Decide on how to use the DRL during AIT / AIV

[hugobuddel]

At METIS CM#16 there was a discussion about the relation between PIP and AIT. This is an overview issue to decide how to use the pipeline during AIT / AIV. Perhaps this issues should be moved to the DRLVT repo once we have it.

Related issue #31 describes what the other subsystems need from PIP during AIT and #32 describes what want to test ourselves during AIT.

There are several problems to address:

• We decided to run 'METIS' as close to the real thing as soon as possible. This would require close interoporation between ScopeSim, the Telescope Simulator, the ICS, etc.
• We should have a shared environment to test on. We decided to use the ICS stack, Wim De Meester can give us a VM image or something like that.
• There will be no official way to run workflows at AIT: we will not create Reflex workflows, and EDPS might not yet exist. We decided to unofficially use the EDPS framework.
• Some recipes will be in the Python prototype stage. We decided to unofficially use the Python capabilities of EDPS/PyESOREX/PyCPL to make them work.

TODO pre-FDR:

• Write this all down in either the DRLVT and/or in an internal note and leave the specifics (e.g. that we want to use EDPS etc) from the official ESO documents.
• Discuss with Wim De Meester / Remko etc.
• Perhaps do some feasibility experiments w.r.t. the post-FDR actions.

TODO, mostly post-FDR:

• Get a VM / Docker image from Wim De Meester.
• Experiment with EDPS / PyESOREX / PyCPL
• Get some early recipes ready, either in C or in Python.
• Ultimately: do a full end-to-end run of planning an observation, performing it, processing the data, do some QA. E.g. with just the flat or the dark