Added a notebook to demostrate the MAGIC-LST1 event coincidence

I added a jupiter notebook to demonstrate the MAGIC-LST1 event coincidence algorithm. The coincident events are searched in each LST subrun file, and it is preferable to submit jobs in parallel. It takes around two minutes to finish the run.

In this algorithm, the maximum number of coincidence is searched within a given time-offset range, and make a coincident event list with the maximizing time-offset. The coincidence window is now fixed as 1 [µs], but this value is empirically decided and has not been optimized yet. Needs to think how to optimize.

Currently, the reconstructed parameters such as energy, direction and hadronness are filled in the different columns in each telescope to avoid the confusion of the estimation method. For example, regarding the hadronness, there are "hadronness_lst" and "hadronness_magic" columns, which are estimated by LST-mono RF and MAGIC-stereo RF, respectively.