Main Content

compiler.build.pythonPackage

Create Python package for deployment outside MATLAB

Since R2021a

Description

compiler.build.pythonPackage(FunctionFiles) creates a Python® package using the MATLAB® functions specified by FunctionFiles.

example

compiler.build.pythonPackage(FunctionFiles,Name,Value) creates a Python package with additional options specified using one or more name-value arguments. Options include the package name, output directory, and additional files to include.

example

compiler.build.pythonPackage(opts) creates a Python package with options specified using a compiler.build.PythonPackageOptions object opts. You cannot specify any other options using name-value arguments.

example

results = compiler.build.pythonPackage(___) returns build information as a compiler.build.Results object using any of the input argument combinations in previous syntaxes. The build information consists of the build type, paths to the compiled files, and build options.

example

Examples

collapse all

Create a Python package using a function file that generates a magic square.

In MATLAB, locate the MATLAB function that you want to deploy as a Python package. For this example, use the file magicsquare.m located in matlabroot\extern\examples\compiler.

appFile = which('magicsquare.m');

Build a Python package using the compiler.build.pythonPackage command.

compiler.build.pythonPackage(appFile);

The build function creates the following files within a folder named magicsquarepythonPackage in your current working directory:

  • GettingStarted.html

  • includedSupportPackages.txt

  • example

  • mccExcludedFiles.log

  • readme.txt

  • requiredMCRProducts.txt

  • pyproject.toml

  • setup.py

  • unresolvedSymbols.txt

Create a Python package and customize it using name-value arguments.

For this example, use the files flames.m and flames.mat located in matlabroot\extern\examples\compiler.

appFile = which('flames.m');
MATFile = which('flames.mat');

Build a Python package using the compiler.build.pythonPackage command. Use name-value arguments to specify the package name, add a MAT-file, and enable verbose output.

compiler.build.pythonPackage(appFile, ...
'PackageName','FlamesApp','AdditionalFiles',MATFile, ...
'Verbose','on');

Create multiple Python packages using a compiler.build.PythonPackageOptions object.

For this example, use the file magicsquare.m located in matlabroot\extern\examples\compiler.

appFile = which('magicsquare.m');

Create a PythonPackageOptions object using appFile. Use name-value arguments to specify a common output directory, disable automatic detection of data files, and enable verbose output.

opts = compiler.build.PythonPackageOptions(appFile, ...
'OutputDir','D:\Documents\MATLAB\work\PythonPackageBatch',...
'AutoDetectDataFiles','off','Verbose','on')
opts =

  PythonPackageOptions with properties:

            FunctionFiles: {'C:\Program Files\MATLAB\R2024b\extern\examples\compiler\magicsquare.m'}
              PackageName: 'example.magicsquare'
    SampleGenerationFiles: {}
          AdditionalFiles: {}
      AutoDetectDataFiles: off
    ExternalEncryptionKey: [0×0 struct]
         ObfuscateArchive: off
          SecretsManifest: ''
          SupportPackages: {'autodetect'}
                  Verbose: on
                OutputDir: 'D:\Documents\MATLAB\work\PythonPackageBatch'

Build the Python package using the PythonPackageOptions object.

compiler.build.pythonPackage(opts);

To compile using the function file hello.m with the same options, use dot notation to modify the FunctionFiles argument of the existing PythonPackageOptions object before running the build function again.

opts.FunctionFiles = which('hello.m');
compiler.build.pythonPackage(opts);

By modifying the FunctionFiles argument and recompiling, you can compile multiple components using the same options object.

Create a Python package and save information about the build type, generated files, included support packages, and build options to a compiler.build.Results object.

Compile using the file magicsquare.m.

results = compiler.build.pythonPackage('magicsquare.m');
results = 

  Results with properties:

            BuildType: 'pythonPackage'
                Files: {3×1 cell}
IncludedSupportPackages: {}
              Options: [1×1 compiler.build.PythonPackageOptions]

The Files property contains the paths to the following:

  • example folder

  • setup.py

  • GettingStarted.html

Input Arguments

collapse all

Files implementing MATLAB functions, specified as a character vector, a string scalar, a string array, or a cell array of character vectors. File paths can be relative to the current working directory or absolute. Files must have one of the following extensions: .m, .p, .mlx, or .mexa64.

Example: ["myfunc1.m","myfunc2.m"]

Data Types: char | string | cell

Python package build options, specified as a compiler.build.PythonPackageOptions object.

Name-Value Arguments

Specify optional pairs of arguments as Name1=Value1,...,NameN=ValueN, where Name is the argument name and Value is the corresponding value. Name-value arguments must appear after other arguments, but the order of the pairs does not matter.

Before R2021a, use commas to separate each name and value, and enclose Name in quotes.

Example: OutputDir='D:\work\myproject'

Additional files and folders to include in the Python package, specified as a character vector, a string scalar, a string array, or a cell array of character vectors. Paths can be relative to the current working directory or absolute.

Example: 'AdditionalFiles',["myvars.mat","data.txt"]

Data Types: char | string | cell

Flag to automatically include data files, specified as 'on' or 'off', or as numeric or logical 1 (true) or 0 (false). A value of 'on' is equivalent to true, and 'off' is equivalent to false. Thus, you can use the value of this property as a logical value. The value is stored as an on/off logical value of type matlab.lang.OnOffSwitchState.

  • If you set this property to 'on', then data files that you provide as inputs to certain functions (such as load and fopen) are automatically included in the package.

  • If you set this property to 'off', then you must add data files to the package using the AdditionalFiles option.

Example: 'AutoDetectDataFiles','off'

Data Types: logical

Since R2024b

Paths to the external AES encryption key and MEX key loader files, specified as a scalar struct with exactly two row char vector or string scalar fields named EncryptionKeyFile and RuntimeKeyLoaderFile, respectively. Both struct fields are required. File paths can be relative to the current working directory or absolute.

For example, specify the encryption key as encrypt.key and loader file as loader.mexw64 using struct keyValueStruct.

keyValueStruct.EncryptionKeyFile='encrypt.key'; keyValueStruct.RuntimeKeyLoaderFile='loader.mexw64'

The encryption key file must be in one of the following supported formats:

  • Binary 256-bit AES key, with a 32 byte file size

  • Hex encoded AES key, with a 64 byte file size

The MEX file loader retrieves the decryption key at runtime and must be an interface with the following arguments:

  • prhs[0] — Input, char array specified as the static value 'get'

  • prhs[1] — Input, char array specified as the CTF component UUID

  • plhs[0] — Output, 32 byte UINT8 numeric array or 64 byte hex encoded char array, depending on the key format

Avoid sharing the same key across multiple CTFs.

Example: 'ExternalEncryptionKey',keyValueStruct

Data Types: struct

Name of the Python package, specified as a character vector or a string scalar. Specify 'PackageName' as a namespace, which is a period-separated list, such as companyname.groupname.component. The name of the generated package is set to the last entry of the period-separated list. The name must begin with a letter and contain only alphabetic characters and periods.

If not specified, PackageName defaults to the name of the first MATLAB file listed in the FunctionFiles argument.

Example: 'PackageName','mathworks.pythonpackage.mymagic'

Data Types: char | string

MATLAB sample files used to generate sample Python files for functions included with the package, specified as a character vector, a string scalar, a string array, or a cell array of character vectors. Paths can be relative to the current working directory or absolute. Files must have a .m extension.

Example: 'SampleGenerationFiles',["sample1.m","sample2.m"]

Data Types: char | string | cell

Since R2024b

Path to a secret manifest JSON file that specifies the secret keys to be embedded in the deployable archive, specified as a character vector or a string scalar. The path can be relative to the current working directory or absolute.

If your MATLAB code calls the getSecret, getSecretMetadata, or isSecret function, you must specify the secret keys to embed in the deployable archive in a JSON secret manifest file. If your code calls getSecret and you do not specify the SecretsManifest option, MATLAB Compiler™ issues a warning and generates a template JSON file in the output folder named <component_name>_secrets_manifest.json. Modify this file by specifying the secret key names in the Embedded field.

The setSecret function is not deployable. To embed secret keys in a deployable archive, you must call setSecret in MATLAB before you build the archive.

For more information on deployment using secrets, see Handle Sensitive Information in Deployed Applications.

Example: 'SecretsManifest','D:\Documents\MATLAB\work\mycomponent\mycomponent_secrets_manifest.json'

Data Types: char | string

Flag to obfuscate the deployable archive, specified as 'on' or 'off', or as numeric or logical 1 (true) or 0 (false). A value of 'on' is equivalent to true, and 'off' is equivalent to false. Thus, you can use the value of this property as a logical value. The value is stored as an on/off logical value of type matlab.lang.OnOffSwitchState.

  • If you set this property to 'on', then folder structures and file names in the deployable archive are obfuscated from the end user, and user code and data contained in MATLAB files are placed into a user package within the archive. Additionally, all .m files are converted to P-files before packaging. This option is equivalent to using mcc with -j and -s specified.

  • If you set this property to 'off', then the deployable archive is not obfuscated. This is the default behavior.

Example: 'ObfuscateArchive','on'

Data Types: logical

Path to the output directory where the build files are saved, specified as a character vector or a string scalar. The path can be relative to the current working directory or absolute.

The default name of the build folder is the package name appended with pythonPackage.

Example: 'OutputDir','D:\Documents\MATLAB\work\mymagicpythonPackage'

Data Types: char | string

Support packages to include, specified as one of the following options:

  • 'autodetect' (default) — The dependency analysis process detects and includes the required support packages automatically.

  • 'none' — No support packages are included. Using this option can cause runtime errors.

  • A string scalar, character vector, or cell array of character vectors — Only the specified support packages are included. To list installed support packages or those used by a specific file, see compiler.codetools.deployableSupportPackages.

Example: 'SupportPackages',{'Deep Learning Toolbox Converter for TensorFlow Models','Deep Learning Toolbox Model for Places365-GoogLeNet Network'}

Data Types: char | string | cell

Flag to control build verbosity, specified as 'on' or 'off', or as numeric or logical 1 (true) or 0 (false). A value of 'on' is equivalent to true, and 'off' is equivalent to false. Thus, you can use the value of this property as a logical value. The value is stored as an on/off logical value of type matlab.lang.OnOffSwitchState.

  • If you set this property to 'on', then the MATLAB command window displays progress information indicating compiler output during the build process.

  • If you set this property to 'off', then the command window does not display progress information.

Example: 'Verbose','on'

Data Types: logical

Output Arguments

collapse all

Build results, returned as a compiler.build.Results object. The Results object contains:

  • The build type, which is 'pythonPackage'

  • Paths to the following items:

    • example folder

    • setup.py

    • GettingStarted.html

  • A list of included support packages

  • Build options, specified as a PythonPackageOptions object

Version History

Introduced in R2021a