Hello, I need to create manually a langfuse trace after I created an opentelemetry output saved in a json file (I created a custom exporter). The opentelemetry output is a list of disctionaries that look exactly like those in here.

Each opentelemetry span/observation has start_time and end_time field, but the methods .start_span() and .start_observation() do not have as input argument a start_time timestamp. Whereas there is the possibility to end the span/observation with the method end() that takes as input a timestamp.

from langfuse import get_client
 
langfuse = get_client()
trace_id = langfuse.create_trace_id()

my_otel_spans = [....]
root_otel_span = my_otel_spans[0]
child_otel_span =  my_otel_spans[1]

root = langfuse.start_span(name=root_otel_span["name"], ...)
child = span.start_observation(name=child_otel_span["name"], ...)

child.end(end_time=child_otel_span["end_time"])
root.end(end_time=root_otel_span["end_time"])

May you, please, add the stat_time timestamp into such methods? If you add such argument, then the latency for each span/observation should be correct. The field may be optional, so it would not change the default flow.
Thanks in advance for your attention.