BristolTopGroup / DailyPythonScripts

Python scripts for the daily tasks in particle physics
Other
8 stars 10 forks source link

Towards a long-term DPS version #322

Open kreczko opened 8 years ago

kreczko commented 8 years ago

In the light of the latest set of problems, as well as the old ones I propose a plan for the future version of DPS. Similar to NTP the proposed changes focus on maintainability and usability.

New conda setup

The old setup procedure still requires ROOT from CMSSW. Since ROOT is now available in conda, we can use that (provided it has all the needed flags). I started the implementation in #314 but this included too many changes. The new changes should come in two steps 1) travis setup 2) soolin & condor setup

New DPS structure

DPS started as a collection of scripts but has greatly outgrown its initial purpose. What I suggest is possibly the biggest change to DPS: make it into a python package.

What I mean by that:

At this point we should also decide how we want to name the module: dps, bris_dps, etc.

NTP console extension

The NTP console is a separate project incubating in NTP that can be used to control more than NTP. DPS is a more obvious and easier candidate for "consolification" than NTP. We already have some of the workflows separated in different scripts with appropriate parameters. This could be the final glue to bring NTP and DPS together in one workspace.

One console to rule them all ;).

kreczko commented 8 years ago

Once #323 is merged, part 1 will be finished.