This provides an SDK to trace your Python application calls LLMs
$ pip install git+https://github.com/hawku-com/middle_ai_py_sdk.git
from middle_ai_sdk.tracer import Tracer
tracer = Tracer('app_reference')
def foo() -> None:
model_params = { "abc": { "def": '1', "ghi": '2'}, "jkl": '3' }
span = tracer.start_trace('trace_name', 'llm_model_name', model_params, 'prompt', 'user_id', 'thread_id')
...
tracer.end_trace(span, llm_output)