How to load and save an architecture object

The first step of configuration is to load an architecture object from a set of CTF-based recorded data onto memory. You can update the architecture object as the following section explains. After you finish updating it, you can save it to a yaml-based file, called "architecture file" to reuse the updated object.

An architecture file has structure of a targeted application. dear_ros_node_viewer help you to comprehend structure of the application with the architecture file.

Python API

CARET serves Python-based APIs to load and save an architecture object.

All of the following code snippets can be executed after load environment variables with source /path/to/ros2_caret_ws/install/setup.bash.

Load from a set of CTF-based recorded data

You can load an architecture object with Architecture constructor.

from caret_analyze import Architecture

arch = Architecture('lttng', '/path/to/ctf-based_recorded_data')

You will find caret_analyze.architecture.Architecture-based object, named 'arch'.

To load recorded data of multi-host system, you can pass the list of recorded data to Architecture object.

arch = Architecture('lttng', ['/path/to/host0/data', '/path/to/host1/data'])

Loading an architecture object from CTF-based recorded data tends to be time-consuming task.

Load from a YAML-based architecture file

As I mentioned, CARET serves a function to stored an architecture object into YAML-based architecture file for reusability. It can save loading time and preserve update of the object. You can load it from YAML-based file with only replacing 'lttng' of the first argument of Architecture constructor by 'yaml'.

from caret_analyze import Architecture

arch = Architecture('yaml', '/path/to/architecture.yaml')

It is recommended for you to use YAML-based file to benefit from reusability unless structure of targeted application is changed.

Save

CARET provides Architecture.export method to save an architecture object as follow.

# arch is caret_analyze.architecture.architecture.Architecture-based object


arch.export('/path/to/destination/architecture.yaml')

! readlink -f /path/to/destination//architecture.yaml
# /path/to/destination/architecture.yaml

The argument of arch.export() is string type and means file path to store the arch object. In this sample, architecture.yaml will be created in /path/to/destination directory if the destination path is writable or another file of the same name does exist.

arch.export() has the second argument, force, to allow you to overwrite the arch object into an existing file. The following sample shows how to overwrite.

arch.export('/path/to/destination/architecture.yaml', force=True)

! readlink -f /path/to/destination//architecture.yaml
# /path/to/destination/architecture.yaml

force=True option erases the existing architecture object.

CLI

Create an architecture file via CLI

With the functions I introduced above, you can create a YAML-based file including an architecture object. CARET serves CLI to create it as well. create_architecture_file command plays role of it.

The following sample code shows how to use create_architecture_file command.

source /path/to/ros2_caret_ws/install/setup.bash

ros2 caret create_architecture_file /path/to/ctf-based_recorded_data -o /path/to/destination/architecture.yaml

readlink -f /path/to/destination/architecture.yaml
# /path/to/destination/architecture.yaml

To create YAML-based file of multi-host system, you can pass the list of recorded data to create_architecture_file command.

ros2 caret create_architecture_file /path/to/host0/data /path/to/host1/data -o /path/to/destination/architecture.yaml