You need to add the Python package vitis
that includes the Python APIs to the Python script. The example below uses the Python
APIs in a Python script to manage an AI Engine component.
# Add package: Vitis Python CLI
import vitis
# Create a Vitis client object -
client = vitis.create_client()
# Set Vitis Workspace
client.set_workspace(path=<workspace_location>)
# Create aie component.
aie_test_comp = client.create_aie_component(name = comp_name, platform = <absolute path of platform>, template = "empty_aie_component")
# Import source files to the component
aie_test_comp.import_files(from_loc = <absolute path of source folder>, files = ['graph.cpp', 'graph.h', 'kernels.h', 'include.h', 'classify.cc', 'hb27_2i.cc'])
# Set top file from the imported source
aie_test_comp.update_top_level_file(top_level_file = 'graph.cpp')
# Build component on target Hardware
aie_test_comp.build(target="hw")
# Print component information
aie_test_comp.report()
The following command launches the Vitis Unified IDE, and executes the Python script:
vitis -s aie_comp.py
The vitis -s command manages the Vitis component (for example, creation, building or editing configuration
file).
Launch the Vitis IDE. Display the AI Engine components in the
workspace using the vitis -w command.
vitis -w <workspace_location>
Vitis creates the AI Engine component and generates the libadf.a file. The component build completes successfully.

To simulate and debug the AI Engine component, refer to the next section.