Can anyone explain how to generate custom JSON for Kafka using nvmsgconv?

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU): NVIDIA GeForce GTX 1650
• DeepStream Version: 6.1
• TensorRT Version: 8.4.3-1
• NVIDIA GPU Driver Version (valid for GPU only): 510.47
• Issue Type( questions, new requirements, bugs): Question

Hi, I’m new to DeepStream and Kafka. I have multiple cameras which have multiple analytics like line-crossing, overcrowding, direction and roi. I want to send this data as a custom JSON to Kafka. The documentation given is very confusing. (Schema customization - Gst-nvmsgconv)

Can someone clarify the following:

  1. Which files should I modify to create custom schemas and payloads?
  2. What is the role of msgconv_config.txt (given in deepstream-test4) and cfg_kafka.txt (also in deepstream-test4)
3

please refer to How to generate metadata on an object-by-object basis in deepstream-test5 app

4

I already saw this example but unfortunately, the steps aren’t quite clear for me. I ended up using ‘kafka-python’ library and processed all my outputs in Python itself. It’s much easier to get the job done without using nvmsgconv and nvmsgbroker. Also I could now send data to multiple topics for a single stream which to my best information is not supported by the plugins that come with DeepStream. So, I don’t see why using those plugins might be a better option than using ‘kafka-python’.

Nvidia should consider letting users to customize the JSON within Python.