Setup

Requirements

To call Python^® from MATLAB, you need to have both installed on your system.

Configuration

Access settings and the status of the Python interpreter:

>> pe = pyenv 

Specify which version to use:

>> pe = pyenv("Version","3.9") 

Ensure the PYTHONHOME environment variable aligns with your Python version. To view its value in MATLAB:

>> getenv("PYTHONHOME") 

Ensure the appropriate versions are on the system path:

>> getenv("PATH")

Run Python Code in MATLAB

Live Editor Task

The Run Python Code task lets you interactively run Python code or files, then generate the code.

Call Python Statements

To execute a few lines of code:

>> outvars = pyrun(code,outputs) 
>> x = pyrun("a = b*c","a", b = 5, c = 10)

Call Python Script

To execute code organized in a script:

>> outvars = pyrunfile(file,outputs) 

Call Python Modules and Functions

>> py.module _ name.function _ name 
>> x = py.math.sqrt(42) 

Pass Keyword Arguments

Either call directly or use pyargs:

>>> foo(5,bar=42) 
>> py.foo(5,bar=42) 
>> py.foo(5,pyargs('bar',42)) 

Reload Modules

Reload the module after making updates:

>> py.importlib.reload(module)

Data Type Conversions

Data types will be automatically converted where possible.

Data Science Libraries

Apache Parquet

Use Apache^® Parquet to efficiently transfer data.

From MATLAB:

>> tbl = parquetread(fname)
>> parquetwrite(tbl,fname) 

From Python:

>>> df = pandas.read _ parquet(fname) 
>>> pandas.Dataframe.to _ parquet(df) 

Deep Learning

Access models in MATLAB with importers for TensorFlow^™, PyTorch^®, and ONNX^™.

>> net = importKerasNetwork(model)

Integrate Python and Simulink

You can use the Python Importer to import Python modules and packages to Simulink.

Open the importer: 

>> obj = Simulink. PythonImporter(); 
>> obj.view(); 

MATLAB Function Block and MATLAB System Block

Implement Python modules in Simulink using MATLAB Function block or MATLAB System block.