[REMI] Add PPT reconstruction from trigger signal

Description

!737 (merged) removed support for pulse detection via trigger signals recorded on the digitizer and solely relying on the presence of the pulse pattern table. Unfortunately it now became necessary to run the reconstruction on 2019 data, on which it previously never ran. Here the timeserver data was not yet saved (and would've been in a different format anyway). Rather than running an old version of pycalibration or re-add various branches to support trigger edges directly, I decided to instead add a cell in the beginning to construct a fake PPT from the trigger signal.

At the moment I assume a constant pulse pattern for the entire run for simplicity. This seems to hold true for p2159 sofar. For compatibility with pasha, this single pattern is wrapped in a FakeKeyDataFunctor to iterate over it.

While I'm at it, also fixed the metadata written to files by explicitly passing a sequence number as well as run and proposal (due to the old 2019 files missing it).

As with all REMI features, this is not entirely plug'n'play for now and will always require careful set-up for each proposal it it used upon.

How Has This Been Tested?

run = 100  # Run ID.
in_folder = '/gpfs/exfel/exp/SQS/201802/p002159/raw'
calib_config_path = '/gpfs/exfel/exp/SQS/201802/p002159/usr/remi_proc_config.yaml'

instrument_sase = 3

reconstruct_ppt = True
trigger_edge_channel = '4_D'
trigger_edge_offset = -200

Types of changes

• New feature (non-breaking change which adds functionality)

Reviewers

@ahmedk @kluyvert