Added generate_petro_jobs_for_field_update

[oddvarlia]

Purpose: RMS script to generate petrophysical property jobs for RMS.
Application: A preparation step to setup a workflow in FMU where both field parameters related to petrophysical variables and gaussian random fields for APS is to be updated as field parameters in ERT simultaneously,

Simplified description of a workflow in RMS where this is relevant:
Initial ensemble:
The typical workflow in RMS will be to simulate initial realizations of the petrophysical variable separately for each individual facies and conditioned to only blocked well observations of the petrophysical variables belonging to only one facies. To do this we want to setup RMS petrophysical jobs not using facies grid realization as input but use only one facies for all grid cells in each zone. The petrophysical variables should have identical distribution as was specified in the original petrophysical job for the given facies.

Updated ensemble: The updated petrophysical variables are imported from files updated by ERT.

A python job in RMS will use updated facies realization as filter to put the petrophysical variables for each facies into one grid parameter. The rest of the workflow in RMS will be unchanged.

The RMS python job defined by the script generate_petro_jobs_for_field_update.py will be used as a preparation script and generate RMS jobs for the petrophysical modelling of each individual facies. The input will be an existing petrophysical job using facies realization. The output will be a set of RMS petrophysical jobs, one per facies simulating the petrophysical variables conditioned to only one facies each. This will simplify and reduce manual work needed to setup the necessary jobs. Manual work will then be reduced to put the generated petrophysical jobs into the workflow instead of manually generating each of these jobs.

Closes: #263

[jcrivenaes]

Some changes; do that first and I can take a second look afterwards

[jcrivenaes]

In general OK. If you you have time, add docstrings to functions to explain short what it will do, and we now have a practice by using type-annotations and mypy checks. If not time now, you can add this as an issue for later work.