ModelicaSystem.py

# -*- coding: utf-8 -*-
"""
Definition of main class to run Modelica simulations - ModelicaSystem.
"""

import abc
import ast
from dataclasses import dataclass
import itertools
import logging
import numbers
import os
import pathlib
import queue
import re
import textwrap
import threading
from typing import Any, cast, Optional, Tuple
import warnings
import xml.etree.ElementTree as ET

import numpy as np

from OMPython.OMCSession import (
    ModelExecutionData,
    ModelExecutionException,

    OMCSessionException,
    OMCSessionLocal,

    OMPathABC,

    OMSessionABC,
    OMSessionRunner,
)

# define logger using the current module name as ID
logger = logging.getLogger(__name__)


class ModelicaSystemError(Exception):
    """
    Exception used in ModelicaSystem classes.
    """


@dataclass
class LinearizationResult:
    """Modelica model linearization results.

    Attributes:
        n: number of states
        m: number of inputs
        p: number of outputs
        A: state matrix (n x n)
        B: input matrix (n x m)
        C: output matrix (p x n)
        D: feedthrough matrix (p x m)
        x0: fixed point
        u0: input corresponding to the fixed point
        stateVars: names of state variables
        inputVars: names of inputs
        outputVars: names of outputs
    """

    n: int
    m: int
    p: int

    A: list
    B: list
    C: list
    D: list

    x0: list[float]
    u0: list[float]

    stateVars: list[str]
    inputVars: list[str]
    outputVars: list[str]

    def __iter__(self):
        """Allow unpacking A, B, C, D = result."""
        yield self.A
        yield self.B
        yield self.C
        yield self.D

    def __getitem__(self, index: int):
        """Allow accessing A, B, C, D via result[0] through result[3].

        This is needed for backwards compatibility, because
        ModelicaSystem.linearize() used to return [A, B, C, D].
        """
        return {0: self.A, 1: self.B, 2: self.C, 3: self.D}[index]


class ModelExecutionCmd:
    """
    All information about a compiled model executable. This should include data about all structured parameters, i.e.
    parameters which need a recompilation of the model. All non-structured parameters can be easily changed without
    the need for recompilation.
    """

    def __init__(
            self,
            runpath: os.PathLike,
            cmd_prefix: list[str],
            cmd_local: bool = False,
            cmd_windows: bool = False,
            timeout: float = 10.0,
            model_name: Optional[str] = None,
    ) -> None:
        if model_name is None:
            raise ModelExecutionException("Missing model name!")

        self._cmd_local = cmd_local
        self._cmd_windows = cmd_windows
        self._cmd_prefix = cmd_prefix
        self._runpath = pathlib.PurePosixPath(runpath)
        self._model_name = model_name
        self._timeout = timeout

        # dictionaries of command line arguments for the model executable
        self._args: dict[str, str | None] = {}
        # 'override' argument needs special handling, as it is a dict on its own saved as dict elements following the
        # structure: 'key' => 'key=value'
        self._arg_override: dict[str, str] = {}

    def arg_set(
            self,
            key: str,
            val: Optional[str | dict[str, Any] | numbers.Number] = None,
    ) -> None:
        """
        Set one argument for the executable model.

        Args:
            key: identifier / argument name to be used for the call of the model executable.
            val: value for the given key; None for no value and for key == 'override' a dictionary can be used which
              indicates variables to override
        """

        def override2str(
                orkey: str,
                orval: str | bool | numbers.Number,
        ) -> str:
            """
            Convert a value for 'override' to a string taking into account differences between Modelica and Python.
            """
            # check oval for any string representations of numbers (or bool) and convert these to Python representations
            if isinstance(orval, str):
                try:
                    val_evaluated = ast.literal_eval(orval)
                    if isinstance(val_evaluated, (numbers.Number, bool)):
                        orval = val_evaluated
                except (ValueError, SyntaxError):
                    pass

            if isinstance(orval, str):
                val_str = orval.strip()
            elif isinstance(orval, bool):
                val_str = 'true' if orval else 'false'
            elif isinstance(orval, numbers.Number):
                val_str = str(orval)
            else:
                raise ModelExecutionException(f"Invalid value for override key {orkey}: {type(orval)}")

            return f"{orkey}={val_str}"

        if not isinstance(key, str):
            raise ModelExecutionException(f"Invalid argument key: {repr(key)} (type: {type(key)})")
        key = key.strip()

        if isinstance(val, dict):
            if key != 'override':
                raise ModelExecutionException("Dictionary input only possible for key 'override'!")

            for okey, oval in val.items():
                if not isinstance(okey, str):
                    raise ModelExecutionException("Invalid key for argument 'override': "
                                                  f"{repr(okey)} (type: {type(okey)})")

                if not isinstance(oval, (str, bool, numbers.Number, type(None))):
                    raise ModelExecutionException(f"Invalid input for 'override'.{repr(okey)}: "
                                                  f"{repr(oval)} (type: {type(oval)})")

                if okey in self._arg_override:
                    if oval is None:
                        logger.info(f"Remove model executable override argument: {repr(self._arg_override[okey])}")
                        del self._arg_override[okey]
                        continue

                    logger.info(f"Update model executable override argument: {repr(okey)} = {repr(oval)} "
                                f"(was: {repr(self._arg_override[okey])})")

                if oval is not None:
                    self._arg_override[okey] = override2str(orkey=okey, orval=oval)

            argval = ','.join(sorted(self._arg_override.values()))
        elif val is None:
            argval = None
        elif isinstance(val, str):
            argval = val.strip()
        elif isinstance(val, numbers.Number):
            argval = str(val)
        else:
            raise ModelExecutionException(f"Invalid argument value for {repr(key)}: {repr(val)} (type: {type(val)})")

        if key in self._args:
            logger.warning(f"Override model executable argument: {repr(key)} = {repr(argval)} "
                           f"(was: {repr(self._args[key])})")
        self._args[key] = argval

    def arg_get(self, key: str) -> Optional[str | dict[str, str | bool | numbers.Number]]:
        """
        Return the value for the given key
        """
        if key in self._args:
            return self._args[key]

        return None

    def args_set(
            self,
            args: dict[str, Optional[str | dict[str, Any] | numbers.Number]],
    ) -> None:
        """
        Define arguments for the model executable.
        """
        for arg in args:
            self.arg_set(key=arg, val=args[arg])

    def get_cmd_args(self) -> list[str]:
        """
        Get a list with the command arguments for the model executable.
        """

        cmdl = []
        for key in sorted(self._args):
            if self._args[key] is None:
                cmdl.append(f"-{key}")
            else:
                cmdl.append(f"-{key}={self._args[key]}")

        return cmdl

    def definition(self) -> ModelExecutionData:
        """
        Define all needed data to run the model executable. The data is stored in an OMCSessionRunData object.
        """
        # ensure that a result filename is provided
        result_file = self.arg_get('r')
        if not isinstance(result_file, str):
            result_file = (self._runpath / f"{self._model_name}.mat").as_posix()

        # as this is the local implementation, pathlib.Path can be used
        cmd_path = self._runpath

        cmd_library_path = None
        if self._cmd_local and self._cmd_windows:
            cmd_library_path = ""

            # set the process environment from the generated .bat file in windows which should have all the dependencies
            # for this pathlib.PurePosixPath() must be converted to a pathlib.Path() object, i.e. WindowsPath
            path_bat = pathlib.Path(cmd_path) / f"{self._model_name}.bat"
            if not path_bat.is_file():
                raise ModelExecutionException("Batch file (*.bat) does not exist " + str(path_bat))

            content = path_bat.read_text(encoding='utf-8')
            for line in content.splitlines():
                match = re.match(pattern=r"^SET PATH=([^%]*)", string=line, flags=re.IGNORECASE)
                if match:
                    cmd_library_path = match.group(1).strip(';')  # Remove any trailing semicolons
            my_env = os.environ.copy()
            my_env["PATH"] = cmd_library_path + os.pathsep + my_env["PATH"]

            cmd_model_executable = cmd_path / f"{self._model_name}.exe"
        else:
            # for Linux the paths to the needed libraries should be included in the executable (using rpath)
            cmd_model_executable = cmd_path / self._model_name

        # define local(!) working directory
        cmd_cwd_local = None
        if self._cmd_local:
            cmd_cwd_local = cmd_path.as_posix()

        omc_run_data = ModelExecutionData(
            cmd_path=cmd_path.as_posix(),
            cmd_model_name=self._model_name,
            cmd_args=self.get_cmd_args(),
            cmd_result_file=result_file,
            cmd_prefix=self._cmd_prefix,
            cmd_library_path=cmd_library_path,
            cmd_model_executable=cmd_model_executable.as_posix(),
            cmd_cwd_local=cmd_cwd_local,
            cmd_timeout=self._timeout,
        )

        return omc_run_data

    @staticmethod
    def parse_simflags(simflags: str) -> dict[str, Optional[str | dict[str, Any] | numbers.Number]]:
        """
        Parse a simflag definition; this is deprecated!

        The return data can be used as input for self.args_set().
        """
        warnings.warn(
            message="The argument 'simflags' is depreciated and will be removed in future versions; "
                    "please use 'simargs' instead",
            category=DeprecationWarning,
            stacklevel=2,
        )

        simargs: dict[str, Optional[str | dict[str, Any] | numbers.Number]] = {}

        args = [s for s in simflags.split(' ') if s]
        for arg in args:
            if arg[0] != '-':
                raise ModelExecutionException(f"Invalid simulation flag: {arg}")
            arg = arg[1:]
            parts = arg.split('=')
            if len(parts) == 1:
                simargs[parts[0]] = None
            elif parts[0] == 'override':
                override = '='.join(parts[1:])

                override_dict = {}
                for item in override.split(','):
                    kv = item.split('=')
                    if not 0 < len(kv) < 3:
                        raise ModelExecutionException(f"Invalid value for '-override': {override}")
                    if kv[0]:
                        try:
                            override_dict[kv[0]] = kv[1]
                        except (KeyError, IndexError) as ex:
                            raise ModelExecutionException(f"Invalid value for '-override': {override}") from ex

                simargs[parts[0]] = override_dict

        return simargs


class ModelicaSystemABC(metaclass=abc.ABCMeta):
    """
    Base class to simulate a Modelica models.
    """

    def __init__(
            self,
            session: OMSessionABC,
            work_directory: Optional[str | os.PathLike] = None,
    ) -> None:
        """Create a ModelicaSystem instance. To define the model use model() or convertFmu2Mo().

        Args:
            work_directory: Path to a directory to be used for temporary
              files like the model executable. If left unspecified, a tmp
              directory will be created.
            session: definition of a (local) OMC session to be used. If
              unspecified, a new local session will be created.
        """

        self._quantities: list[dict[str, Any]] = []
        self._params: dict[str, str] = {}  # even numerical values are stored as str
        self._inputs: dict[str, list[tuple[float, float]]] = {}
        self._outputs: dict[str, np.float64] = {}  # numpy.float64 as it allows to define None values
        self._continuous: dict[str, np.float64] = {}  # numpy.float64 as it allows to define None values
        self._simulate_options: dict[str, str] = {}
        self._override_variables: dict[str, str] = {}
        self._simulate_options_override: dict[str, str] = {}
        self._linearization_options: dict[str, str] = {
            'startTime': str(0.0),
            'stopTime': str(1.0),
            'stepSize': str(0.002),
            'tolerance': str(1e-8),
        }
        self._optimization_options = self._linearization_options | {
            'numberOfIntervals': str(500),
        }
        self._linearized_inputs: list[str] = []  # linearization input list
        self._linearized_outputs: list[str] = []  # linearization output list
        self._linearized_states: list[str] = []  # linearization states list

        self._session = session

        # get OpenModelica version
        version_str = self._session.get_version()
        self._version = self._parse_om_version(version=version_str)

        self._simulated = False  # True if the model has already been simulated
        self._result_file: Optional[OMPathABC] = None  # for storing result file

        self._work_dir: OMPathABC = self.setWorkDirectory(work_directory)

        self._model_name: Optional[str] = None
        self._libraries: Optional[list[str | tuple[str, str]]] = None
        self._file_name: Optional[OMPathABC] = None
        self._variable_filter: Optional[str] = None

    def get_session(self) -> OMSessionABC:
        """
        Return the OMC session used for this class.
        """
        return self._session

    def get_model_name(self) -> str:
        """
        Return the defined model name.
        """
        if not isinstance(self._model_name, str):
            raise ModelicaSystemError("No model name defined!")

        return self._model_name

    def setWorkDirectory(self, work_directory: Optional[str | os.PathLike] = None) -> OMPathABC:
        """
        Define the work directory for the ModelicaSystem / OpenModelica session. The model is build within this
        directory. If no directory is defined a unique temporary directory is created.
        """
        if work_directory is not None:
            workdir = self._session.omcpath(work_directory).absolute()
            if not workdir.is_dir():
                raise IOError(f"Provided work directory does not exists: {work_directory}!")
        else:
            workdir = self._session.omcpath_tempdir().absolute()
            if not workdir.is_dir():
                raise IOError(f"{workdir} could not be created")

        logger.info("Define work dir as %s", workdir)
        self._session.set_workdir(workdir=workdir)

        # set the class variable _work_dir ...
        self._work_dir = workdir
        # ... and also return the defined path
        return workdir

    def getWorkDirectory(self) -> OMPathABC:
        """
        Return the defined working directory for this ModelicaSystem / OpenModelica session.
        """
        return self._work_dir

    def check_model_executable(self):
        """
        Check if the model executable is working
        """
        # check if the executable exists ...
        om_cmd = ModelExecutionCmd(
            runpath=self.getWorkDirectory(),
            cmd_local=self._session.model_execution_local,
            cmd_windows=self._session.model_execution_windows,
            cmd_prefix=self._session.model_execution_prefix(cwd=self.getWorkDirectory()),
            model_name=self._model_name,
        )
        # ... by running it - output help for command help
        om_cmd.arg_set(key="help", val="help")
        cmd_definition = om_cmd.definition()
        returncode = cmd_definition.run()
        if returncode != 0:
            raise ModelicaSystemError("Model executable not working!")

    def _xmlparse(self, xml_file: OMPathABC):
        if not xml_file.is_file():
            raise ModelicaSystemError(f"XML file not generated: {xml_file}")

        xml_content = xml_file.read_text()
        tree = ET.ElementTree(ET.fromstring(xml_content))
        root = tree.getroot()
        for attr in root.iter('DefaultExperiment'):
            for key in ("startTime", "stopTime", "stepSize", "tolerance",
                        "solver", "outputFormat"):
                self._simulate_options[key] = str(attr.get(key))

        for sv in root.iter('ScalarVariable'):
            translations = {
                "alias": "alias",
                "aliasvariable": "aliasVariable",
                "causality": "causality",
                "changeable": "isValueChangeable",
                "description": "description",
                "name": "name",
                "variability": "variability",
            }

            scalar: dict[str, Any] = {}
            for key_dst, key_src in translations.items():
                val = sv.get(key_src)
                scalar[key_dst] = None if val is None else str(val)

            ch = list(sv)
            for att in ch:
                scalar["start"] = att.get('start')
                scalar["min"] = att.get('min')
                scalar["max"] = att.get('max')
                scalar["unit"] = att.get('unit')

            # save parameters in the corresponding class variables
            if scalar["variability"] == "parameter":
                if scalar["name"] in self._override_variables:
                    self._params[scalar["name"]] = self._override_variables[scalar["name"]]
                else:
                    self._params[scalar["name"]] = scalar["start"]
            if scalar["variability"] == "continuous":
                self._continuous[scalar["name"]] = np.float64(scalar["start"])
            if scalar["causality"] == "input":
                self._inputs[scalar["name"]] = scalar["start"]
            if scalar["causality"] == "output":
                self._outputs[scalar["name"]] = np.float64(scalar["start"])

            self._quantities.append(scalar)

    def getQuantities(self, names: Optional[str | list[str]] = None) -> list[dict]:
        """
        This method returns list of dictionaries. It displays details of
        quantities such as name, value, changeable, and description.

        Examples:
            >>> mod.getQuantities()
            [
              {
                'alias': 'noAlias',
                'aliasvariable': None,
                'causality': 'local',
                'changeable': 'true',
                'description': None,
                'max': None,
                'min': None,
                'name': 'x',
                'start': '1.0',
                'unit': None,
                'variability': 'continuous',
              },
              {
                'name': 'der(x)',
                # ...
              },
              # ...
            ]

            >>> getQuantities("y")
            [{
              'name': 'y', # ...
            }]

            >>> getQuantities(["y","x"])
            [
              {
                'name': 'y', # ...
              },
              {
                'name': 'x', # ...
              }
            ]
        """
        if names is None:
            return self._quantities

        if isinstance(names, str):
            r = [x for x in self._quantities if x["name"] == names]
            if r == []:
                raise KeyError(names)
            return r

        if isinstance(names, list):
            return [x for y in names for x in self._quantities if x["name"] == y]

        raise ModelicaSystemError("Unhandled input for getQuantities()")

    def getContinuousInitial(
            self,
            names: Optional[str | list[str]] = None,
    ) -> dict[str, np.float64] | list[np.float64]:
        """
        Get (initial) values of continuous signals.

        Args:
            names: Either None (default), a string with the continuous signal
              name, or a list of signal name strings.
        Returns:
            If `names` is None, a dict in the format
            {signal_name: signal_value} is returned.
            If `names` is a string, a single element list [signal_value] is
            returned.
            If `names` is a list, a list with one value for each signal name
            in names is returned: [signal1_value, signal2_value, ...].

        Examples:
            >>> mod.getContinuousInitial()
            {'x': '1.0', 'der(x)': None, 'y': '-0.4'}
            >>> mod.getContinuousInitial("y")
            ['-0.4']
            >>> mod.getContinuousInitial(["y","x"])
            ['-0.4', '1.0']
        """
        if names is None:
            return self._continuous
        if isinstance(names, str):
            return [self._continuous[names]]
        if isinstance(names, list):
            return [self._continuous[x] for x in names]

        raise ModelicaSystemError("Unhandled input for getContinousInitial()")

    def getParameters(
            self,
            names: Optional[str | list[str]] = None,
    ) -> dict[str, str] | list[str]:
        """Get parameter values.

        Args:
            names: Either None (default), a string with the parameter name,
              or a list of parameter name strings.
        Returns:
            If `names` is None, a dict in the format
            {parameter_name: parameter_value} is returned.
            If `names` is a string, a single element list is returned.
            If `names` is a list, a list with one value for each parameter name
            in names is returned.
            In all cases, parameter values are returned as strings.

        Examples:
            >>> mod.getParameters()
            {'Name1': '1.23', 'Name2': '4.56'}
            >>> mod.getParameters("Name1")
            ['1.23']
            >>> mod.getParameters(["Name1","Name2"])
            ['1.23', '4.56']
        """
        if names is None:
            return self._params
        if isinstance(names, str):
            return [self._params[names]]
        if isinstance(names, list):
            return [self._params[x] for x in names]

        raise ModelicaSystemError("Unhandled input for getParameters()")

    def getInputs(
            self,
            names: Optional[str | list[str]] = None,
    ) -> dict[str, list[tuple[float, float]]] | list[list[tuple[float, float]]]:
        """Get values of input signals.

        Args:
            names: Either None (default), a string with the input name,
              or a list of input name strings.
        Returns:
            If `names` is None, a dict in the format
            {input_name: input_value} is returned.
            If `names` is a string, a single element list [input_value] is
            returned.
            If `names` is a list, a list with one value for each input name
            in names is returned: [input1_values, input2_values, ...].
            In all cases, input values are returned as a list of tuples,
            where the first element in the tuple is the time and the second
            element is the input value.

        Examples:
            >>> mod.getInputs()
            {'Name1': [(0.0, 0.0), (1.0, 1.0)], 'Name2': None}
            >>> mod.getInputs("Name1")
            [[(0.0, 0.0), (1.0, 1.0)]]
            >>> mod.getInputs(["Name1","Name2"])
            [[(0.0, 0.0), (1.0, 1.0)], None]
        """
        if names is None:
            return self._inputs
        if isinstance(names, str):
            return [self._inputs[names]]
        if isinstance(names, list):
            return [self._inputs[x] for x in names]

        raise ModelicaSystemError("Unhandled input for getInputs()")

    def getOutputsInitial(
            self,
            names: Optional[str | list[str]] = None,
    ) -> dict[str, np.float64] | list[np.float64]:
        """
        Get (initial) values of output signals.

        Args:
            names: Either None (default), a string with the output name,
              or a list of output name strings.
        Returns:
            If `names` is None, a dict in the format
            {output_name: output_value} is returned.
            If `names` is a string, a single element list [output_value] is
            returned.
            If `names` is a list, a list with one value for each output name
            in names is returned: [output1_value, output2_value, ...].

        Examples:
            >>> mod.getOutputsInitial()
            {'out1': '-0.4', 'out2': '1.2'}
            >>> mod.getOutputsInitial("out1")
            ['-0.4']
            >>> mod.getOutputsInitial(["out1","out2"])
            ['-0.4', '1.2']
        """
        if names is None:
            return self._outputs
        if isinstance(names, str):
            return [self._outputs[names]]
        if isinstance(names, list):
            return [self._outputs[x] for x in names]

        raise ModelicaSystemError("Unhandled input for getOutputsInitial()")

    def getSimulationOptions(
            self,
            names: Optional[str | list[str]] = None,
    ) -> dict[str, str] | list[str]:
        """Get simulation options such as stopTime and tolerance.

        Args:
            names: Either None (default), a string with the simulation option
              name, or a list of option name strings.

        Returns:
            If `names` is None, a dict in the format
            {option_name: option_value} is returned.
            If `names` is a string, a single element list [option_value] is
            returned.
            If `names` is a list, a list with one value for each option name
            in names is returned: [option1_value, option2_value, ...].
            Option values are always returned as strings.

        Examples:
            >>> mod.getSimulationOptions()
            {'startTime': '0', 'stopTime': '1.234',
             'stepSize': '0.002', 'tolerance': '1.1e-08', 'solver': 'dassl', 'outputFormat': 'mat'}
            >>> mod.getSimulationOptions("stopTime")
            ['1.234']
            >>> mod.getSimulationOptions(["tolerance", "stopTime"])
            ['1.1e-08', '1.234']
        """
        if names is None:
            return self._simulate_options
        if isinstance(names, str):
            return [self._simulate_options[names]]
        if isinstance(names, list):
            return [self._simulate_options[x] for x in names]

        raise ModelicaSystemError("Unhandled input for getSimulationOptions()")

    def getLinearizationOptions(
            self,
            names: Optional[str | list[str]] = None,
    ) -> dict[str, str] | list[str]:
        """Get simulation options used for linearization.

        Args:
            names: Either None (default), a string with the linearization option
              name, or a list of option name strings.

        Returns:
            If `names` is None, a dict in the format
            {option_name: option_value} is returned.
            If `names` is a string, a single element list [option_value] is
            returned.
            If `names` is a list, a list with one value for each option name
            in names is returned: [option1_value, option2_value, ...].

            The option values are always returned as strings.

        Examples:
            >>> mod.getLinearizationOptions()
            {'startTime': '0.0', 'stopTime': '1.0', 'stepSize': '0.002', 'tolerance': '1e-08'}
            >>> mod.getLinearizationOptions("stopTime")
            ['1.0']
            >>> mod.getLinearizationOptions(["tolerance", "stopTime"])
            ['1e-08', '1.0']
        """
        if names is None:
            return self._linearization_options
        if isinstance(names, str):
            return [self._linearization_options[names]]
        if isinstance(names, list):
            return [self._linearization_options[x] for x in names]

        raise ModelicaSystemError("Unhandled input for getLinearizationOptions()")

    def getOptimizationOptions(
            self,
            names: Optional[str | list[str]] = None,
    ) -> dict[str, str] | list[str]:
        """Get simulation options used for optimization.

        Args:
            names: Either None (default), a string with the optimization option
              name, or a list of option name strings.

        Returns:
            If `names` is None, a dict in the format
            {option_name: option_value} is returned.
            If `names` is a string, a single element list [option_value] is
            returned.
            If `names` is a list, a list with one value for each option name
            in names is returned: [option1_value, option2_value, ...].

            The option values are always returned as string.

        Examples:
            >>> mod.getOptimizationOptions()
            {'startTime': 0.0, 'stopTime': 1.0, 'numberOfIntervals': 500, 'stepSize': 0.002, 'tolerance': 1e-08}
            >>> mod.getOptimizationOptions("stopTime")
            [1.0]
            >>> mod.getOptimizationOptions(["tolerance", "stopTime"])
            [1e-08, 1.0]
        """
        if names is None:
            return self._optimization_options
        if isinstance(names, str):
            return [self._optimization_options[names]]
        if isinstance(names, list):
            return [self._optimization_options[x] for x in names]

        raise ModelicaSystemError("Unhandled input for getOptimizationOptions()")

    @staticmethod
    def _parse_om_version(version: str) -> tuple[int, int, int]:
        """
        Evaluate an OMC version string and return a tuple of (epoch, major, minor).
        """
        match = re.search(pattern=r"v?(\d+)\.(\d+)\.(\d+)", string=version)
        if not match:
            raise ValueError(f"Version not found in: {version}")
        major, minor, patch = map(int, match.groups())

        return major, minor, patch

    def _process_override_data(
            self,
            om_cmd: ModelExecutionCmd,
            override_file: OMPathABC,
            override_var: dict[str, str],
            override_sim: dict[str, str],
    ) -> None:
        """
        Define the override parameters. As the definition of simulation specific override parameter changes with OM
        1.26.0, version specific code is needed. Please keep in mind, that this will fail if OMC is not used to run the
        model executable.
        """
        if len(override_var) == 0 and len(override_sim) == 0:
            return

        override_content = ""
        if override_var:
            override_content += "\n".join([f"{key}={value}" for key, value in override_var.items()]) + "\n"

        # simulation options are not read from override file from version >= 1.26.0,
        # pass them to simulation executable directly as individual arguments
        # see https://github.com/OpenModelica/OpenModelica/pull/14813
        if override_sim:
            if self._version >= (1, 26, 0):
                for key, opt_value in override_sim.items():
                    om_cmd.arg_set(key=key, val=str(opt_value))
            else:
                override_content += "\n".join([f"{key}={value}" for key, value in override_sim.items()]) + "\n"

        if override_content:
            override_file.write_text(override_content)
            om_cmd.arg_set(key="overrideFile", val=override_file.as_posix())

    def simulate_cmd(
            self,
            result_file: OMPathABC,
            simflags: Optional[str] = None,
            simargs: Optional[dict[str, Optional[str | dict[str, Any] | numbers.Number]]] = None,
    ) -> ModelExecutionCmd:
        """
        This method prepares the simulates model according to the simulation options. It returns an instance of
        ModelicaSystemCmd which can be used to run the simulation.

        Due to the tempdir being unique for the ModelicaSystem instance, *NEVER* use this to create several simulations
        with the same instance of ModelicaSystem! Restart each simulation process with a new instance of ModelicaSystem.

        However, if only non-structural parameters are used, it is possible to reuse an existing instance of
        ModelicaSystem to create several version ModelicaSystemCmd to run the model using different settings.

        Parameters
        ----------
        result_file
        simflags
        simargs

        Returns
        -------
            An instance if ModelicaSystemCmd to run the requested simulation.
        """

        om_cmd = ModelExecutionCmd(
            runpath=self.getWorkDirectory(),
            cmd_local=self._session.model_execution_local,
            cmd_windows=self._session.model_execution_windows,
            cmd_prefix=self._session.model_execution_prefix(cwd=self.getWorkDirectory()),
            model_name=self._model_name,
        )

        # always define the result file to use
        om_cmd.arg_set(key="r", val=result_file.as_posix())

        # allow runtime simulation flags from user input
        if simflags is not None:
            om_cmd.args_set(args=om_cmd.parse_simflags(simflags=simflags))

        if simargs:
            om_cmd.args_set(args=simargs)

        self._process_override_data(
            om_cmd=om_cmd,
            override_file=result_file.parent / f"{result_file.stem}_override.txt",
            override_var=self._override_variables,
            override_sim=self._simulate_options_override,
        )

        if self._inputs:  # if model has input quantities
            for key, val in self._inputs.items():
                if val is None:
                    val = [(float(self._simulate_options["startTime"]), 0.0),
                           (float(self._simulate_options["stopTime"]), 0.0)]
                    self._inputs[key] = val
                if float(self._simulate_options["startTime"]) != val[0][0]:
                    raise ModelicaSystemError(f"startTime not matched for Input {key}!")
                if float(self._simulate_options["stopTime"]) != val[-1][0]:
                    raise ModelicaSystemError(f"stopTime not matched for Input {key}!")

            # csvfile is based on name used for result file
            csvfile = result_file.parent / f"{result_file.stem}.csv"
            # write csv file and store the name
            csvfile = self._createCSVData(csvfile=csvfile)

            om_cmd.arg_set(key="csvInput", val=csvfile.as_posix())

        return om_cmd

    def simulate(
            self,
            resultfile: Optional[str | os.PathLike] = None,
            simflags: Optional[str] = None,
            simargs: Optional[dict[str, Optional[str | dict[str, Any] | numbers.Number]]] = None,
    ) -> None:
        """Simulate the model according to simulation options.

        See setSimulationOptions().

        Args:
            resultfile: Path to a custom result file
            simflags: String of extra command line flags for the model binary.
              This argument is deprecated, use simargs instead.
            simargs: Dict with simulation runtime flags.

        Examples:
            mod.simulate()
            mod.simulate(resultfile="a.mat")
            # set runtime simulation flags, deprecated
            mod.simulate(simflags="-noEventEmit -noRestart -override=e=0.3,g=10")
            # using simargs
            mod.simulate(simargs={"noEventEmit": None, "noRestart": None, "override": "override": {"e": 0.3, "g": 10}})
        """

        if resultfile is None:
            # default result file generated by OM
            self._result_file = self.getWorkDirectory() / f"{self._model_name}_res.mat"
        elif isinstance(resultfile, OMPathABC):
            self._result_file = resultfile
        else:
            self._result_file = self._session.omcpath(resultfile)
            if not self._result_file.is_absolute():
                self._result_file = self.getWorkDirectory() / resultfile

        if not isinstance(self._result_file, OMPathABC):
            raise ModelicaSystemError(f"Invalid result file path: {self._result_file} - must be an OMCPath object!")

        om_cmd = self.simulate_cmd(
            result_file=self._result_file,
            simflags=simflags,
            simargs=simargs,
        )

        # delete resultfile ...
        if self._result_file.is_file():
            self._result_file.unlink()
        # ... run simulation ...
        cmd_definition = om_cmd.definition()
        returncode = cmd_definition.run()
        # and check returncode *AND* resultfile
        if returncode != 0 and self._result_file.is_file():
            # check for an empty (=> 0B) result file which indicates a crash of the model executable
            # see: https://github.com/OpenModelica/OMPython/issues/261
            #      https://github.com/OpenModelica/OpenModelica/issues/13829
            if self._result_file.size() == 0:
                self._result_file.unlink()
                raise ModelicaSystemError("Empty result file - this indicates a crash of the model executable!")

            logger.warning(f"Return code = {returncode} but result file exists!")

        self._simulated = True

    @staticmethod
    def _prepare_input_data(
            input_args: Any,
            input_kwargs: dict[str, Any],
    ) -> dict[str, str]:
        """
        Convert raw input to a structured dictionary {'key1': 'value1', 'key2': 'value2'}.
        """

        def prepare_str(str_in: str) -> dict[str, str]:
            str_in = str_in.replace(" ", "")
            key_val_list: list[str] = str_in.split("=")
            if len(key_val_list) != 2:
                raise ModelicaSystemError(f"Invalid 'key=value' pair: {str_in}")

            input_data_from_str: dict[str, str] = {key_val_list[0]: key_val_list[1]}

            return input_data_from_str

        input_data: dict[str, str] = {}

        for input_arg in input_args:
            if isinstance(input_arg, str):
                warnings.warn(message="The definition of values to set should use a dictionary, "
                                      "i.e. {'key1': 'val1', 'key2': 'val2', ...}. Please convert all cases which "
                                      "use a string ('key=val') or list ['key1=val1', 'key2=val2', ...]",
                              category=DeprecationWarning,
                              stacklevel=3)
                input_data = input_data | prepare_str(input_arg)
            elif isinstance(input_arg, list):
                warnings.warn(message="The definition of values to set should use a dictionary, "
                                      "i.e. {'key1': 'val1', 'key2': 'val2', ...}. Please convert all cases which "
                                      "use a string ('key=val') or list ['key1=val1', 'key2=val2', ...]",
                              category=DeprecationWarning,
                              stacklevel=3)

                for item in input_arg:
                    if not isinstance(item, str):
                        raise ModelicaSystemError(f"Invalid input data type for set*() function: {type(item)}!")
                    input_data = input_data | prepare_str(item)
            elif isinstance(input_arg, dict):
                input_data = input_data | input_arg
            else:
                raise ModelicaSystemError(f"Invalid input data type for set*() function: {type(input_arg)}!")

        if len(input_kwargs):
            for key, val in input_kwargs.items():
                # ensure all values are strings to align it on one type: dict[str, str]
                if not isinstance(val, str):
                    # spaces have to be removed as setInput() could take list of tuples as input and spaces would
                    # result in an error on recreating the input data
                    str_val = str(val).replace(' ', '')
                else:
                    str_val = val
                if ' ' in key or ' ' in str_val:
                    raise ModelicaSystemError(f"Spaces not allowed in key/value pairs: {repr(key)} = {repr(val)}!")
                input_data[key] = str_val

        return input_data

    def _set_method_helper(
            self,
            inputdata: dict[str, str],
            classdata: dict[str, Any],
            datatype: str,
            overridedata: Optional[dict[str, str]] = None,
    ) -> bool:
        """
        Helper function for:
        * setParameter()
        * setContinuous()
        * setSimulationOptions()
        * setLinearizationOption()
        * setOptimizationOption()
        * setInputs()

        Parameters
        ----------
        inputdata
            string or list of string given by user
        classdata
            dict() containing the values of different variables (eg: parameter, continuous, simulation parameters)
        datatype
            type identifier (eg; continuous, parameter, simulation, linearization, optimization)
        overridedata
            dict() which stores the new override variables list,
        """

        for key, val in inputdata.items():
            if key not in classdata:
                raise ModelicaSystemError(f"Invalid variable for type {repr(datatype)}: {repr(key)}")

            if datatype == "parameter" and not self.isParameterChangeable(key):
                raise ModelicaSystemError(f"It is not possible to set the parameter {repr(key)}. It seems to be "
                                          "structural, final, protected, evaluated or has a non-constant binding. "
                                          "Use sendExpression(...) and rebuild the model using buildModel() API; "
                                          "command to set the parameter before rebuilding the model: "
                                          "sendExpression(expr=\"setParameterValue("
                                          f"{self._model_name}, {key}, {val if val is not None else '<?value?>'}"
                                          ")\").")

            classdata[key] = val
            if overridedata is not None:
                overridedata[key] = val

        return True

    def isParameterChangeable(
            self,
            name: str,
    ) -> bool:
        """
        Return if the parameter defined by name is changeable (= non-structural; can be modified without the need to
        recompile the model).
        """
        q = self.getQuantities(name)
        if q[0]["changeable"] == "false":
            return False
        return True

    def setContinuous(
            self,
            *args: Any,
            **kwargs: dict[str, Any],
    ) -> bool:
        """
        This method is used to set continuous values. It can be called:
        with a sequence of continuous name and assigning corresponding values as arguments as show in the example below:
        usage
        >>> setContinuous("Name=value")  # depreciated
        >>> setContinuous(["Name1=value1","Name2=value2"])  # depreciated

        >>> setContinuous(Name1="value1", Name2="value2")
        >>> param = {"Name1": "value1", "Name2": "value2"}
        >>> setContinuous(**param)
        """
        inputdata = self._prepare_input_data(input_args=args, input_kwargs=kwargs)

        return self._set_method_helper(
            inputdata=inputdata,
            classdata=self._continuous,
            datatype="continuous",
            overridedata=self._override_variables)

    def setParameters(
            self,
            *args: Any,
            **kwargs: dict[str, Any],
    ) -> bool:
        """
        This method is used to set parameter values. It can be called:
        with a sequence of parameter name and assigning corresponding value as arguments as show in the example below:
        usage
        >>> setParameters("Name=value")  # depreciated
        >>> setParameters(["Name1=value1","Name2=value2"])  # depreciated

        >>> setParameters(Name1="value1", Name2="value2")
        >>> param = {"Name1": "value1", "Name2": "value2"}
        >>> setParameters(**param)
        """
        inputdata = self._prepare_input_data(input_args=args, input_kwargs=kwargs)

        return self._set_method_helper(
            inputdata=inputdata,
            classdata=self._params,
            datatype="parameter",
            overridedata=self._override_variables)

    def setSimulationOptions(
            self,
            *args: Any,
            **kwargs: dict[str, Any],
    ) -> bool:
        """
        This method is used to set simulation options. It can be called:
        with a sequence of simulation options name and assigning corresponding values as arguments as show in the
        example below:
        usage
        >>> setSimulationOptions("Name=value")  # depreciated
        >>> setSimulationOptions(["Name1=value1","Name2=value2"])  # depreciated

        >>> setSimulationOptions(Name1="value1", Name2="value2")
        >>> param = {"Name1": "value1", "Name2": "value2"}
        >>> setSimulationOptions(**param)
        """
        inputdata = self._prepare_input_data(input_args=args, input_kwargs=kwargs)

        return self._set_method_helper(
            inputdata=inputdata,
            classdata=self._simulate_options,
            datatype="simulation-option",
            overridedata=self._simulate_options_override)

    def setLinearizationOptions(
            self,
            *args: Any,
            **kwargs: dict[str, Any],
    ) -> bool:
        """
        This method is used to set linearization options. It can be called:
        with a sequence of linearization options name and assigning corresponding value as arguments as show in the
        example below
        usage
        >>> setLinearizationOptions("Name=value")  # depreciated
        >>> setLinearizationOptions(["Name1=value1","Name2=value2"])  # depreciated

        >>> setLinearizationOptions(Name1="value1", Name2="value2")
        >>> param = {"Name1": "value1", "Name2": "value2"}
        >>> setLinearizationOptions(**param)
        """
        inputdata = self._prepare_input_data(input_args=args, input_kwargs=kwargs)

        return self._set_method_helper(
            inputdata=inputdata,
            classdata=self._linearization_options,
            datatype="Linearization-option",
            overridedata=None)

    def setOptimizationOptions(
            self,
            *args: Any,
            **kwargs: dict[str, Any],
    ) -> bool:
        """
        This method is used to set optimization options. It can be called:
        with a sequence of optimization options name and assigning corresponding values as arguments as show in the
        example below:
        usage
        >>> setOptimizationOptions("Name=value")  # depreciated
        >>> setOptimizationOptions(["Name1=value1","Name2=value2"])  # depreciated

        >>> setOptimizationOptions(Name1="value1", Name2="value2")
        >>> param = {"Name1": "value1", "Name2": "value2"}
        >>> setOptimizationOptions(**param)
        """
        inputdata = self._prepare_input_data(input_args=args, input_kwargs=kwargs)

        return self._set_method_helper(
            inputdata=inputdata,
            classdata=self._optimization_options,
            datatype="optimization-option",
            overridedata=None)

    def setInputs(
            self,
            *args: Any,
            **kwargs: dict[str, Any],
    ) -> bool:
        """
        This method is used to set input values. It can be called with a sequence of input name and assigning
        corresponding values as arguments as show in the example below. Compared to other set*() methods this is a
        special case as value could be a list of tuples - these are converted to a string in _prepare_input_data()
        and restored here via ast.literal_eval().

        >>> setInputs("Name=value")  # depreciated
        >>> setInputs(["Name1=value1","Name2=value2"])  # depreciated

        >>> setInputs(Name1="value1", Name2="value2")
        >>> param = {"Name1": "value1", "Name2": "value2"}
        >>> setInputs(**param)
        """
        inputdata = self._prepare_input_data(input_args=args, input_kwargs=kwargs)

        for key, val in inputdata.items():
            if key not in self._inputs:
                raise ModelicaSystemError(f"{key} is not an input")

            if not isinstance(val, str):
                raise ModelicaSystemError(f"Invalid data in input for {repr(key)}: {repr(val)}")

            val_evaluated = ast.literal_eval(val)

            if isinstance(val_evaluated, (int, float)):
                self._inputs[key] = [(float(self._simulate_options["startTime"]), float(val)),
                                     (float(self._simulate_options["stopTime"]), float(val))]
            elif isinstance(val_evaluated, list):
                if not all([isinstance(item, tuple) for item in val_evaluated]):
                    raise ModelicaSystemError("Value for setInput() must be in tuple format; "
                                              f"got {repr(val_evaluated)}")
                if val_evaluated != sorted(val_evaluated, key=lambda x: x[0]):
                    raise ModelicaSystemError("Time value should be in increasing order; "
                                              f"got {repr(val_evaluated)}")

                for item in val_evaluated:
                    if item[0] < float(self._simulate_options["startTime"]):
                        raise ModelicaSystemError(f"Time value in {repr(item)} of {repr(val_evaluated)} is less "
                                                  "than the simulation start time")
                    if len(item) != 2:
                        raise ModelicaSystemError(f"Value {repr(item)} of {repr(val_evaluated)} "
                                                  "is in incorrect format!")

                self._inputs[key] = val_evaluated
            else:
                raise ModelicaSystemError(f"Data cannot be evaluated for {repr(key)}: {repr(val)}")

        return True

    def _createCSVData(self, csvfile: Optional[OMPathABC] = None) -> OMPathABC:
        """
        Create a csv file with inputs for the simulation/optimization of the model. If csvfile is provided as argument,
        this file is used; else a generic file name is created.
        """
        start_time: float = float(self._simulate_options["startTime"])
        stop_time: float = float(self._simulate_options["stopTime"])

        # Replace None inputs with a default constant zero signal
        inputs: dict[str, list[tuple[float, float]]] = {}
        for input_name, input_signal in self._inputs.items():
            if input_signal is None:
                inputs[input_name] = [(start_time, 0.0), (stop_time, 0.0)]
            else:
                inputs[input_name] = input_signal

        # Collect all unique timestamps across all input signals
        all_times = np.array(
            sorted({t for signal in inputs.values() for t, _ in signal}),
            dtype=float
        )

        # Interpolate missing values
        interpolated_inputs: dict[str, np.ndarray] = {}
        for signal_name, signal_values in inputs.items():
            signal = np.array(signal_values)
            interpolated_inputs[signal_name] = np.interp(
                x=all_times,
                xp=signal[:, 0],  # times
                fp=signal[:, 1],  # values
            )

        # Write CSV file
        input_names = list(interpolated_inputs.keys())
        header = ['time'] + input_names + ['end']

        csv_rows = [header]
        for i, t in enumerate(all_times):
            row = [
                t,  # time
                *(interpolated_inputs[name][i] for name in input_names),  # input values
                0,  # trailing 'end' column
            ]
            csv_rows.append(row)

        if csvfile is None:
            csvfile = self.getWorkDirectory() / f'{self._model_name}.csv'

        # basic definition of a CSV file using csv_rows as input
        csv_content = "\n".join([",".join(map(str, row)) for row in csv_rows]) + "\n"

        csvfile.write_text(csv_content)

        return csvfile

    def linearize(
            self,
            lintime: Optional[float] = None,
            simflags: Optional[str] = None,
            simargs: Optional[dict[str, Optional[str | dict[str, Any] | numbers.Number]]] = None,
    ) -> LinearizationResult:
        """Linearize the model according to linearization options.

        See setLinearizationOptions.

        Args:
            lintime: Override "stopTime" value.
            simflags: String of extra command line flags for the model binary.
              This argument is deprecated, use simargs instead.
            simargs: A dict with command line flags and possible options; example: "simargs={'csvInput': 'a.csv'}"

        Returns:
            A LinearizationResult object is returned. This allows several
            uses:
            * `(A, B, C, D) = linearize()` to get just the matrices,
            * `result = linearize(); result.A` to get everything and access the
              attributes one by one,
            * `result = linearize(); A = result[0]` mostly just for backwards
              compatibility, because linearize() used to return `[A, B, C, D]`.
        """
        if len(self._quantities) == 0:
            # if self._quantities has no content, the xml file was not parsed; see self._xmlparse()
            raise ModelicaSystemError(
                "Linearization cannot be performed as the model is not build, "
                "use ModelicaSystemOMC() to build the model first"
            )

        om_cmd = ModelExecutionCmd(
            runpath=self.getWorkDirectory(),
            cmd_local=self._session.model_execution_local,
            cmd_windows=self._session.model_execution_windows,
            cmd_prefix=self._session.model_execution_prefix(cwd=self.getWorkDirectory()),
            model_name=self._model_name,
        )

        self._process_override_data(
            om_cmd=om_cmd,
            override_file=self.getWorkDirectory() / f'{self._model_name}_override_linear.txt',
            override_var=self._override_variables,
            override_sim=self._linearization_options,
        )

        if self._inputs:
            for data in self._inputs.values():
                if data is not None:
                    for value in data:
                        if value[0] < float(self._simulate_options["startTime"]):
                            raise ModelicaSystemError('Input time value is less than simulation startTime')
            csvfile = self._createCSVData()
            om_cmd.arg_set(key="csvInput", val=csvfile.as_posix())

        if lintime is None:
            lintime = float(self._linearization_options["stopTime"])
        if (float(self._linearization_options["startTime"]) > lintime
                or float(self._linearization_options["stopTime"]) < lintime):
            raise ModelicaSystemError(f"Invalid linearisation time: {lintime=}; "
                                      f"expected value: {self._linearization_options['startTime']} "
                                      f"<= lintime <= {self._linearization_options['stopTime']}")
        om_cmd.arg_set(key="l", val=str(lintime))

        # allow runtime simulation flags from user input
        if simflags is not None:
            om_cmd.args_set(args=om_cmd.parse_simflags(simflags=simflags))

        if simargs:
            om_cmd.args_set(args=simargs)

        # the file create by the model executable which contains the matrix and linear inputs, outputs and states
        linear_file = self.getWorkDirectory() / "linearized_model.py"
        linear_file.unlink(missing_ok=True)

        cmd_definition = om_cmd.definition()
        returncode = cmd_definition.run()
        if returncode != 0:
            raise ModelicaSystemError(f"Linearize failed with return code: {returncode}")
        if not linear_file.is_file():
            raise ModelicaSystemError(f"Linearization failed: {linear_file} not found!")

        self._simulated = True

        # extract data from the python file with the linearized model using the ast module - this allows to get the
        # needed information without executing the created code
        linear_data = {}
        linear_file_content = linear_file.read_text()
        try:
            # ignore possible typing errors below (mypy) - these are caught by the try .. except .. block
            linear_file_ast = ast.parse(linear_file_content)
            for body_part in linear_file_ast.body[0].body:  # type: ignore
                if not isinstance(body_part, ast.Assign):
                    continue

                target = body_part.targets[0].id  # type: ignore
                value_ast = ast.literal_eval(body_part.value)

                linear_data[target] = value_ast
        except (AttributeError, IndexError, ValueError, SyntaxError, TypeError) as ex:
            raise ModelicaSystemError(f"Error parsing linearization file {linear_file}: {ex}") from ex

        # remove the file
        linear_file.unlink()

        self._linearized_inputs = linear_data["inputVars"]
        self._linearized_outputs = linear_data["outputVars"]
        self._linearized_states = linear_data["stateVars"]

        return LinearizationResult(
            n=linear_data["n"],
            m=linear_data["m"],
            p=linear_data["p"],
            x0=linear_data["x0"],
            u0=linear_data["u0"],
            A=linear_data["A"],
            B=linear_data["B"],
            C=linear_data["C"],
            D=linear_data["D"],
            stateVars=linear_data["stateVars"],
            inputVars=linear_data["inputVars"],
            outputVars=linear_data["outputVars"],
        )

    def getLinearInputs(self) -> list[str]:
        """Get names of input variables of the linearized model."""
        return self._linearized_inputs

    def getLinearOutputs(self) -> list[str]:
        """Get names of output variables of the linearized model."""
        return self._linearized_outputs

    def getLinearStates(self) -> list[str]:
        """Get names of state variables of the linearized model."""
        return self._linearized_states


class ModelicaSystemOMC(ModelicaSystemABC):
    """
    Class to simulate a Modelica model using OpenModelica via OMCSession.
    """

    def __init__(
            self,
            command_line_options: Optional[list[str]] = None,
            work_directory: Optional[str | os.PathLike] = None,
            omhome: Optional[str] = None,
            session: Optional[OMSessionABC] = None,
    ) -> None:
        """Create a ModelicaSystem instance. To define the model use model() or convertFmu2Mo().

        Args:
            command_line_options: List with extra command line options as elements. The list elements are
              provided to omc via setCommandLineOptions(). If set, the default values will be overridden.
              To disable any command line options, use an empty list.
            work_directory: Path to a directory to be used for temporary
              files like the model executable. If left unspecified, a tmp
              directory will be created.
            omhome: path to OMC to be used when creating the OMC session (see OMCSession).
            session: definition of a (local) OMC session to be used. If
              unspecified, a new local session will be created.
        """

        if session is None:
            session = OMCSessionLocal(omhome=omhome)

        super().__init__(
            session=session,
            work_directory=work_directory,
        )

        # set commandLineOptions using default values or the user defined list
        if command_line_options is None:
            # set default command line options to improve the performance of linearization and to avoid recompilation if
            # the simulation executable is reused in linearize() via the runtime flag '-l'
            command_line_options = [
                "--linearizationDumpLanguage=python",
                "--generateSymbolicLinearization",
            ]
        for opt in command_line_options:
            self.set_command_line_options(command_line_option=opt)

    def model(
            self,
            model_name: Optional[str] = None,
            model_file: Optional[str | os.PathLike] = None,
            libraries: Optional[list[str | tuple[str, str]]] = None,
            variable_filter: Optional[str] = None,
            build: bool = True,
    ) -> None:
        """Load and build a Modelica model.

        This method loads the model file and builds it if requested (build == True).

        Args:
            model_file: Path to the model file. Either absolute or relative to
              the current working directory.
            model_name: The name of the model class. If it is contained within
              a package, "PackageName.ModelName" should be used.
            libraries: List of libraries to be loaded before the model itself is
              loaded. Two formats are supported for the list elements:
              lmodel=["Modelica"] for just the library name
              and lmodel=[("Modelica","3.2.3")] for specifying both the name
              and the version.
            variable_filter: A regular expression. Only variables fully
              matching the regexp will be stored in the result file.
              Leaving it unspecified is equivalent to ".*".
            build: Boolean controlling whether the model should be
              built when constructor is called. If False, the constructor
              simply loads the model without compiling.

        Examples:
            mod = ModelicaSystemOMC()
            # and then one of the lines below
            mod.model(name="modelName", file="ModelicaModel.mo", )
            mod.model(name="modelName", file="ModelicaModel.mo", libraries=["Modelica"])
            mod.model(name="modelName", file="ModelicaModel.mo", libraries=[("Modelica","3.2.3"), "PowerSystems"])
        """

        if self._model_name is not None:
            raise ModelicaSystemError("Can not reuse this instance of ModelicaSystem "
                                      f"defined for {repr(self._model_name)}!")

        if model_name is None or not isinstance(model_name, str):
            raise ModelicaSystemError("A model name must be provided!")

        if libraries is None:
            libraries = []

        if not isinstance(libraries, list):
            raise ModelicaSystemError(f"Invalid input type for libraries: {type(libraries)} - list expected!")

        # set variables
        self._model_name = model_name  # Model class name
        self._libraries = libraries  # may be needed if model is derived from other model
        self._variable_filter = variable_filter

        if self._libraries:
            self._loadLibrary(libraries=self._libraries)

        self._file_name = None
        if model_file is not None:
            file_path = pathlib.Path(model_file)
            # special handling for OMCProcessLocal - consider a relative path
            if isinstance(self._session, OMCSessionLocal) and not file_path.is_absolute():
                file_path = pathlib.Path.cwd() / file_path
            if not file_path.is_file():
                raise IOError(f"Model file {file_path} does not exist!")

            self._file_name = self.getWorkDirectory() / file_path.name
            if (isinstance(self._session, OMCSessionLocal)
                    and file_path.as_posix() == self._file_name.as_posix()):
                pass
            elif self._file_name.is_file():
                raise IOError(f"Simulation model file {self._file_name} exist - not overwriting!")
            else:
                content = file_path.read_text(encoding='utf-8')
                self._file_name.write_text(content)

        if self._file_name is not None:
            self._loadFile(fileName=self._file_name)

        if build:
            self.buildModel(variable_filter)

    def set_command_line_options(self, command_line_option: str):
        """
        Set the provided command line option via OMC setCommandLineOptions().
        """
        expr = f'setCommandLineOptions("{command_line_option}")'
        self.sendExpression(expr=expr)

    def _loadFile(self, fileName: OMPathABC):
        # load file
        self.sendExpression(expr=f'loadFile("{fileName.as_posix()}")')

    # for loading file/package, loading model and building model
    def _loadLibrary(self, libraries: list):
        # load Modelica standard libraries or Modelica files if needed
        for element in libraries:
            if element is not None:
                if isinstance(element, str):
                    if element.endswith(".mo"):
                        api_call = "loadFile"
                    else:
                        api_call = "loadModel"
                    self._requestApi(apiName=api_call, entity=element)
                elif isinstance(element, tuple):
                    if not element[1]:
                        expr_load_lib = f"loadModel({element[0]})"
                    else:
                        expr_load_lib = f'loadModel({element[0]}, {{"{element[1]}"}})'
                    self.sendExpression(expr=expr_load_lib)
                else:
                    raise ModelicaSystemError("loadLibrary() failed, Unknown type detected: "
                                              f"{element} is of type {type(element)}, "
                                              "The following patterns are supported:\n"
                                              '1)["Modelica"]\n'
                                              '2)[("Modelica","3.2.3"), "PowerSystems"]\n')

    def buildModel(self, variableFilter: Optional[str] = None):
        filter_def: Optional[str] = None
        if variableFilter is not None:
            filter_def = variableFilter
        elif self._variable_filter is not None:
            filter_def = self._variable_filter

        if filter_def is not None:
            var_filter = f'variableFilter="{filter_def}"'
        else:
            var_filter = 'variableFilter=".*"'

        build_model_result = self._requestApi(apiName="buildModel", entity=self._model_name, properties=var_filter)
        logger.debug("OM model build result: %s", build_model_result)

        # check if the executable exists ...
        self.check_model_executable()

        xml_file = self._session.omcpath(build_model_result[0]).parent / build_model_result[1]
        self._xmlparse(xml_file=xml_file)

    def sendExpression(self, expr: str, parsed: bool = True) -> Any:
        """
        Wrapper for OMCSession.sendExpression().
        """
        try:
            retval = self._session.sendExpression(expr=expr, parsed=parsed)
        except OMCSessionException as ex:
            raise ModelicaSystemError(f"Error executing {repr(expr)}: {ex}") from ex

        logger.debug(f"Result of executing {repr(expr)}: {textwrap.shorten(repr(retval), width=100)}")

        return retval

    # request to OMC
    def _requestApi(
            self,
            apiName: str,
            entity: Optional[str] = None,
            properties: Optional[str] = None,
    ) -> Any:
        if entity is not None and properties is not None:
            expr = f'{apiName}({entity}, {properties})'
        elif entity is not None and properties is None:
            if apiName in ("loadFile", "importFMU"):
                expr = f'{apiName}("{entity}")'
            else:
                expr = f'{apiName}({entity})'
        else:
            expr = f'{apiName}()'

        return self.sendExpression(expr=expr)

    def getContinuousFinal(
            self,
            names: Optional[str | list[str]] = None,
    ) -> dict[str, np.float64] | list[np.float64]:
        """
        Get (final) values of continuous signals (at stopTime).

        Args:
            names: Either None (default), a string with the continuous signal
              name, or a list of signal name strings.
        Returns:
            If `names` is None, a dict in the format
            {signal_name: signal_value} is returned.
            If `names` is a string, a single element list [signal_value] is
            returned.
            If `names` is a list, a list with one value for each signal name
            in names is returned: [signal1_value, signal2_value, ...].

        Examples:
            >>> mod.getContinuousFinal()
            {'x': np.float64(0.68), 'der(x)': np.float64(-0.24), 'y': np.float64(-0.24)}
            >>> mod.getContinuousFinal("x")
            [np.float64(0.68)]
            >>> mod.getContinuousFinal(["y","x"])
            [np.float64(-0.24), np.float64(0.68)]
        """
        if not self._simulated:
            raise ModelicaSystemError("Please use getContinuousInitial() before the simulation was started!")

        def get_continuous_solution(name_list: list[str]) -> None:
            for name in name_list:
                if name in self._continuous:
                    value = self.getSolutions(name)
                    self._continuous[name] = np.float64(value[0][-1])
                else:
                    raise KeyError(f"{names} is not continuous")

        if names is None:
            get_continuous_solution(name_list=list(self._continuous.keys()))
            return self._continuous

        if isinstance(names, str):
            get_continuous_solution(name_list=[names])
            return [self._continuous[names]]

        if isinstance(names, list):
            get_continuous_solution(name_list=names)
            values = []
            for name in names:
                values.append(self._continuous[name])
            return values

        raise ModelicaSystemError("Unhandled input for getContinousFinal()")

    def getContinuous(
            self,
            names: Optional[str | list[str]] = None,
    ) -> dict[str, np.float64] | list[np.float64]:
        """Get values of continuous signals.

        If called before simulate(), the initial values are returned.
        If called after simulate(), the final values (at stopTime) are returned.
        The return format is always numpy.float64.

        Args:
            names: Either None (default), a string with the continuous signal
              name, or a list of signal name strings.
        Returns:
            If `names` is None, a dict in the format
            {signal_name: signal_value} is returned.
            If `names` is a string, a single element list [signal_value] is
            returned.
            If `names` is a list, a list with one value for each signal name
            in names is returned: [signal1_value, signal2_value, ...].

        Examples:
            Before simulate():
            >>> mod.getContinuous()
            {'x': '1.0', 'der(x)': None, 'y': '-0.4'}
            >>> mod.getContinuous("y")
            ['-0.4']
            >>> mod.getContinuous(["y","x"])
            ['-0.4', '1.0']

            After simulate():
            >>> mod.getContinuous()
            {'x': np.float64(0.68), 'der(x)': np.float64(-0.24), 'y': np.float64(-0.24)}
            >>> mod.getContinuous("x")
            [np.float64(0.68)]
            >>> mod.getContinuous(["y","x"])
            [np.float64(-0.24), np.float64(0.68)]
        """
        if not self._simulated:
            return self.getContinuousInitial(names=names)

        return self.getContinuousFinal(names=names)

    def getOutputsFinal(
            self,
            names: Optional[str | list[str]] = None,
    ) -> dict[str, np.float64] | list[np.float64]:
        """Get (final) values of output signals (at stopTime).

        Args:
            names: Either None (default), a string with the output name,
              or a list of output name strings.
        Returns:
            If `names` is None, a dict in the format
            {output_name: output_value} is returned.
            If `names` is a string, a single element list [output_value] is
            returned.
            If `names` is a list, a list with one value for each output name
            in names is returned: [output1_value, output2_value, ...].

        Examples:
            >>> mod.getOutputsFinal()
            {'out1': np.float64(-0.1234), 'out2': np.float64(2.1)}
            >>> mod.getOutputsFinal("out1")
            [np.float64(-0.1234)]
            >>> mod.getOutputsFinal(["out1","out2"])
            [np.float64(-0.1234), np.float64(2.1)]
        """
        if not self._simulated:
            raise ModelicaSystemError("Please use getOuputsInitial() before the simulation was started!")

        def get_outputs_solution(name_list: list[str]) -> None:
            for name in name_list:
                if name in self._outputs:
                    value = self.getSolutions(name)
                    self._outputs[name] = np.float64(value[0][-1])
                else:
                    raise KeyError(f"{names} is not a valid output")

        if names is None:
            get_outputs_solution(name_list=list(self._outputs.keys()))
            return self._outputs

        if isinstance(names, str):
            get_outputs_solution(name_list=[names])
            return [self._outputs[names]]

        if isinstance(names, list):
            get_outputs_solution(name_list=names)
            values = []
            for name in names:
                values.append(self._outputs[name])
            return values

        raise ModelicaSystemError("Unhandled input for getOutputs()")

    def getOutputs(
            self,
            names: Optional[str | list[str]] = None,
    ) -> dict[str, np.float64] | list[np.float64]:
        """Get values of output signals.

        If called before simulate(), the initial values are returned.
        If called after simulate(), the final values (at stopTime) are returned.
        The return format is always numpy.float64.

        Args:
            names: Either None (default), a string with the output name,
              or a list of output name strings.
        Returns:
            If `names` is None, a dict in the format
            {output_name: output_value} is returned.
            If `names` is a string, a single element list [output_value] is
            returned.
            If `names` is a list, a list with one value for each output name
            in names is returned: [output1_value, output2_value, ...].

        Examples:
            Before simulate():
            >>> mod.getOutputs()
            {'out1': '-0.4', 'out2': '1.2'}
            >>> mod.getOutputs("out1")
            ['-0.4']
            >>> mod.getOutputs(["out1","out2"])
            ['-0.4', '1.2']

            After simulate():
            >>> mod.getOutputs()
            {'out1': np.float64(-0.1234), 'out2': np.float64(2.1)}
            >>> mod.getOutputs("out1")
            [np.float64(-0.1234)]
            >>> mod.getOutputs(["out1","out2"])
            [np.float64(-0.1234), np.float64(2.1)]
        """
        if not self._simulated:
            return self.getOutputsInitial(names=names)

        return self.getOutputsFinal(names=names)

    def plot(
            self,
            plotdata: str,
            resultfile: Optional[str | os.PathLike] = None,
    ) -> None:
        """
        Plot a variable using OMC; this will work for local OMC usage only (OMCProcessLocal). The reason is that the
        plot is created by OMC which needs access to the local display. This is not the case for docker and WSL.
        """

        if not isinstance(self._session, OMCSessionLocal):
            raise ModelicaSystemError("Plot is using the OMC plot functionality; "
                                      "thus, it is only working if OMC is running locally!")

        if resultfile is not None:
            plot_result_file = self._session.omcpath(resultfile)
        elif self._result_file is not None:
            plot_result_file = self._result_file
        else:
            raise ModelicaSystemError("No resultfile available - either run simulate() before plotting "
                                      "or provide a result file!")

        if not plot_result_file.is_file():
            raise ModelicaSystemError(f"Provided resultfile {repr(plot_result_file.as_posix())} does not exists!")

        expr = f'plot({plotdata}, fileName="{plot_result_file.as_posix()}")'
        self.sendExpression(expr=expr)

    def getSolutions(
            self,
            varList: Optional[str | list[str]] = None,
            resultfile: Optional[str | os.PathLike] = None,
    ) -> tuple[str] | np.ndarray:
        """Extract simulation results from a result data file.

        Args:
            varList: Names of variables to be extracted. Either unspecified to
              get names of available variables, or a single variable name
              as a string, or a list of variable names.
            resultfile: Path to the result file. If unspecified, the result
              file created by simulate() is used.

        Returns:
            If varList is None, a tuple with names of all variables
            is returned.
            If varList is a string, a 1D numpy array is returned.
            If varList is a list, a 2D numpy array is returned.

        Examples:
            >>> mod.getSolutions()
            ('a', 'der(x)', 'time', 'x')
            >>> mod.getSolutions("x")
            np.array([[1.        , 0.90483742, 0.81873075]])
            >>> mod.getSolutions(["x", "der(x)"])
            np.array([[1.        , 0.90483742 , 0.81873075],
                      [-1.       , -0.90483742, -0.81873075]])
            >>> mod.getSolutions(resultfile="c:/a.mat")
            ('a', 'der(x)', 'time', 'x')
            >>> mod.getSolutions("x", resultfile="c:/a.mat")
            np.array([[1.        , 0.90483742, 0.81873075]])
            >>> mod.getSolutions(["x", "der(x)"], resultfile="c:/a.mat")
            np.array([[1.        , 0.90483742 , 0.81873075],
                      [-1.       , -0.90483742, -0.81873075]])
        """
        if resultfile is None:
            if self._result_file is None:
                raise ModelicaSystemError("No result file found. Run simulate() first.")
            result_file = self._result_file
        else:
            result_file = self._session.omcpath(resultfile)

        # check if the result file exits
        if not result_file.is_file():
            raise ModelicaSystemError(f"Result file does not exist {result_file.as_posix()}")

        # get absolute path
        result_file = result_file.absolute()

        result_vars = self.sendExpression(expr=f'readSimulationResultVars("{result_file.as_posix()}")')
        self.sendExpression(expr="closeSimulationResultFile()")
        if varList is None:
            return result_vars

        if isinstance(varList, str):
            var_list_checked = [varList]
        elif isinstance(varList, list):
            var_list_checked = varList
        else:
            raise ModelicaSystemError("Unhandled input for getSolutions()")

        for var in var_list_checked:
            if var == "time":
                continue
            if var not in result_vars:
                raise ModelicaSystemError(f"Requested data {repr(var)} does not exist")
        variables = ",".join(var_list_checked)
        res = self.sendExpression(expr=f'readSimulationResult("{result_file.as_posix()}",{{{variables}}})')
        np_res = np.array(res)
        self.sendExpression(expr="closeSimulationResultFile()")
        return np_res

    def convertMo2Fmu(
            self,
            version: str = "2.0",
            fmuType: str = "me_cs",
            fileNamePrefix: Optional[str] = None,
            includeResources: bool = True,
    ) -> OMPathABC:
        """Translate the model into a Functional Mockup Unit.

        Args:
            See https://build.openmodelica.org/Documentation/OpenModelica.Scripting.translateModelFMU.html

        Returns:
            str: Path to the created '*.fmu' file.

        Examples:
            >>> mod.convertMo2Fmu()
            '/tmp/tmpmhfx9umo/CauerLowPassAnalog.fmu'
            >>> mod.convertMo2Fmu(version="2.0", fmuType="me|cs|me_cs", fileNamePrefix="<default>",
                                  includeResources=True)
            '/tmp/tmpmhfx9umo/CauerLowPassAnalog.fmu'
        """

        if fileNamePrefix is None:
            if self._model_name is None:
                fileNamePrefix = "<default>"
            else:
                fileNamePrefix = self._model_name
        include_resources_str = "true" if includeResources else "false"

        properties = (f'version="{version}", fmuType="{fmuType}", '
                      f'fileNamePrefix="{fileNamePrefix}", includeResources={include_resources_str}')
        fmu = self._requestApi(apiName='buildModelFMU', entity=self._model_name, properties=properties)
        fmu_path = self._session.omcpath(fmu)

        # report proper error message
        if not fmu_path.is_file():
            raise ModelicaSystemError(f"Missing FMU file: {fmu_path.as_posix()}")

        return fmu_path

    # to convert FMU to Modelica model
    def convertFmu2Mo(
            self,
            fmu: os.PathLike,
    ) -> OMPathABC:
        """
        In order to load FMU, at first it needs to be translated into Modelica model. This method is used to generate
        Modelica model from the given FMU. It generates "fmuName_me_FMU.mo".
        Currently, it only supports Model Exchange conversion.
        usage
        >>> convertFmu2Mo("c:/BouncingBall.Fmu")
        """

        fmu_path = self._session.omcpath(fmu)

        if not fmu_path.is_file():
            raise ModelicaSystemError(f"Missing FMU file: {fmu_path.as_posix()}")

        filename = self._requestApi(apiName='importFMU', entity=fmu_path.as_posix())
        filepath = self.getWorkDirectory() / filename

        # report proper error message
        if not filepath.is_file():
            raise ModelicaSystemError(f"Missing file {filepath.as_posix()}")

        self.model(
            model_name=f"{fmu_path.stem}_me_FMU",
            model_file=filepath,
        )

        return filepath

    def optimize(self) -> dict[str, Any]:
        """Perform model-based optimization.

        Optimization options set by setOptimizationOptions() are used.

        Returns:
            A dict with various values is returned. One of these values is the
            path to the result file.

        Examples:
            >>> mod.optimize()
            {'messages': 'LOG_SUCCESS | info | The initialization finished successfully without homotopy method. ...'
             'resultFile': '/tmp/tmp68guvjhs/BangBang2021_res.mat',
             'simulationOptions': 'startTime = 0.0, stopTime = 1.0, numberOfIntervals = '
                                  "1000, tolerance = 1e-8, method = 'optimization', "
                                  "fileNamePrefix = 'BangBang2021', options = '', "
                                  "outputFormat = 'mat', variableFilter = '.*', cflags = "
                                  "'', simflags = '-s=\\'optimization\\' "
                                  "-optimizerNP=\\'1\\''",
             'timeBackend': 0.008684897,
             'timeCompile': 0.7546678929999999,
             'timeFrontend': 0.045438053000000006,
             'timeSimCode': 0.0018537170000000002,
             'timeSimulation': 0.266354356,
             'timeTemplates': 0.002007785,
             'timeTotal': 1.079097854}
        """
        properties = ','.join(f"{key}={val}" for key, val in self._optimization_options.items())
        self.set_command_line_options("-g=Optimica")
        return self._requestApi(apiName='optimize', entity=self._model_name, properties=properties)


class ModelicaSystem(ModelicaSystemOMC):
    """
    Compatibility class.
    """


class ModelicaDoEABC(metaclass=abc.ABCMeta):
    """
    Base class to run DoEs based on a (Open)Modelica model using ModelicaSystem

    Example
    -------
    ```
    import OMPython
    import pathlib


    def run_doe():
        mypath = pathlib.Path('.')

        model = mypath / "M.mo"
        model.write_text(
            "    model M\n"
            "      parameter Integer p=1;\n"
            "      parameter Integer q=1;\n"
            "      parameter Real a = -1;\n"
            "      parameter Real b = -1;\n"
            "      Real x[p];\n"
            "      Real y[q];\n"
            "    equation\n"
            "      der(x) = a * fill(1.0, p);\n"
            "      der(y) = b * fill(1.0, q);\n"
            "    end M;\n"
        )

        param = {
            # structural
            'p': [1, 2],
            'q': [3, 4],
            # non-structural
            'a': [5, 6],
            'b': [7, 8],
        }

        resdir = mypath / 'DoE'
        resdir.mkdir(exist_ok=True)

        mod = OMPython.ModelicaSystemOMC()
        mod.model(
            model_name="M",
            model_file=model.as_posix(),
        )
        doe_mod = OMPython.ModelicaSystemDoE(
            mod=mod,
            parameters=param,
            resultpath=resdir,
            simargs={"override": {'stopTime': 1.0}},
        )
        doe_mod.prepare()
        doe_def = doe_mod.get_doe_definition()
        doe_mod.simulate()
        doe_sol = doe_mod.get_doe_solutions()

        # ... work with doe_def and doe_sol ...


    if __name__ == "__main__":
        run_doe()
    ```

    """

    # Dictionary keys used in simulation dict (see _sim_dict or get_doe()). These dict keys contain a space and, thus,
    # cannot be used as OM variable identifiers. They are defined here as reference for any evaluation of the data.
    DICT_ID_STRUCTURE: str = 'ID structure'
    DICT_ID_NON_STRUCTURE: str = 'ID non-structure'
    DICT_RESULT_AVAILABLE: str = 'result available'

    def __init__(
            self,
            # ModelicaSystem definition to use
            mod: ModelicaSystemABC,
            # simulation specific input
            # TODO: add more settings (simulation options, input options, ...)
            simargs: Optional[dict[str, Optional[str | dict[str, str] | numbers.Number]]] = None,
            # DoE specific inputs
            resultpath: Optional[str | os.PathLike] = None,
            parameters: Optional[dict[str, list[str] | list[int] | list[float]]] = None,
    ) -> None:
        """
        Initialisation of ModelicaSystemDoE. The parameters are based on: ModelicaSystem.__init__() and
        ModelicaSystem.simulate(). Additionally, the path to store the result files is needed (= resultpath) as well as
        a list of parameters to vary for the Doe (= parameters). All possible combinations are considered.
        """
        if not isinstance(mod, ModelicaSystemABC):
            raise ModelicaSystemError("Missing definition of ModelicaSystem!")

        self._mod = mod
        self._model_name = mod.get_model_name()

        self._simargs = simargs

        if resultpath is None:
            self._resultpath = self.get_session().omcpath_tempdir()
        else:
            self._resultpath = self.get_session().omcpath(resultpath).resolve()
        if not self._resultpath.is_dir():
            raise ModelicaSystemError("Argument resultpath must be set to a valid path within the environment used "
                                      f"for the OpenModelica session: {resultpath}!")

        if isinstance(parameters, dict):
            self._parameters = parameters
        else:
            self._parameters = {}

        self._doe_def: Optional[dict[str, dict[str, Any]]] = None
        self._doe_cmd: Optional[dict[str, ModelExecutionData]] = None

    def get_session(self) -> OMSessionABC:
        """
        Return the OMC session used for this class.
        """
        return self._mod.get_session()

    def get_resultpath(self) -> OMPathABC:
        """
        Get the path there the result data is saved.
        """
        return self._resultpath

    def prepare(self) -> int:
        """
        Prepare the DoE by evaluating the parameters. Each structural parameter requires a new instance of
        ModelicaSystem while the non-structural parameters can just be set on the executable.

        The return value is the number of simulation defined.
        """

        doe_sim = {}
        doe_def = {}

        param_structure = {}
        param_non_structure = {}
        for param_name in self._parameters.keys():
            changeable = self._mod.isParameterChangeable(name=param_name)
            logger.info(f"Parameter {repr(param_name)} is changeable? {changeable}")

            if changeable:
                param_non_structure[param_name] = self._parameters[param_name]
            else:
                param_structure[param_name] = self._parameters[param_name]

        param_structure_combinations = list(itertools.product(*param_structure.values()))
        param_non_structural_combinations = list(itertools.product(*param_non_structure.values()))

        for idx_pc_structure, pc_structure in enumerate(param_structure_combinations):
            sim_param_structure = self._prepare_structure_parameters(
                idx_pc_structure=idx_pc_structure,
                pc_structure=pc_structure,
                param_structure=param_structure,
            )

            for idx_non_structural, pk_non_structural in enumerate(param_non_structural_combinations):
                sim_param_non_structural = {}
                for idx, pk in enumerate(param_non_structure.keys()):
                    sim_param_non_structural[pk] = cast(Any, pk_non_structural[idx])

                resfilename = f"DOE_{idx_pc_structure:09d}_{idx_non_structural:09d}.mat"
                logger.info(f"use result file {repr(resfilename)} "
                            f"for structural parameters: {sim_param_structure} "
                            f"and non-structural parameters: {sim_param_non_structural}")
                resultfile = self._resultpath / resfilename

                df_data = (
                        {
                            self.DICT_ID_STRUCTURE: idx_pc_structure,
                        }
                        | sim_param_structure
                        | {
                            self.DICT_ID_NON_STRUCTURE: idx_non_structural,
                        }
                        | sim_param_non_structural
                        | {
                            self.DICT_RESULT_AVAILABLE: False,
                        }
                )

                self._mod.setParameters(sim_param_non_structural)
                mscmd = self._mod.simulate_cmd(
                    result_file=resultfile,
                )
                if self._simargs is not None:
                    mscmd.args_set(args=self._simargs)
                cmd_definition = mscmd.definition()
                del mscmd

                doe_sim[resfilename] = cmd_definition
                doe_def[resfilename] = df_data

        logger.info(f"Prepared {len(doe_sim)} simulation definitions for the defined DoE.")
        self._doe_cmd = doe_sim
        self._doe_def = doe_def

        return len(doe_sim)

    @abc.abstractmethod
    def _prepare_structure_parameters(
            self,
            idx_pc_structure: int,
            pc_structure: Tuple,
            param_structure: dict[str, list[str] | list[int] | list[float]],
    ) -> dict[str, str | int | float]:
        """
        Handle structural parameters. This should be implemented by the derived class
        """

    def get_doe_definition(self) -> Optional[dict[str, dict[str, Any]]]:
        """
        Get the defined DoE as a dict, where each key is the result filename and the value is a dict of simulation
        settings including structural and non-structural parameters.

        The following code snippet can be used to convert the data to a pandas dataframe:

        ```
        import pandas as pd

        doe_dict = doe_mod.get_doe_definition()
        doe_df = pd.DataFrame.from_dict(data=doe_dict, orient='index')
        ```

        """
        return self._doe_def

    def get_doe_command(self) -> Optional[dict[str, ModelExecutionData]]:
        """
        Get the definitions of simulations commands to run for this DoE.
        """
        return self._doe_cmd

    def simulate(
            self,
            num_workers: int = 3,
    ) -> bool:
        """
        Simulate the DoE using the defined number of workers.

        Returns True if all simulations were done successfully, else False.
        """

        if self._doe_cmd is None or self._doe_def is None:
            raise ModelicaSystemError("DoE preparation missing - call prepare() first!")

        doe_cmd_total = len(self._doe_cmd)
        doe_def_total = len(self._doe_def)

        if doe_cmd_total != doe_def_total:
            raise ModelicaSystemError(f"Mismatch between number simulation commands ({doe_cmd_total}) "
                                      f"and simulation definitions ({doe_def_total}).")

        doe_task_query: queue.Queue = queue.Queue()
        if self._doe_cmd is not None:
            for doe_cmd in self._doe_cmd.values():
                doe_task_query.put(doe_cmd)

        if not isinstance(self._doe_def, dict) or len(self._doe_def) == 0:
            raise ModelicaSystemError("Missing Doe Summary!")

        def worker(worker_id, task_queue):
            while True:
                try:
                    # Get the next task from the queue
                    cmd_definition = task_queue.get(block=False)
                except queue.Empty:
                    logger.info(f"[Worker {worker_id}] No more simulations to run.")
                    break

                if cmd_definition is None:
                    raise ModelicaSystemError("Missing simulation definition!")

                resultfile = cmd_definition.cmd_result_file
                resultpath = self.get_session().omcpath(resultfile)

                logger.info(f"[Worker {worker_id}] Performing task: {resultpath.name}")

                try:
                    returncode = cmd_definition.run()
                    logger.info(f"[Worker {worker_id}] Simulation {resultpath.name} "
                                f"finished with return code: {returncode}")
                except ModelicaSystemError as ex:
                    logger.warning(f"Simulation error for {resultpath.name}: {ex}")

                # Mark the task as done
                task_queue.task_done()

                sim_query_done = doe_cmd_total - doe_task_query.qsize()
                logger.info(f"[Worker {worker_id}] Task completed: {resultpath.name} "
                            f"({doe_cmd_total - sim_query_done}/{doe_cmd_total} = "
                            f"{(doe_cmd_total - sim_query_done) / doe_cmd_total * 100:.2f}% of tasks left)")

        # Create and start worker threads
        logger.info(f"Start simulations for DoE with {doe_cmd_total} simulations "
                    f"using {num_workers} workers ...")
        threads = []
        for i in range(num_workers):
            thread = threading.Thread(target=worker, args=(i, doe_task_query))
            thread.start()
            threads.append(thread)

        # Wait for all threads to complete
        for thread in threads:
            thread.join()

        doe_def_done = 0
        for resultfilename in self._doe_def:
            resultfile = self._resultpath / resultfilename

            # include check for an empty (=> 0B) result file which indicates a crash of the model executable
            # see: https://github.com/OpenModelica/OMPython/issues/261
            # https://github.com/OpenModelica/OpenModelica/issues/13829
            if resultfile.is_file() and resultfile.size() > 0:
                self._doe_def[resultfilename][self.DICT_RESULT_AVAILABLE] = True
                doe_def_done += 1

        logger.info(f"All workers finished ({doe_def_done} of {doe_def_total} simulations with a result file).")

        return doe_def_total == doe_def_done


class ModelicaDoEOMC(ModelicaDoEABC):
    """
    Class to run DoEs based on a (Open)Modelica model using ModelicaSystemOMC

    The example is the same as defined for ModelicaDoEABC
    """

    def __init__(
            self,
            # ModelicaSystem definition to use
            mod: ModelicaSystemOMC,
            # simulation specific input
            # TODO: add more settings (simulation options, input options, ...)
            simargs: Optional[dict[str, Optional[str | dict[str, str] | numbers.Number]]] = None,
            # DoE specific inputs
            resultpath: Optional[str | os.PathLike] = None,
            parameters: Optional[dict[str, list[str] | list[int] | list[float]]] = None,
    ) -> None:

        if not isinstance(mod, ModelicaSystemOMC):
            raise ModelicaSystemError(f"Invalid definition for mod: {type(mod)} - expect ModelicaSystemOMC!")

        super().__init__(
            mod=mod,
            simargs=simargs,
            resultpath=resultpath,
            parameters=parameters,
        )

    def _prepare_structure_parameters(
            self,
            idx_pc_structure: int,
            pc_structure: Tuple,
            param_structure: dict[str, list[str] | list[int] | list[float]],
    ) -> dict[str, str | int | float]:
        build_dir = self._resultpath / f"DOE_{idx_pc_structure:09d}"
        build_dir.mkdir()
        self._mod.setWorkDirectory(work_directory=build_dir)

        # need to repeat this check to make the linters happy
        if not isinstance(self._mod, ModelicaSystemOMC):
            raise ModelicaSystemError(f"Invalid definition for mod: {type(self._mod)} - expect ModelicaSystemOMC!")

        sim_param_structure = {}
        for idx_structure, pk_structure in enumerate(param_structure.keys()):
            sim_param_structure[pk_structure] = pc_structure[idx_structure]

            pk_value = pc_structure[idx_structure]
            if isinstance(pk_value, str):
                pk_value_str = self.get_session().escape_str(pk_value)
                expr = f"setParameterValue({self._model_name}, {pk_structure}, \"{pk_value_str}\")"
            elif isinstance(pk_value, bool):
                pk_value_bool_str = "true" if pk_value else "false"
                expr = f"setParameterValue({self._model_name}, {pk_structure}, {pk_value_bool_str});"
            else:
                expr = f"setParameterValue({self._model_name}, {pk_structure}, {pk_value})"
            res = self._mod.sendExpression(expr=expr)
            if not res:
                raise ModelicaSystemError(f"Cannot set structural parameter {self._model_name}.{pk_structure} "
                                          f"to {pk_value} using {repr(expr)}")

        self._mod.buildModel()

        return sim_param_structure

    def get_doe_solutions(
            self,
            var_list: Optional[list] = None,
    ) -> Optional[tuple[str] | dict[str, dict[str, np.ndarray]]]:
        """
        Wrapper for doe_get_solutions()
        """
        if not isinstance(self._mod, ModelicaSystemOMC):
            raise ModelicaSystemError(f"Invalid definition for mod: {type(self._mod)} - expect ModelicaSystemOMC!")

        return doe_get_solutions(
            msomc=self._mod,
            resultpath=self._resultpath,
            doe_def=self.get_doe_definition(),
            var_list=var_list,
        )


def doe_get_solutions(
        msomc: ModelicaSystemOMC,
        resultpath: OMPathABC,
        doe_def: Optional[dict] = None,
        var_list: Optional[list] = None,
) -> Optional[tuple[str] | dict[str, dict[str, np.ndarray]]]:
    """
    Get all solutions of the DoE run. The following return values are possible:

    * A list of variables if val_list == None

    * The Solutions as dict[str, pd.DataFrame] if a value list (== val_list) is defined.

    The following code snippet can be used to convert the solution data for each run to a pandas dataframe:

    ```
    import pandas as pd

    doe_sol = doe_mod.get_doe_solutions()
    for key in doe_sol:
        data = doe_sol[key]['data']
        if data:
            doe_sol[key]['df'] = pd.DataFrame.from_dict(data=data)
        else:
            doe_sol[key]['df'] = None
    ```

    """
    if not isinstance(doe_def, dict):
        return None

    if len(doe_def) == 0:
        raise ModelicaSystemError("No result files available - all simulations did fail?")

    sol_dict: dict[str, dict[str, Any]] = {}
    for resultfilename in doe_def:
        resultfile = resultpath / resultfilename

        sol_dict[resultfilename] = {}

        if not doe_def[resultfilename][ModelicaDoEABC.DICT_RESULT_AVAILABLE]:
            msg = f"No result file available for {resultfilename}"
            logger.warning(msg)
            sol_dict[resultfilename]['msg'] = msg
            sol_dict[resultfilename]['data'] = {}
            continue

        if var_list is None:
            var_list_row = list(msomc.getSolutions(resultfile=resultfile))
        else:
            var_list_row = var_list

        try:
            sol = msomc.getSolutions(varList=var_list_row, resultfile=resultfile)
            sol_data = {var: sol[idx] for idx, var in enumerate(var_list_row)}
            sol_dict[resultfilename]['msg'] = 'Simulation available'
            sol_dict[resultfilename]['data'] = sol_data
        except ModelicaSystemError as ex:
            msg = f"Error reading solution for {resultfilename}: {ex}"
            logger.warning(msg)
            sol_dict[resultfilename]['msg'] = msg
            sol_dict[resultfilename]['data'] = {}

    return sol_dict


class ModelicaSystemDoE(ModelicaDoEOMC):
    """
    Compatibility class.
    """


class ModelicaSystemRunner(ModelicaSystemABC):
    """
    Class to simulate a Modelica model using a pre-compiled model binary.
    """

    def __init__(
            self,
            work_directory: Optional[str | os.PathLike] = None,
            session: Optional[OMSessionABC] = None,
    ) -> None:
        if session is None:
            session = OMSessionRunner()

        if not isinstance(session, OMSessionRunner):
            raise ModelicaSystemError("Only working if OMCsessionDummy is used!")

        super().__init__(
            work_directory=work_directory,
            session=session,
        )

    def setup(
            self,
            model_name: Optional[str] = None,
            variable_filter: Optional[str] = None,
    ) -> None:
        """
        Needed definitions to set up the runner class. This class expects the model (defined by model_name) to exists
        within the working directory. At least two files are needed:

        * model executable (as '<model_name>' or '<model_name>.exe'; in case of Windows additional '<model_name>.bat'
          is expected to evaluate the path to needed dlls
        * the model initialization file (as '<model_name>_init.xml')
        """

        if self._model_name is not None:
            raise ModelicaSystemError("Can not reuse this instance of ModelicaSystem "
                                      f"defined for {repr(self._model_name)}!")

        if model_name is None or not isinstance(model_name, str):
            raise ModelicaSystemError("A model name must be provided!")

        # set variables
        self._model_name = model_name  # Model class name
        self._variable_filter = variable_filter

        # test if the model can be executed
        self.check_model_executable()

        # read XML file
        xml_file = self._session.omcpath(self.getWorkDirectory()) / f"{self._model_name}_init.xml"
        self._xmlparse(xml_file=xml_file)


class ModelicaDoERunner(ModelicaDoEABC):
    """
    Class to run DoEs based on a (Open)Modelica model using ModelicaSystemRunner

    The example is the same as defined for ModelicaDoEABC
    """

    def __init__(
            self,
            # ModelicaSystem definition to use
            mod: ModelicaSystemABC,
            # simulation specific input
            # TODO: add more settings (simulation options, input options, ...)
            simargs: Optional[dict[str, Optional[str | dict[str, str] | numbers.Number]]] = None,
            # DoE specific inputs
            resultpath: Optional[str | os.PathLike] = None,
            parameters: Optional[dict[str, list[str] | list[int] | list[float]]] = None,
    ) -> None:
        if not isinstance(mod, ModelicaSystemABC):
            raise ModelicaSystemError(f"Invalid definition for ModelicaSystem*: {type(mod)}!")

        super().__init__(
            mod=mod,
            simargs=simargs,
            resultpath=resultpath,
            parameters=parameters,
        )

    def _prepare_structure_parameters(
            self,
            idx_pc_structure: int,
            pc_structure: Tuple,
            param_structure: dict[str, list[str] | list[int] | list[float]],
    ) -> dict[str, str | int | float]:
        if len(param_structure.keys()) > 0:
            raise ModelicaSystemError(f"{self.__class__.__name__} can not handle structure parameters as it uses a "
                                      "pre-compiled binary of model.")

        return {}