Moved flush to after_call

[dsblank]

Details

This PR adds @track(flush=True) so that you don't have to do it manually after the function has been called.

Before:

from opik import track, flush_tracker

@track()
def streaming_function(input):
    messages = [{"role": "user", "content": input}]
    response = framework.completion(
        model="gpt-3.5-turbo",
        messages=messages,
    )
    return response
streaming_function("Why?")
flush_tracker()

After:

from opik import track

@track(flush=True)
def streaming_function(input):
    messages = [{"role": "user", "content": input}]
    response = framework.completion(
        model="gpt-3.5-turbo",
        messages=messages,
    )
    return response
streaming_function("Why?")

Issues

None, except for adding some convenience.

Testing

No tests added.

Documentation

Docstring for flush added.

[jverre]

This look good to me, thanks :) Let's wait for @japdubengsub to review and approve before we merge

@dsblank Do you think we should update the docs ? Could be good to add a paragraph in the log_traces page

[japdubengsub]

Unfortunately, I wasn't part of the initial discussion on this task since I joined the project a bit later, so I might be slightly out of context.
Everything looks okay!
My only concern is that this might negatively impact batching mode. However, since both this mode and batching are disabled by default, there shouldn't be any significant issues. :)

[andrescrz]

Can you add tests?

[dsblank]

@andrescrz, I can't find a test of tracker.flush_tracker and I don't have enough information to test it myself. Can you point me to something that I could use as a base to copy?

[andrescrz]

@andrescrz, I can't find a test of tracker.flush_tracker and I don't have enough information to test it myself. Can you point me to something that I could use as a base to copy?

As discussed offline, this was tested manually prior to merge, so it's fine to go from my end. We can add a test if any issue or at a later stage.