20]], dtype=np.uint8)
array([[ 8,  6,  9,  7], # random
       [ 1, 16,  9, 12]], dtype=uint8)