new baby and omero groups

and pyyaml (unused at, least directly).

without py-find-1st, speed may be compromised.

renamed get_retained to apply_cutoff

tracking is done by Baby

Solution is to re-rerun postprocessing.

Fixed signal to work on old legacy h5 files, where channels are stored differently.

Missing brackets were giving non-monotonic time values.

channels can can be specified in the general dict, overriding OMERO

Necessary to calculate the noise for the mother and bud.

time_settings/ntimepoints should be int not List[int].

Works for legacy log files only; no examples yet of new log file.

Removed OMERO_channels and spatial_positions from pipeline.py

If magnification differs from the default, re-scale tile_size and
parameters for segmenting traps.