# Copyright 2021 IRT Saint Exupéry, https://www.irt-saintexupery.com
#
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU Lesser General Public
# License version 3 as published by the Free Software Foundation.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
# Lesser General Public License for more details.
#
# You should have received a copy of the GNU Lesser General Public License
# along with this program; if not, write to the Free Software Foundation,
# Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
# Contributors:
# INITIAL AUTHORS - initial API and implementation and/or initial
# documentation
# :author: Francois Gallard
# OTHER AUTHORS - MACROSCOPIC CHANGES
"""Abstraction of processes."""
from __future__ import annotations
import collections
import logging
import pickle
import sys
from collections import defaultdict
from copy import deepcopy
from multiprocessing import Manager
from multiprocessing import Value
from multiprocessing import cpu_count
from pathlib import Path
from timeit import default_timer as timer
from typing import TYPE_CHECKING
from typing import Any
from typing import ClassVar
from typing import Final
from numpy import concatenate
from numpy import empty
from numpy import ndarray
from numpy import zeros
from scipy.sparse import csr_array
from strenum import StrEnum
from gemseo.caches.cache_factory import CacheFactory
from gemseo.caches.simple_cache import SimpleCache
from gemseo.core.derivatives.derivation_modes import DerivationMode
from gemseo.core.discipline_data import DisciplineData
from gemseo.core.grammars.factory import GrammarFactory
from gemseo.core.namespaces import remove_prefix_from_list
from gemseo.core.serializable import Serializable
from gemseo.disciplines.utils import get_sub_disciplines
from gemseo.utils.derivatives.approximation_modes import ApproximationMode
from gemseo.utils.derivatives.derivatives_approx import EPSILON
from gemseo.utils.derivatives.derivatives_approx import DisciplineJacApprox
from gemseo.utils.enumeration import merge_enums
from gemseo.utils.multiprocessing import get_multi_processing_manager
from gemseo.utils.string_tools import MultiLineString
from gemseo.utils.string_tools import pretty_str
if TYPE_CHECKING:
from collections.abc import Iterable
from collections.abc import Iterator
from collections.abc import Mapping
from collections.abc import MutableMapping
from numpy.typing import NDArray
from gemseo.core.cache import AbstractCache
from gemseo.core.data_processor import DataProcessor
from gemseo.core.derivatives.jacobian_operator import JacobianOperator
from gemseo.core.execution_sequence import AtomicExecSequence
from gemseo.core.grammars.base_grammar import BaseGrammar
from gemseo.core.grammars.defaults import Defaults
LOGGER = logging.getLogger(__name__)
[docs]
def default_dict_factory() -> dict:
"""Instantiate a defaultdict(None) object."""
return defaultdict(None)
[docs]
class MDODiscipline(Serializable):
"""A software integrated in the workflow.
To be used,
subclass :class:`.MDODiscipline`
and implement the :meth:`._run` method which defines the execution of the software.
Typically,
:meth:`._run` gets the input data stored in the :attr:`.local_data`,
passes them to the callable computing the output data, e.g. a software,
and stores these output data in the :attr:`.local_data`.
Then,
the end-user calls the :meth:`.execute` method
with optional ``input_data``;
if not,
:attr:`.default_inputs` are used.
This :meth:`.execute` method uses name grammars
to check the variable names and types of
both the passed input data before calling :meth:`.run`
and the returned output data before they are stored in the :attr:`.cache`.
A grammar can be either a :class:`.SimpleGrammar` or a :class:`.JSONGrammar`,
or your own which derives from :class:`.BaseGrammar`.
"""
[docs]
class ExecutionStatus(StrEnum):
"""The execution statuses of a discipline."""
VIRTUAL = "VIRTUAL"
PENDING = "PENDING"
DONE = "DONE"
RUNNING = "RUNNING"
FAILED = "FAILED"
LINEARIZE = "LINEARIZE"
[docs]
class GrammarType(StrEnum):
"""The name of the grammar class."""
JSON = "JSONGrammar"
SIMPLE = "SimpleGrammar"
SIMPLER = "SimplerGrammar"
PYDANTIC = "PydanticGrammar"
[docs]
class CacheType(StrEnum):
"""The name of the cache class."""
SIMPLE = "SimpleCache"
HDF5 = "HDF5Cache"
MEMORY_FULL = "MemoryFullCache"
NONE = ""
"""No cache is used."""
[docs]
class InitJacobianType(StrEnum):
"""The way to initialize Jacobian matrices."""
EMPTY = "empty"
"""The Jacobian is initialized as an empty NumPy ndarray."""
DENSE = "dense"
"""The Jacobian is initialized as a NumPy ndarray filled in with zeros."""
SPARSE = "sparse"
"""The Jacobian is initialized as a SciPy CSR array with zero elements."""
ApproximationMode = ApproximationMode
LinearizationMode = merge_enums(
"LinearizationMode",
StrEnum,
DerivationMode,
ApproximationMode,
)
[docs]
class ReExecutionPolicy(StrEnum):
"""The re-execution policy of a discipline."""
DONE = "RE_EXEC_DONE"
NEVER = "RE_EXEC_NEVER"
input_grammar: BaseGrammar
"""The input grammar."""
output_grammar: BaseGrammar
"""The output grammar."""
data_processor: DataProcessor
"""A tool to pre- and post-process discipline data."""
residual_variables: dict[str, str]
"""The output variables mapping to their inputs, to be considered as residuals; they
shall be equal to zero."""
run_solves_residuals: bool
"""Whether the run method shall solve the residuals."""
jac: MutableMapping[
str, MutableMapping[str, ndarray | csr_array | JacobianOperator]
]
"""The Jacobians of the outputs wrt inputs.
The structure is ``{output: {input: matrix}}``.
"""
exec_for_lin: bool
"""Whether the last execution was due to a linearization."""
name: str
"""The name of the discipline."""
__linear_relationships: Mapping[str, Iterable[str]]
"""The linear relationships between discipline inputs and outputs.
For each output define one or many inputs that in linear relationship with it.
"""
cache: AbstractCache | None
"""The cache containing one or several executions of the discipline according to the
cache policy."""
activate_counters: ClassVar[bool] = True
"""Whether to activate the counters (execution time, calls and linearizations)."""
activate_input_data_check: ClassVar[bool] = True
"""Whether to check the input data respect the input grammar."""
activate_output_data_check: ClassVar[bool] = True
"""Whether to check the output data respect the output grammar."""
activate_cache: bool = True
"""Whether to cache the discipline evaluations by default."""
re_exec_policy: ReExecutionPolicy
"""The policy to re-execute the same discipline."""
GRAMMAR_DIRECTORY: ClassVar[str | None] = None
"""The directory in which to search for the grammar files if not the class one."""
virtual_execution: ClassVar[bool] = False
"""Whether to skip the :meth:`._run` method during execution and return the
:attr:`.default_outputs`, whatever the inputs."""
N_CPUS: Final[int] = cpu_count()
"""The number of available CPUs."""
_ATTR_NOT_TO_SERIALIZE: ClassVar[set[str]] = {
"_status_observers",
}
__mp_manager: Manager | None = None
time_stamps: ClassVar[dict[str, float] | None] = None
"""The mapping from discipline name to their execution time."""
def __init__(
self,
name: str | None = None,
input_grammar_file: str | Path | None = None,
output_grammar_file: str | Path | None = None,
auto_detect_grammar_files: bool = False,
grammar_type: GrammarType = GrammarType.JSON,
cache_type: CacheType = CacheType.SIMPLE,
cache_file_path: str | Path | None = None,
) -> None:
"""
Args:
name: The name of the discipline.
If ``None``, use the class name.
input_grammar_file: The input grammar file path.
If ``None`` and ``auto_detect_grammar_files=True``,
look for ``"ClassName_input.json"``
in the :attr:`.GRAMMAR_DIRECTORY` if any
or in the directory of the discipline class module.
If ``None`` and ``auto_detect_grammar_files=False``,
do not initialize the input grammar from a schema file.
output_grammar_file: The output grammar file path.
If ``None`` and ``auto_detect_grammar_files=True``,
look for ``"ClassName_output.json"``
in the :attr:`.GRAMMAR_DIRECTORY` if any
or in the directory of the discipline class module.
If ``None`` and ``auto_detect_grammar_files=False``,
do not initialize the output grammar from a schema file.
auto_detect_grammar_files: Whether to
look for ``"ClassName_{input,output}.json"``
in the :attr:`.GRAMMAR_DIRECTORY` if any
or in the directory of the discipline class module
when ``{input,output}_grammar_file`` is ``None``.
grammar_type: The type of the input and output grammars.
cache_type: The type of cache.
cache_file_path: The HDF file path
when ``grammar_type`` is :attr:`.MDODiscipline.CacheType.HDF5`.
""" # noqa: D205, D212, D415
self.data_processor = None
self.input_grammar = None
self.output_grammar = None
# Allow to re-execute the same discipline twice, only if did not fail
# and not running
self.re_exec_policy = self.ReExecutionPolicy.DONE
# : list of outputs that shall be null, to be considered as residuals
self.residual_variables = {}
self._disciplines = []
self.run_solves_residuals = False
self._differentiated_inputs = [] # : inputs to differentiate
# : outputs to be used for differentiation
self._differentiated_outputs = []
self._n_calls = None # : number of calls to execute()
self._exec_time = None # : cumulated execution time
# : number of calls to linearize()
self._n_calls_linearize = None
self._in_data_hash_dict = {}
self.jac = {} # : Jacobians of outputs wrt inputs dictionary
# : True if linearize() has already been called
self._is_linearized = False
self._jac_approx = None # Jacobian's approximation object
self._linearize_on_last_state = False # If true, the linearization
# is performed on the state computed by the disciplines
# (MDAs for instance) otherwise, the inputs that are also outputs
# are reset at the previous inputs value be fore calling
# _compute_jacobian
# the last execution was due to a linearization
self.exec_for_lin = False
if name is None:
self.name = self.__class__.__name__
else:
self.name = name
self.cache = None
if not self.activate_cache:
cache_type = self.CacheType.NONE
if cache_type is not self.CacheType.NONE:
self.cache = self.__create_new_cache(
cache_type, hdf_file_path=cache_file_path, hdf_node_path=self.name
)
self._cache_was_loaded = False
self._linearization_mode = self.LinearizationMode.AUTO
self._grammar_type = grammar_type
if auto_detect_grammar_files:
if input_grammar_file is None:
input_grammar_file = self.auto_get_grammar_file()
if output_grammar_file is None:
output_grammar_file = self.auto_get_grammar_file(is_input=False)
self._instantiate_grammars(
input_grammar_file, output_grammar_file, self._grammar_type
)
self._local_data = None
self.__set_local_data({})
self._status = self.ExecutionStatus.PENDING
if self.activate_counters:
self._init_shared_memory_attrs()
self._status_observers = []
self.__linear_relationships = {}
def __str__(self) -> str:
return self.name
@property
def linear_relationships(self) -> Mapping[str, Iterable[str]]:
"""The linear relationships between inputs and outputs."""
return self.__linear_relationships
[docs]
def set_linear_relationships(
self,
outputs: Iterable[str] = (),
inputs: Iterable[str] = (),
) -> None:
"""Set linear relationships between discipline inputs and outputs.
Args:
outputs: The discipline output(s) in a linear relation with the input(s).
If empty, all discipline outputs are considered.
inputs: The discipline input(s) in a linear relation with the output(s).
If empty, all discipline inputs are considered.
"""
if not outputs:
outputs = self.get_output_data_names()
if not inputs:
inputs = self.get_input_data_names()
for out in outputs:
self.__linear_relationships[out] = inputs
@property
def _string_representation(self) -> MultiLineString:
"""The string representation of the object."""
mls = MultiLineString()
mls.add(self.name)
mls.indent()
mls.add("Inputs: {}", pretty_str(self.get_input_data_names()))
mls.add("Outputs: {}", pretty_str(self.get_output_data_names()))
return mls
def __repr__(self) -> str:
return str(self._string_representation)
def _repr_html_(self) -> str:
return self._string_representation._repr_html_()
def _init_shared_memory_attrs(self) -> None:
self._n_calls = Value("i", 0)
self._exec_time = Value("d", 0.0)
self._n_calls_linearize = Value("i", 0)
@property
def disciplines(self) -> list[MDODiscipline]:
"""The sub-disciplines, if any."""
return self._disciplines
@property
def local_data(self) -> DisciplineData:
"""The current input and output data."""
return self._local_data
@local_data.setter
def local_data(self, data: MutableMapping[str, Any]) -> None:
self.__set_local_data(data)
def __set_local_data(self, data: MutableMapping[str, Any]) -> None:
self._local_data = DisciplineData(
data,
input_to_namespaced=self.input_grammar.to_namespaced,
output_to_namespaced=self.output_grammar.to_namespaced,
)
@property
def n_calls(self) -> int | None:
"""The number of times the discipline was executed.
This property is multiprocessing safe.
Raises:
RuntimeError: When the discipline counters are disabled.
"""
if self.activate_counters:
return self._n_calls.value
return None
@n_calls.setter
def n_calls(
self,
value: int,
) -> None:
if not self.activate_counters:
msg = "The discipline counters are disabled."
raise RuntimeError(msg)
self._n_calls.value = value
@property
def exec_time(self) -> float | None:
"""The cumulated execution time of the discipline.
This property is multiprocessing safe.
Raises:
RuntimeError: When the discipline counters are disabled.
"""
if self.activate_counters:
return self._exec_time.value
return None
@exec_time.setter
def exec_time(
self,
value: float,
) -> None:
if not self.activate_counters:
msg = "The discipline counters are disabled."
raise RuntimeError(msg)
self._exec_time.value = value
@property
def n_calls_linearize(self) -> int | None:
"""The number of times the discipline was linearized.
This property is multiprocessing safe.
Raises:
RuntimeError: When the discipline counters are disabled.
"""
if self.activate_counters:
return self._n_calls_linearize.value
return None
@n_calls_linearize.setter
def n_calls_linearize(
self,
value: int,
) -> None:
if not self.activate_counters:
msg = "The discipline counters are disabled."
raise RuntimeError(msg)
self._n_calls_linearize.value = value
@property
def grammar_type(self) -> GrammarType:
"""The type of grammar to be used for inputs and outputs declaration."""
return self._grammar_type
[docs]
def auto_get_grammar_file(
self,
is_input: bool = True,
name: str | None = None,
comp_dir: str | Path | None = None,
) -> Path:
"""Use a naming convention to associate a grammar file to the discipline.
Search in the directory ``comp_dir`` for
either an input grammar file named ``name + "_input.json"``
or an output grammar file named ``name + "_output.json"``.
Args:
is_input: Whether to search for an input or output grammar file.
name: The name to be searched in the file names.
If ``None``,
use the name of the discipline class.
comp_dir: The directory in which to search the grammar file.
If ``None``,
use the :attr:`.GRAMMAR_DIRECTORY` if any,
or the directory of the discipline class module.
Returns:
The grammar file path.
"""
cls = self.__class__
initial_name = name or cls.__name__
classes = [cls] + [
base for base in cls.__bases__ if issubclass(base, MDODiscipline)
]
names = [initial_name] + [cls.__name__ for cls in classes[1:]]
in_or_out = "in" if is_input else "out"
for cls, name in zip(classes, names):
grammar_file_path = self.__get_grammar_file_path(
cls, comp_dir, in_or_out, name
)
if grammar_file_path.is_file():
return grammar_file_path
file_name = f"{initial_name}_{in_or_out}put.json"
msg = f"The grammar file {file_name} is missing."
raise FileNotFoundError(msg)
@staticmethod
def __get_grammar_file_path(
cls: type[MDODiscipline],
directory_path: str | Path | None,
in_or_out: str,
name: str,
) -> Path:
"""Return the grammar file path.
Args:
cls: The class for which the grammar file is searched.
directory_path: The initial directory path if any.
in_or_out: The suffix to look for in the file name, either "in" or "out".
name: The name to be searched in the file names.
Returns:
The grammar file path.
"""
if directory_path is None:
if cls.GRAMMAR_DIRECTORY is not None:
directory_path = Path(cls.GRAMMAR_DIRECTORY)
else:
class_module = sys.modules[cls.__module__]
directory_path = Path(class_module.__file__).parent.absolute()
else:
directory_path = Path(directory_path)
return directory_path / f"{name}_{in_or_out}put.json"
[docs]
def add_differentiated_outputs(
self,
outputs: Iterable[str] | None = None,
) -> None:
"""Add the outputs for differentiation.
The outputs that do not represent continuous numbers are filtered out.
Args:
outputs: The output variables to be differentiated.
If ``None``, all the outputs of the discipline are used.
Raises:
ValueError: When ``outputs `` are not in the output grammar.
"""
if not (outputs is None or self.is_all_outputs_existing(outputs)):
msg = (
f"Cannot differentiate the discipline {self.name} w.r.t. the outputs "
"that are not among the discipline outputs: "
f"{self.get_output_data_names()}."
)
raise ValueError(msg)
if outputs is None:
outputs = self.get_output_data_names()
self._differentiated_outputs = list(
set(self._differentiated_outputs).union(
filter(self.output_grammar.data_converter.is_continuous, outputs)
)
)
def __create_new_cache(
self,
class_name: str,
**kwargs: bool | float | str,
) -> AbstractCache:
"""Create a cache object.
Args:
class_name: The name of the cache class.
**kwargs: The arguments to instantiate a cache object.
Returns:
The cache object.
"""
if class_name != self.CacheType.HDF5:
for key in ("hdf_file_path", "hdf_node_path"):
if key in kwargs:
del kwargs[key]
if class_name != self.CacheType.MEMORY_FULL:
key = "is_memory_shared"
if key in kwargs:
del kwargs[key]
return CacheFactory().create(class_name, name=self.name, **kwargs)
[docs]
def set_cache_policy(
self,
cache_type: CacheType = CacheType.SIMPLE,
cache_tolerance: float = 0.0,
cache_hdf_file: str | Path | None = None,
cache_hdf_node_path: str | None = None,
is_memory_shared: bool = True,
) -> None:
"""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 :class:`.SimpleCache` recording the last execution
or a cache storing all executions,
e.g. :class:`.MemoryFullCache` and :class:`.HDF5Cache`.
Caching data can be either in-memory,
e.g. :class:`.SimpleCache` and :class:`.MemoryFullCache`,
or on the disk,
e.g. :class:`.HDF5Cache`.
The attribute :attr:`.CacheFactory.caches` provides the available caches types.
Args:
cache_type: The type of cache.
cache_tolerance: The maximum relative norm
of the difference between two input arrays
to consider that two input arrays are equal.
cache_hdf_file: The path to the HDF file to store the data;
this argument is mandatory when the
:attr:`.MDODiscipline.CacheType.HDF5` policy is used.
cache_hdf_node_path: The name of the HDF file node
to store the discipline data,
possibly passed as a path ``root_name/.../group_name/.../node_name``.
If ``None``, :attr:`.MDODiscipline.name` is used.
is_memory_shared: Whether to store the data with a shared memory dictionary,
which makes the cache compatible with multiprocessing.
.. warning:
If set to False,
and multiple disciplines point
to the same cache or the process is multiprocessed,
there may be duplicate computations
because the cache will not be shared among the processes.
"""
if cache_type == self.CacheType.NONE:
self.cache = None
return
if self.cache.__class__.__name__ != cache_type or not (
cache_type == self.CacheType.HDF5
and cache_hdf_file == self.cache.hdf_file.hdf_file_path
and cache_hdf_node_path == self.cache.hdf_node_path
):
self.cache = self.__create_new_cache(
cache_type,
tolerance=cache_tolerance,
hdf_file_path=cache_hdf_file,
hdf_node_path=cache_hdf_node_path or self.name,
is_memory_shared=is_memory_shared,
)
else:
LOGGER.warning(
(
"The cache policy is already set to %s "
"with the file path %r and node name %r; "
"call discipline.cache.clear() to clear the cache."
),
cache_type,
cache_hdf_file,
cache_hdf_node_path,
)
[docs]
def get_sub_disciplines(self, recursive: bool = False) -> list[MDODiscipline]:
"""Determine the sub-disciplines.
This method lists the sub-disciplines' disciplines. It will list up to one level
of disciplines contained inside another one unless the ``recursive`` argument is
set to ``True``.
Args:
recursive: If ``True``, the method will look inside any discipline that has
other disciplines inside until it reaches a discipline without
sub-disciplines, in this case the return value will not include any
discipline that has sub-disciplines. If ``False``, the method will list
up to one level of disciplines contained inside another one, in this
case the return value may include disciplines that contain
sub-disciplines.
Returns:
The sub-disciplines.
"""
return get_sub_disciplines(self._disciplines, recursive)
[docs]
def get_expected_workflow(self) -> AtomicExecSequence:
"""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.
.. seealso::
MDOFormulation.get_expected_workflow
Returns:
The expected execution sequence.
"""
# avoid circular dependency
from gemseo.core.execution_sequence import ExecutionSequenceFactory
return ExecutionSequenceFactory.atom(self)
[docs]
def get_expected_dataflow(
self,
) -> list[tuple[MDODiscipline, MDODiscipline, list[str]]]:
"""Return the expected data exchange sequence.
This method is used for the XDSM representation.
The default expected data exchange sequence is an empty list.
.. seealso::
MDOFormulation.get_expected_dataflow
Returns:
The data exchange arcs.
"""
return []
[docs]
def get_disciplines_in_dataflow_chain(self) -> list[MDODiscipline]:
"""Return the disciplines that must be shown as blocks in the XDSM.
By default, only the discipline itself is shown.
This function can be differently implemented for any type of inherited
discipline.
Returns:
The disciplines shown in the XDSM chain.
"""
return [self]
def _instantiate_grammars(
self,
input_grammar_file: str | Path | None,
output_grammar_file: str | Path | None,
grammar_type: GrammarType = GrammarType.JSON,
) -> None:
"""Create the input and output grammars.
Args:
input_grammar_file: The input grammar file path.
If ``None``, do not initialize the input grammar from a schema file.
output_grammar_file: The output grammar file path.
If ``None``, do not initialize the output grammar from a schema file.
grammar_type: The type of the input and output grammars.
"""
factory = GrammarFactory()
self.input_grammar = factory.create(
grammar_type,
name=f"{self.name}_discipline_input",
file_path=input_grammar_file,
)
self.output_grammar = factory.create(
grammar_type,
name=f"{self.name}_discipline_output",
file_path=output_grammar_file,
)
def _run(self) -> None:
"""Define the execution of the process, given that data has been checked.
To be overloaded by subclasses.
"""
raise NotImplementedError
def _filter_inputs(
self,
input_data: Mapping[str, Any] | None = None,
) -> DisciplineData:
"""Filter data with the discipline inputs and use the default values if missing.
Args:
input_data: The data to be filtered.
Returns:
The values of the input variables based on the provided data.
Raises:
TypeError: When the input data are not passed as a dictionary.
"""
if input_data is None:
return deepcopy(self.default_inputs)
if not isinstance(input_data, collections.abc.Mapping):
msg = f"Input data must be of dict type, got {type(input_data)} instead."
raise TypeError(msg)
full_input_data = DisciplineData({})
for input_name in self.input_grammar:
input_value = input_data.get(input_name)
if input_value is not None:
full_input_data[input_name] = input_value
else:
input_value = self.input_grammar.defaults.get(input_name)
if input_value is not None:
full_input_data[input_name] = input_value
return full_input_data
def _filter_local_data(self) -> None:
"""Filter the local data after execution.
This method removes data that are neither inputs nor outputs.
"""
for key in self._local_data.keys() - self.get_input_output_data_names():
del self._local_data[key]
def _check_status_before_run(self) -> None:
"""Check the status of the discipline.
Check the status of the discipline depending on
:attr:`.MDODiscipline.re_execute_policy`.
If ``re_exec_policy == ReExecutionPolicy.NEVER``,
the status shall be either :attr:`.MDODiscipline.ExecutionStatus.PENDING`
or :attr:`.MDODiscipline.VIRTUAL`.
If ``self.re_exec_policy == ReExecutionPolicy.NEVER``,
- if status is :attr:`.MDODiscipline.ExecutionStatus.DONE`,
:meth:`.MDODiscipline.reset_statuses_for_run`.
- otherwise status must be :attr:`.MDODiscipline.VIRTUAL`
or :attr:`.MDODiscipline.ExecutionStatus.PENDING`.
Raises:
ValueError:
If the discipline status and the re-execution policy are no consistent.
"""
if self.status not in {
self.ExecutionStatus.PENDING,
self.ExecutionStatus.VIRTUAL,
self.ExecutionStatus.DONE,
} or (
self.status == self.ExecutionStatus.DONE
and self.re_exec_policy == self.ReExecutionPolicy.NEVER
):
msg = (
f"Trying to run a discipline {type(self)} with status: {self.status} "
f"while re_exec_policy is {self.re_exec_policy}."
)
raise ValueError(msg)
def __create_input_data_for_cache(
self,
input_data: DisciplineData,
) -> DisciplineData:
"""Prepare the input data for caching.
Args:
input_data: The values of the inputs.
Returns:
The input data to be cached.
"""
in_and_out = set(self.get_input_data_names()).intersection(
self.get_output_data_names()
)
input_data_ = input_data.copy()
for key in in_and_out:
val = input_data.get(key)
if val is not None:
# If also an output, use a copy of the original input value
input_data_[key] = deepcopy(val)
# Non simple caches require NumPy arrays.
if not isinstance(self.cache, SimpleCache):
to_array = self.input_grammar.data_converter.convert_value_to_array
for name, value in input_data_.items():
input_data_[name] = to_array(name, value)
return input_data_
[docs]
def execute(
self,
input_data: Mapping[str, Any] | None = None,
) -> DisciplineData:
"""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 :attr:`.MDODiscipline.default_inputs`.
* Checks whether the last execution of the discipline was called
with identical inputs, i.e. cached in :attr:`.MDODiscipline.cache`;
if so, directly returns ``self.cache.get_output_cache(inputs)``.
* Caches the inputs.
* Checks the input data against :attr:`.MDODiscipline.input_grammar`.
* If :attr:`.MDODiscipline.data_processor` is not None, runs the preprocessor.
* Updates the status to :attr:`.MDODiscipline.ExecutionStatus.RUNNING`.
* Calls the :meth:`.MDODiscipline._run` method, that shall be defined.
* If :attr:`.MDODiscipline.data_processor` is not None, runs the postprocessor.
* Checks the output data.
* Caches the outputs.
* Updates the status to :attr:`.MDODiscipline.ExecutionStatus.DONE`
or :attr:`.MDODiscipline.ExecutionStatus.FAILED`.
* Updates summed execution time.
Args:
input_data: The input data needed to execute the discipline
according to the discipline input grammar.
If ``None``, use the :attr:`.MDODiscipline.default_inputs`.
Returns:
The discipline local data after execution.
"""
# Load the default_inputs if the user did not provide all required data
input_data = self._filter_inputs(input_data)
cached_local_data = self.__get_cache_data(input_data)
if cached_local_data is not None:
return cached_local_data
# Cache was not loaded, see self.linearize
self._cache_was_loaded = False
if self.activate_input_data_check:
self.check_input_data(input_data)
# Keep a pristine copy of the inputs before it is eventually changed.
inputs_for_cache = self.__create_input_data_for_cache(input_data)
processor = self.data_processor
if processor is not None:
self.__set_local_data(processor.pre_process_data(input_data))
else:
self.__set_local_data(input_data)
self._is_linearized = False
if self.activate_counters:
self.__increment_n_calls()
t_0 = timer()
self._check_status_before_run()
self.status = self.ExecutionStatus.RUNNING
if not self.virtual_execution:
try:
# Effectively run the discipline, the _run method has to be
# Defined by the subclasses
self._run()
except Exception:
self.status = self.ExecutionStatus.FAILED
# Update the status but
# raise the same exception
raise
else:
self.store_local_data(**self.default_outputs)
self.status = self.ExecutionStatus.DONE
if self.activate_counters:
self.__increment_exec_time(t_0)
# If the data processor is set, post process the data after _run
# See gemseo.core.data_processor module
if processor is not None:
self.__set_local_data(processor.post_process_data(self._local_data))
# Filter data that is neither outputs nor inputs
self._filter_local_data()
if self.activate_output_data_check:
self.check_output_data()
self.__cache_outputs(inputs_for_cache)
return self._local_data
def __get_cache_data(self, input_data: DisciplineData) -> DisciplineData | None:
"""Return the cached local data if cached.
Args:
input_data: The input data bound to the cached output data.
Returns:
The cached local data if cached, ``None`` otherwise.
"""
if self.cache is None:
return None
# Check if the cache already contains the outputs associated to these inputs.
_, out_cached, out_jac = self.cache[input_data]
if not out_cached:
return None
# Non simple caches require NumPy arrays.
if not isinstance(self.cache, SimpleCache):
# Do not modify the cache entry which is mutable.
out_cached = out_cached.copy()
to_value = self.output_grammar.data_converter.convert_array_to_value
for name, value in out_cached.items():
out_cached[name] = to_value(name, value)
self.__update_local_data_from_cache(input_data, out_cached, out_jac)
return self._local_data
def __cache_outputs(self, input_data: DisciplineData) -> None:
"""Store the output data in the cache.
Args:
input_data: The input data bound to the output data.
"""
if self.cache is None:
return
out_names = self.get_output_data_names()
if out_names:
outputs_for_cache = self._local_data.copy(keys=out_names)
# Non simple caches require NumPy arrays.
if not isinstance(self.cache, SimpleCache):
to_array = self.output_grammar.data_converter.convert_value_to_array
for name, value in outputs_for_cache.items():
outputs_for_cache[name] = to_array(name, value)
self.cache.cache_outputs(input_data, outputs_for_cache)
# Some disciplines are always linearized during execution, cache the
# jac in this case
if self._is_linearized:
self.cache.cache_jacobian(input_data, self.jac)
def __update_local_data_from_cache(
self,
input_data: DisciplineData,
out_cached: DisciplineData,
out_jac: dict[str, dict[str, ndarray]],
) -> None:
"""Update the local data from the cache.
Args:
input_data: The input data.
out_cached: The output data retrieved from the cache.
out_jac: The Jacobian data retrieved from the cache.
"""
self.__set_local_data(input_data)
self._local_data.update(out_cached)
if out_jac is not None:
self.jac = out_jac
self._is_linearized = True
else: # Erase jacobian which is unknown
self.jac.clear()
self._is_linearized = False
self.check_output_data()
self._cache_was_loaded = True
def __increment_n_calls(self) -> None:
"""Increment by 1 the number of executions.."""
with self._n_calls.get_lock():
self._n_calls.value += 1
def __increment_n_calls_lin(self) -> None:
"""Increment by 1 the number of linearizations."""
with self._n_calls_linearize.get_lock():
self._n_calls_linearize.value += 1
def __increment_exec_time(
self,
t_0: float,
linearize: bool = False,
) -> None:
"""Increment the execution time of the discipline.
The execution can be either an evaluation or a linearization.
Args:
t_0: The time of the execution start.
linearize: Whether it is a linearization.
"""
curr_t = timer()
with self._exec_time.get_lock():
self._exec_time.value += curr_t - t_0
time_stamps = MDODiscipline.time_stamps
if time_stamps is not None:
disc_stamps = time_stamps.get(self.name, self.__mp_manager.list())
stamp = (t_0, curr_t, linearize)
disc_stamps.append(stamp)
time_stamps[self.name] = disc_stamps
def _retrieve_diff_inouts(
self,
compute_all_jacobians: bool = False,
) -> tuple[list[str], list[str]]:
"""Get the inputs and outputs used in the differentiation of the discipline.
Args:
compute_all_jacobians: Whether to compute the Jacobians of all the output
with respect to all the inputs.
Otherwise,
set the input variables against which to differentiate the output ones
with :meth:`.add_differentiated_inputs`
and set these output variables to differentiate
with :meth:`.add_differentiated_outputs`.
"""
if compute_all_jacobians:
return self.get_input_data_names(), self.get_output_data_names()
return self._differentiated_inputs, self._differentiated_outputs
[docs]
@classmethod
def activate_time_stamps(cls) -> None:
"""Activate the time stamps.
For storing start and end times of execution and linearizations.
"""
MDODiscipline.__mp_manager = manager = get_multi_processing_manager()
MDODiscipline.time_stamps = manager.dict()
[docs]
@classmethod
def deactivate_time_stamps(cls) -> None:
"""Deactivate the time stamps.
For storing start and end times of execution and linearizations.
"""
MDODiscipline.time_stamps = None
[docs]
def linearize(
self,
input_data: Mapping[str, Any] | None = None,
compute_all_jacobians: bool = False,
execute: bool = True,
) -> Mapping[str, Mapping[str, NDArray[float]]]:
"""Compute the Jacobians of some outputs with respect to some inputs.
Args:
input_data: The input data for which to compute the Jacobian.
If ``None``, use the :attr:`.MDODiscipline.default_inputs`.
compute_all_jacobians: Whether to compute the Jacobians of all the output
with respect to all the inputs.
Otherwise,
set the input variables against which to differentiate the output ones
with :meth:`.add_differentiated_inputs`
and set these output variables to differentiate
with :meth:`.add_differentiated_outputs`.
execute: Whether to start by executing the discipline
with the input data for which to compute the Jacobian;
this allows to ensure that the discipline was executed
with the right input data;
it can be almost free if the corresponding output data
have been stored in the :attr:`.cache`.
Returns:
The Jacobian of the discipline
shaped as ``{output_name: {input_name: jacobian_array}}`` where
``jacobian_array[i, j]`` is the partial derivative of ``output_name[i]``
with respect to ``input_name[j]``.
Raises:
ValueError: When either the inputs
for which to differentiate the outputs
or the outputs to differentiate are missing.
"""
# TODO: remove the execution when no option exec_before_lin is set to True.
input_data = self._filter_inputs(input_data)
inputs, outputs = self._retrieve_diff_inouts(compute_all_jacobians)
if self.cache is not None and (not outputs or not inputs):
self.jac = self.cache[input_data].jacobian
return self.jac
if execute:
self.reset_statuses_for_run()
self.exec_for_lin = True
self.execute(input_data)
self.exec_for_lin = False
# The local_data shall be reset to their original values
# in case an input is also an output,
# if we don't want to keep the computed state (as in MDAs).
if not self._linearize_on_last_state:
self._local_data.update(input_data)
# If the caching was triggered,
# check if the jacobian was loaded,
# or the discipline._run method also linearizes the discipline.
if (self._cache_was_loaded or self._is_linearized) and self.jac:
# For cases when linearization is called twice with different I/O
# while cache_was_loaded=True,
# the check_jacobian_shape raises a KeyError.
try:
self._check_jacobian_shape(inputs, outputs)
except KeyError:
# In this case, another computation of Jacobian is triggered.
pass
else:
return self.jac
self.status = self.ExecutionStatus.LINEARIZE
t_0 = timer()
if self._linearization_mode in set(self.ApproximationMode):
# Time already counted in execute()
self.jac = self._jac_approx.compute_approx_jac(outputs, inputs)
else:
self._compute_jacobian(inputs, outputs)
if self.activate_counters:
self.__increment_exec_time(t_0, linearize=True)
if self.activate_counters:
self.__increment_n_calls_lin()
if not compute_all_jacobians:
for output_name in list(self.jac.keys()):
if output_name not in outputs:
del self.jac[output_name]
else:
jac = self.jac[output_name]
for input_name in list(jac.keys()):
if input_name not in inputs:
del jac[input_name]
# The check of the jacobian shape is required only when some of its
# components are requested.
if inputs and outputs:
self._check_jacobian_shape(inputs, outputs)
if self.cache is not None:
self.cache.cache_jacobian(input_data, self.jac)
self.status = self.ExecutionStatus.DONE
return self.jac
[docs]
def set_jacobian_approximation(
self,
jac_approx_type: ApproximationMode = ApproximationMode.FINITE_DIFFERENCES,
jax_approx_step: float = 1e-7,
jac_approx_n_processes: int = 1,
jac_approx_use_threading: bool = False,
jac_approx_wait_time: float = 0,
) -> None:
"""Set the Jacobian approximation method.
Sets the linearization mode to approx_method,
sets the parameters of the approximation for further use
when calling :meth:`.MDODiscipline.linearize`.
Args:
jac_approx_type: The approximation method,
either "complex_step" or "finite_differences".
jax_approx_step: The differentiation step.
jac_approx_n_processes: The maximum simultaneous number of threads,
if ``jac_approx_use_threading`` is True, or processes otherwise,
used to parallelize the execution.
jac_approx_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.
jac_approx_wait_time: The time waited between two forks
of the process / thread.
"""
self._jac_approx = DisciplineJacApprox(
self,
approx_method=jac_approx_type,
step=jax_approx_step,
parallel=jac_approx_n_processes > 1,
n_processes=jac_approx_n_processes,
use_threading=jac_approx_use_threading,
wait_time_between_fork=jac_approx_wait_time,
)
self.linearization_mode = jac_approx_type
[docs]
def set_optimal_fd_step(
self,
outputs: Iterable[str] | None = None,
inputs: Iterable[str] | None = None,
compute_all_jacobians: bool = False,
print_errors: bool = False,
numerical_error: float = EPSILON,
) -> ndarray:
"""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
(round-off when doing f(x+step)-f(x))
are approximately equal.
.. warning::
This calls the discipline execution twice per input variables.
.. seealso::
https://en.wikipedia.org/wiki/Numerical_differentiation
and
"Numerical Algorithms and Digital Representation", Knut Morken ,
Chapter 11, "Numerical Differentiation"
Args:
inputs: The inputs wrt which the outputs are linearized.
If ``None``, use the :attr:`.MDODiscipline._differentiated_inputs`.
outputs: The outputs to be linearized.
If ``None``, use the :attr:`.MDODiscipline._differentiated_outputs`.
compute_all_jacobians: Whether to compute the Jacobians of all the output
with respect to all the inputs.
Otherwise,
set the input variables against which to differentiate the output ones
with :meth:`.add_differentiated_inputs`
and set these output variables to differentiate
with :meth:`.add_differentiated_outputs`.
print_errors: Whether to display the estimated errors.
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.
Returns:
The estimated errors of truncation and cancellation error.
Raises:
ValueError: When the Jacobian approximation method has not been set.
"""
if self._jac_approx is None:
msg = (
"set_jacobian_approximation must be called "
"before setting an optimal step."
)
raise ValueError(msg)
diff_inputs, diff_outputs = self._retrieve_diff_inouts(
compute_all_jacobians=compute_all_jacobians
)
if outputs is None or compute_all_jacobians:
outputs = diff_outputs
if inputs is None or compute_all_jacobians:
inputs = diff_inputs
return self._jac_approx.auto_set_step(
outputs, inputs, print_errors, numerical_error=numerical_error
)
def _check_jacobian_shape(
self,
input_names: Iterable[str],
output_names: Iterable[str],
) -> None:
"""Check that the Jacobian is a dictionary of dictionaries of 2D NumPy arrays.
Args:
input_names: The input names wrt the output names are linearized.
output_names: The output names to be linearized.
Raises:
ValueError:
When the discipline was not linearized.
When the Jacobian is not of the right shape.
KeyError:
When output names are missing in the Jacobian of the discipline.
When input names are missing for an output in the Jacobian of the
discipline.
"""
if not self.jac:
msg = f"The discipline {self.name} was not linearized."
raise ValueError(msg)
unique_output_names = set(output_names)
unique_input_names = set(input_names)
jac_output_names = self.jac.keys()
if not unique_output_names.issubset(jac_output_names):
msg = (
f"Missing output names in Jacobian of discipline {self.name}: "
f"{unique_output_names.difference(jac_output_names)}."
)
raise KeyError(msg)
get_input_size = self.input_grammar.data_converter.get_value_size
get_output_size = self.output_grammar.data_converter.get_value_size
for output_name in output_names:
jac_output = self.jac[output_name]
if not unique_input_names.issubset(jac_output.keys()):
msg = (
f"Missing input names {unique_input_names.difference(jac_output)} "
f"in Jacobian of discipline {self.name}, "
f"for output: {output_name}."
)
raise KeyError(msg)
output_value = self._local_data.get(output_name)
if output_value is None:
# Unknown dimension, don't check the shape.
continue
output_size = get_output_size(output_name, output_value)
for input_name in input_names:
input_value = self._local_data.get(input_name)
if input_value is None:
# Unknown dimension, don't check the shape.
continue
input_size = get_input_size(input_name, input_value)
if jac_output[input_name].shape != (output_size, input_size):
msg = (
f"The shape {jac_output[input_name].shape} "
f"of the Jacobian matrix d{output_name}/d{input_name} "
f"of the discipline {self.name} "
"does not match "
f"(output_size, input_size)=({output_size}, {input_size})."
)
raise ValueError(msg)
# Discard imaginary part of Jacobian
for output_jacobian in self.jac.values():
for input_name, input_output_jacobian in output_jacobian.items():
output_jacobian[input_name] = input_output_jacobian.real
@property
def cache_tol(self) -> float:
"""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)``.
Raises:
ValueError: When the discipline does not have a cache.
"""
if self.cache is None:
msg = f"The discipline {self.name} does not have a cache."
raise ValueError(msg)
return self.cache.tolerance
@cache_tol.setter
def cache_tol(
self,
cache_tol: float,
) -> None:
if self.cache is None:
msg = f"The discipline {self.name} does not have a cache."
raise ValueError(msg)
self._set_cache_tol(cache_tol)
def _set_cache_tol(
self,
cache_tol: float,
) -> None:
"""Set to the cache input tolerance.
To be overloaded by subclasses.
Args:
cache_tol: The cache tolerance.
"""
self.cache.tolerance = cache_tol or 0.0
@property
def default_inputs(self) -> Defaults:
"""The default inputs."""
return self.input_grammar.defaults
@default_inputs.setter
def default_inputs(self, data: Mapping[str, Any]) -> None:
self.input_grammar.defaults = data
@property
def default_outputs(self) -> Defaults:
"""The default outputs used when :attr:`.virtual_execution` is ``True``."""
return self.output_grammar.defaults
@default_outputs.setter
def default_outputs(self, data: Mapping[str, Any]) -> None:
self.output_grammar.defaults = data
[docs]
def add_namespace_to_output(self, name: str, namespace: str) -> None:
"""Add a namespace prefix to an existing output grammar element.
The updated output grammar element name will be
``namespace`` + :data:`~gemseo.core.namespaces.namespaces_separator` + ``name``.
Args:
name: The element name to rename.
namespace: The name of the namespace.
"""
self.output_grammar.add_namespace(name, namespace)
def _compute_jacobian(
self,
inputs: Iterable[str] | None = None,
outputs: Iterable[str] | None = None,
) -> None:
"""Compute the Jacobian matrix.
To be overloaded by subclasses, actual computation of the Jacobian matrix.
Args:
inputs: The inputs wrt the outputs are linearized.
If ``None``,
the linearization should be performed wrt all inputs.
outputs: The outputs to be linearized.
If ``None``,
the linearization should be performed on all outputs.
"""
def _init_jacobian(
self,
inputs: Iterable[str] | None = None,
outputs: Iterable[str] | None = None,
init_type: InitJacobianType = InitJacobianType.DENSE,
fill_missing_keys: bool = False,
) -> tuple[list[str], list[str]]:
"""Initialize the Jacobian dictionary of the form ``{input: {output: matrix}}``.
Args:
inputs: The inputs wrt to which the outputs are linearized.
If ``None``,
the linearization should be performed wrt all inputs.
outputs: The outputs to be linearized.
If ``None``,
the linearization should be performed on all outputs declared
differentiable.
init_type: The type used to initialize the Jacobian.
fill_missing_keys: If ``True``,
just fill the missing items with zeros/empty
but do not override the existing data.
Returns:
The names of the input variables
against which to differentiate the output ones,
and these output variables.
"""
if init_type == self.InitJacobianType.EMPTY:
default_matrix = empty
elif init_type == self.InitJacobianType.DENSE:
default_matrix = zeros
elif init_type == self.InitJacobianType.SPARSE:
default_matrix = csr_array
else:
# Cast the argument so that the enum class raise an explicit error.
self.InitJacobianType(init_type)
input_names = self._differentiated_inputs if inputs is None else inputs
input_names_to_sizes = self.input_grammar.data_converter.compute_names_to_sizes(
input_names,
self._local_data,
)
output_names = self._differentiated_outputs if outputs is None else outputs
output_names_to_sizes = (
self.output_grammar.data_converter.compute_names_to_sizes(
output_names,
self._local_data,
)
)
if fill_missing_keys:
jac = self.jac
# Only fill the missing sub jacobians
for output_name, output_size in output_names_to_sizes.items():
jac_loc = jac.setdefault(output_name, defaultdict(None))
for input_name, input_size in input_names_to_sizes.items():
sub_jac = jac_loc.get(input_name)
if sub_jac is None:
jac_loc[input_name] = default_matrix((output_size, input_size))
else:
# When a key is not in the default dict, ie a function is not in
# the Jacobian; return an empty defaultdict(None)
jac = defaultdict(default_dict_factory)
if input_names:
for output_name, output_size in output_names_to_sizes.items():
jac_loc = jac[output_name]
for input_name, input_size in input_names_to_sizes.items():
jac_loc[input_name] = default_matrix((output_size, input_size))
self.jac = jac
return input_names, output_names
@property
def linearization_mode(self) -> LinearizationMode:
"""The linearization mode among :attr:`.MDODiscipline.LinearizationMode`.
Raises:
ValueError: When the linearization mode is unknown.
"""
return self._linearization_mode
@linearization_mode.setter
def linearization_mode(
self,
linearization_mode: LinearizationMode,
) -> None:
self._linearization_mode = linearization_mode
if (
linearization_mode in set(self.ApproximationMode)
and self._jac_approx is None
):
self.set_jacobian_approximation(linearization_mode)
[docs]
def check_jacobian(
self,
input_data: Mapping[str, ndarray] | None = None,
derr_approx: ApproximationMode = ApproximationMode.FINITE_DIFFERENCES,
step: float = 1e-7,
threshold: float = 1e-8,
linearization_mode: str = "auto",
inputs: Iterable[str] | None = None,
outputs: Iterable[str] | None = None,
parallel: bool = False,
n_processes: int = N_CPUS,
use_threading: bool = False,
wait_time_between_fork: float = 0,
auto_set_step: bool = False,
plot_result: bool = False,
file_path: str | Path = "jacobian_errors.pdf",
show: bool = False,
fig_size_x: float = 10,
fig_size_y: float = 10,
reference_jacobian_path: str | Path | None = None,
save_reference_jacobian: bool = False,
indices: Iterable[int] | None = None,
) -> bool:
"""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.
Args:
input_data: The input data needed to execute the discipline
according to the discipline input grammar.
If ``None``, use the :attr:`.MDODiscipline.default_inputs`.
derr_approx: The approximation method,
either "complex_step" or "finite_differences".
threshold: The acceptance threshold for the Jacobian error.
linearization_mode: the mode of linearization: direct, adjoint
or automated switch depending on dimensions
of inputs and outputs (Default value = 'auto')
inputs: The names of the inputs wrt which to differentiate the outputs.
outputs: The names of the outputs to be differentiated.
step: The differentiation step.
parallel: Whether to differentiate the discipline in parallel.
n_processes: The maximum simultaneous number of threads,
if ``use_threading`` is True, or processes otherwise,
used to parallelize the execution.
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.
wait_time_between_fork: The time waited between two forks
of the process / thread.
auto_set_step: Whether to compute the optimal step
for a forward first order finite differences gradient approximation.
plot_result: Whether to plot the result of the validation
(computed vs approximated Jacobians).
file_path: The path to the output file if ``plot_result`` is ``True``.
show: Whether to open the figure.
fig_size_x: The x-size of the figure in inches.
fig_size_y: The y-size of the figure in inches.
reference_jacobian_path: The path of the reference Jacobian file.
save_reference_jacobian: Whether to save the reference Jacobian.
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``.
Returns:
Whether the analytical Jacobian is correct
with respect to the reference one.
"""
# Do not use self._jac_approx because we may want to check complex
# step approximation with the finite differences for instance
approx = DisciplineJacApprox(
self,
derr_approx,
step,
parallel,
n_processes,
use_threading,
wait_time_between_fork,
)
if inputs is None:
inputs = self.get_input_data_names()
if outputs is None:
outputs = self.get_output_data_names()
if auto_set_step:
approx.auto_set_step(outputs, inputs)
# Differentiate analytically
self.add_differentiated_inputs(inputs)
self.add_differentiated_outputs(outputs)
self.linearization_mode = linearization_mode
self.reset_statuses_for_run()
# Linearize performs execute() if needed
self.linearize(input_data)
return approx.check_jacobian(
self.jac,
outputs,
inputs,
self,
threshold,
plot_result=plot_result,
file_path=file_path,
show=show,
fig_size_x=fig_size_x,
fig_size_y=fig_size_y,
reference_jacobian_path=reference_jacobian_path,
save_reference_jacobian=save_reference_jacobian,
indices=indices,
)
@property
def status(self) -> ExecutionStatus:
"""The status of the discipline.
The status aims at monitoring the process and give the user a simplified view on
the state (the process state = execution or linearize or done) of the
disciplines. The core part of the execution is _run, the core part of linearize
is _compute_jacobian or approximate jacobian computation.
"""
return self._status
[docs]
def set_disciplines_statuses(
self,
status: str,
) -> None:
"""Set the sub-disciplines statuses.
To be implemented in subclasses.
Args:
status: The status.
"""
[docs]
def is_output_existing(
self,
data_name: str,
) -> bool:
"""Test if a variable is a discipline output.
Args:
data_name: The name of the variable.
Returns:
Whether the variable is a discipline output.
"""
return data_name in self.output_grammar
[docs]
def is_all_outputs_existing(
self,
data_names: Iterable[str],
) -> bool:
"""Test if several variables are discipline outputs.
Args:
data_names: The names of the variables.
Returns:
Whether all the variables are discipline outputs.
"""
output_names = self.output_grammar.keys()
return all(data_name in output_names for data_name in data_names)
def _is_status_ok_for_run_again(
self,
status: str,
) -> bool:
"""Check if the discipline can be run again.
Args:
status: The status of the discipline.
Returns:
Whether the discipline can be run again.
"""
return status != self.ExecutionStatus.RUNNING
[docs]
def reset_statuses_for_run(self) -> None:
"""Set all the statuses to :attr:`.MDODiscipline.ExecutionStatus.PENDING`.
Raises:
ValueError: When the discipline cannot be run because of its status.
"""
if not self._is_status_ok_for_run_again(self.status):
msg = f"Cannot run discipline {self.name} with status {self.status}."
raise ValueError(msg)
self.status = self.ExecutionStatus.PENDING
@status.setter
def status(
self,
status: ExecutionStatus,
) -> None:
self._status = status
self.notify_status_observers()
[docs]
def add_status_observer(
self,
obs: Any,
) -> None:
"""Add an observer for the status.
Add an observer for the status
to be notified when self changes of status.
Args:
obs: The observer to add.
"""
if obs not in self._status_observers:
self._status_observers.append(obs)
[docs]
def remove_status_observer(
self,
obs: Any,
) -> None:
"""Remove an observer for the status.
Args:
obs: The observer to remove.
"""
if obs in self._status_observers:
self._status_observers.remove(obs)
[docs]
def notify_status_observers(self) -> None:
"""Notify all status observers that the status has changed."""
for obs in self._status_observers[:]:
obs.update_status(self)
[docs]
def store_local_data(self, **kwargs: Any) -> None:
"""Store discipline data in local data.
Args:
**kwargs: The data to be stored in :attr:`.MDODiscipline.local_data`.
"""
out_ns = self.output_grammar.to_namespaced
if not out_ns:
self._local_data.update(kwargs)
else:
out_names = self.output_grammar.keys()
for key, value in kwargs.items():
if key in out_names:
self._local_data[key] = value
else:
key_with_ns = out_ns.get(key)
if key_with_ns is not None:
self._local_data[key_with_ns] = value
# else out data will be cleared
[docs]
def check_output_data(
self,
raise_exception: bool = True,
) -> None:
"""Check the output data validity.
Args:
raise_exception: Whether to raise an exception when the data is invalid.
"""
self.output_grammar.validate(self._local_data, raise_exception)
[docs]
def get_outputs_asarray(self) -> ndarray:
"""Return the local input data as a large NumPy array.
The order is the one of :meth:`.MDODiscipline.get_all_inputs`.
Returns:
The local input data.
"""
return concatenate(list(self.get_all_outputs()))
# TODO: API: remove and replace with get_data_list_from_dict with a default for the
# the second argument such that it uses self._local_data
# TODO: API: remove and replace with get_data_list_from_dict with a default for the
# the second argument such that it uses self._local_data
[docs]
def get_outputs_by_name(
self,
data_names: Iterable[str],
) -> Iterator[Any]:
"""Return the local data associated with output variables.
Args:
data_names: 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.
"""
try:
return self.get_data_list_from_dict(data_names, self._local_data)
except KeyError as err:
msg = f"Discipline {self.name} has no output named {err}."
raise ValueError(msg) from None
[docs]
def get_output_data_names(self, with_namespaces: bool = True) -> list[str]:
"""Return the names of the output variables.
Args:
with_namespaces: Whether to keep the namespace prefix of the
output names, if any.
Returns:
The names of the output variables.
"""
if with_namespaces:
return list(self.output_grammar.keys())
return remove_prefix_from_list(self.output_grammar.keys())
[docs]
def get_all_outputs(self) -> Iterator[Any]:
"""Return the local output data.
The order is given by :meth:`.MDODiscipline.get_output_data_names`.
Returns:
The local output data.
"""
return self.get_outputs_by_name(self.get_output_data_names())
[docs]
def get_output_data(self, with_namespaces: bool = True) -> dict[str, Any]:
"""Return the local output data as a dictionary.
Args:
with_namespaces: Whether to keep the namespace prefix of the
output names, if any.
Returns:
The local output data.
"""
return self._local_data.copy(
keys=self.output_grammar.keys(),
with_namespace=with_namespaces or not self.output_grammar.to_namespaced,
)
[docs]
def to_pickle(self, file_path: str | Path) -> None:
"""Serialize the discipline and store it in a file.
Args:
file_path: The path to the file to store the discipline.
"""
with Path(file_path).open("wb") as outfobj:
pickler = pickle.Pickler(outfobj, protocol=2)
pickler.dump(self)
[docs]
@staticmethod
def from_pickle(file_path: str | Path) -> MDODiscipline:
"""Deserialize a discipline from a file.
Args:
file_path: The path to the file containing the discipline.
Returns:
The discipline instance.
"""
with Path(file_path).open("rb") as file_:
pickler = pickle.Unpickler(file_)
return pickler.load()
def __setstate__(
self,
state: Mapping[str, Any],
) -> None:
super().__setstate__(state)
# Initialize the attributes that are not serializable nor Synchronized last.
self._status_observers = []
[docs]
def get_local_data_by_name(
self,
data_names: Iterable[str],
) -> Iterator[Any]:
"""Return the local data of the discipline associated with variables names.
Args:
data_names: The names of the variables.
Returns:
The local data associated with the variables names.
Raises:
ValueError: When a name is not a discipline input name.
"""
try:
return self.get_data_list_from_dict(data_names, self._local_data)
except KeyError as err:
msg = f"Discipline {self.name} has no local_data named {err}."
raise ValueError(msg) from None
[docs]
@staticmethod
def is_scenario() -> bool:
"""Whether the discipline is a scenario."""
return False
# TODO: API: let the second argument have None as default value such that
# self._local_data is used as data_dict.
[docs]
@staticmethod
def get_data_list_from_dict(
keys: str | Iterable[str],
data_dict: dict[str, Any],
) -> Any | Iterator[Any]:
"""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.
Args:
keys: One or several names.
data_dict: The mapping from which to get the data.
Returns:
Either a data or a generator of data.
"""
if isinstance(keys, str):
return data_dict[keys]
return (data_dict[name] for name in keys)
