MlFmuBuilder

class mlfmu.api.MlFmuBuilder(fmu_name: str | None = None, interface_file: Path | None = None, ml_model_file: Path | None = None, source_folder: Path | None = None, fmu_output_folder: Path | None = None, build_folder: Path | None = None, root_directory: Path | None = None)

Bases: object

Builder for creating FMUs (Functional Mock-up Units) from machine learning models.

This class is core to executing the different commands in the mlfmu process.

__init__(fmu_name: str | None = None, interface_file: Path | None = None, ml_model_file: Path | None = None, source_folder: Path | None = None, fmu_output_folder: Path | None = None, build_folder: Path | None = None, root_directory: Path | None = None) → None

Initialize a new instance of the MlFmuBuilder class.

Parameters:

• fmu_name (str | None, optional) – The name of the FMU., by default None

• interface_file (Path | None, optional) – The path to the interface JSON file., by default None

• ml_model_file (Path | None, optional) – The path to the machine learning model file., by default None

• source_folder (Path | None, optional) – The folder containing the source code for the FMU., by default None

• fmu_output_folder (Path | None, optional) – The folder where the built FMU will be saved., by default None

• build_folder (Path | None, optional) – The folder where the FMU will be built., by default None

• root_directory (Path | None, optional) – The root directory for the builder., by default None

Methods

__init__([fmu_name, interface_file, ...])	Initialize a new instance of the MlFmuBuilder class.
build()	Build an FMU from the machine learning model file and interface file.
compile()	Compile the FMU from the C++ source code and model description.
default_build_folder()	Return the path to a build folder inside the temp_folder.
default_build_source_folder()	Return the path to a src folder inside the temp_folder.
default_compile_source_folder()	Return the path to the default source folder for the compile process.
default_fmu_output_folder()	Return the path to the default fmu output folder.
default_generate_source_folder()	Return the path to the default source folder for the generate process.
default_interface_file()	Return the path to a interface json file inside self.root_directory if it can be inferred.
default_model_file()	Return the path to a ml model file inside self.root_directory if it can be inferred.
generate()	Generate C++ source code and model description from the machine learning model file and interface file.

Attributes

temp_folder

The temporary folder used for building the FMU.

build() → None

Build an FMU from the machine learning model file and interface file.

Builds an FMU from ml_model_file and interface_file and saves it to fmu_output_folder. If the paths to the necessary files and directories are not given the function will try to find files and directories that match the ones needed.

Raises:

• FileNotFoundError – if ml_model_file or interface_file do not exists or is not set and cannot be easily inferred.

• --- –

compile() → None

Compile the FMU from the C++ source code and model description.

Compiles the FMU from the FMU C++ source code and model description contained in source_folder and saves it to fmu_output_folder.

If the paths to the necessary files and directories are not given the function will try to find files and directories that match the ones needed.

Raises:

FileNotFoundError – if source_folder or fmu_name is not set and cannot be easily inferred.

default_build_folder() → Path

Return the path to a build folder inside the temp_folder. Creates the temp_folder if it is not set.

Returns:

the path to the build folder.

Return type:

Path

default_build_source_folder() → Path

Return the path to a src folder inside the temp_folder. Creates the temp_folder if it is not set.

Returns:

the path to the src folder.

Return type:

Path

default_compile_source_folder() → Path | None

Return the path to the default source folder for the compile process.

Searches inside self.source_folder and self.root_directory for a folder that contains a folder structure and files that is required to be valid ml fmu source code.

Returns:

the path to the default source folder for the compile process.

Return type:

Path

default_fmu_output_folder() → Path

Return the path to the default fmu output folder.

Returns:

the path to the default fmu output folder.

Return type:

Path

default_generate_source_folder() → Path

Return the path to the default source folder for the generate process.

Returns:

the path to the default source folder for the generate process.

Return type:

Path

default_interface_file() → Path | None

Return the path to a interface json file inside self.root_directory if it can be inferred.

Returns:

the path to the interface json file, or None.

Return type:

Path | None

default_model_file() → Path | None

Return the path to a ml model file inside self.root_directory if it can be inferred.

Returns:

the path to the ml model file, or None.

Return type:

Path | None

generate() → None

Generate C++ source code and model description from the machine learning model file and interface file.

Generates FMU C++ source code and model description from ml_model_file and interface_file and saves it to source_folder.

If the paths to the necessary files and directories are not given the function will try to find files and directories that match the ones needed.

Raises:

FileNotFoundError – if ml_model_file or interface_file do not exists or is not set and cannot be easily inferred.

temp_folder: TemporaryDirectory[str]

The temporary folder used for building the FMU.