Added top level tempo object that exposes context specific objects

[axsaucedo]

Ability to expose more friendlier interface to the tempo object which holds insights but also payload information (and could contain metrics, etc).

Example

from tempo.magic import tempo

local_options = RuntimeOptions(**{"insights_options": {"worker_endpoint": DefaultInsightsLocalEndpoint }})

@pipeline(
    name='insights-pipeline',
    uri="s3://tempo/insights-pipeline/resources",
    local_folder=ARTIFACTS_FOLDER,
    runtime_options=runtime_options,
)
class Pipeline:
    
    @predictmethod
    def predict(self, data: np.ndarray, parameters: dict) -> np.ndarray:
        if parameters.get("log"):
            tempo.insights.log_request()
            tempo.insights.log_response()
            tempo.insights.log(parameters)
        return data