pipeless-ai 0.1.7

A framework to build and deploy multimodal perception apps in minutes without worrying about multimedia pipelines

Pipeless Core

Pipeless is a framework to build and deploy multimodal perception apps in minutes without worrying about multimedia pipelines. Pipeless Core is the core component of the framework.

The Pipeless core is split into several components:

• input: Receives the media streams, demux and decode the streams.
• worker: Receives raw media frames, either audio or video frames, and processes them according to the user provided app
• output: Receives the processed raw media frames, encodes and mux them into the proper container format for the output protocol provided by the user

System Dependencies

• Gstreamer 1.20.3. Verify with gst-launch-1.0 --gst-version. Installation instructions here

Python dependencies

• Poetry: find the installation instructions here

Installation

pip install pipeless-ai

Development

To test your changes run the following command from the src directory:

python -m pipeless.pipeless <component> [app_path]

• <component> can be input, worker, output, all (default)
• app_path is required for worker component and must be the path to the app.py (including app.py)

For simplicity, it will load a mock configuration (hardcoded) at src/pipeless/pipeless.py that you can edit for your use case. The hardcoded configuration will only be used when launching the components with the command above, it won't affect if testing with the CLI.

In order to debug, you can set the configuration log_level to DEBUG. If you find an error related to GStreamer and no usefull information has been logged, try using the env var GST_DEBUG=5 to enable GStreamer debug logs. Refer to this page for more information about GStreamer debugging.

Manual Testing

In order to test your changes, start the virtualenv:

poetry shell

After that, go to the cli directory and run poetry install to install the pipeless-ai-cli (CLI) component. Then go to the core directory and run poetry install to install the pipeless-ai module. This will override the upstream dependency from the CLi componet to use your local one.

Verify you environment by ensuring the pipeless modules are pointing to your local directories instead of the PyPi public modules:

pip list | grep pipeless

With that, you should be able to run pipeless run with your changes.