Ignoring my setup details since it isn’t pertinent to my question.
I’ve been experimenting with Graph Composer after wrangling with Deepstream and Gstreamer. They’re both wonderful, yet complex, so I appreciated what Composer sets out to do; however I’m not clear how this can meet my production needs. I need to set up various camera streams and then manipulate them with gstreamer and other C libraries. What is the recommended way to get the output of a Graph Composer graph back into my C code so I can continue using that data in other programs?
Seems you have some misunderstanding for Graph Composer. It is not a code generation tool but some higher level framework which can use DeepStream inside, it is impossible to convert the yaml graph back to c code.
What kind of data do you want to use in other programs? The inferencing outputs?
Im trying to get the inference data and metadata out of gstreamer and into a middleware layer. I understand that Composer doesn’t generate code. So one example would be running an inference, then using a tee to send the inferences video stream over RTP while also on the other side taking all the data about the inference and loading that into my events layer in C (Rust really since that’s the code base, but since Deepstream doesn’t have a Rust API I’m doing video inference in C and then sending the data out). The former I can see this working for, but I’m not clear if the latter is possible. I guess I’m trying to get my head around the use cases where Composer is the best choice vs coding.
There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks
Seems you are talking about the IPC between the DeepStream application process and your own RUST app process.