Prototype of new alignment configuration

Overview

This MR contains a prototype for configuring alignment jobs using a Moore-like configuration. Its main purpose is documenting the work done during past hackathons and to serve as basis for the next steps. It includes track and vertex selections, adding constraints and defining the elements to be aligned. The fundamental change is making the Alignment project depend on Moore instead of Phys.

Some documentation how to use the new configuration is provided in Alignment/Humboldt/README.md

Next steps and open points are summarised in #17 and #20

Configurations

The configurations added are

Alignment/Humboldt/options/AlignVPHalvesModules.py #example Velo alignment job
Alignment/Humboldt/options/AlignVPHalvesModules_PrKalman.py #example Velo job using PrKalman (see #16)
Alignment/Humboldt/options/AlignFTStationsLayers.py #example SciFi job
Alignment/Humboldt/options/AlignVPHalvesModules_Analyzer.py #run analyser part of example alignment job
Alignment/Humboldt/options/AlignVPHalvesModules_Iterator.py #run iterator part of example alignment job

Helper functions to set up new configuration are in Alignment/Humboldt/python/Humboldt/utils.py.

Central track and particle selections are started to be defined in

Alignment/Humboldt/python/Humboldt/TrackSelections.py
Alignment/Humboldt/python/Humboldt/ParticleSelections.py

New functionality (WIP)

• add AlignIterator taking parts from OnlineAlignment to perform iterator part offline

Tests

• add Alignment/Humboldt/tests/qmtest/humboldt.qms/align-vp-halves-modules.qmt to test new configuration
• add Alignment/Humboldt/tests/qmtest/humboldt.qms/analyzer-vp-halves-modules.qmt to test Analyzer part
• add Alignment/Humboldt/tests/qmtest/humboldt.qms/iterator-vp-halves-modules.qmt to test Iterator part
• add Alignment/Humboldt/tests/qmtest/humboldt.qms/align-vp-halves-modules-splititer.qmt to test gaudisplititer.py (WIP)

Uses Rec!2495 (merged)

Alignment/Escher/options/velo.py

@@ -95,12 +95,11 @@ def createAlignUpdateTool(surveyconstraintTool=AlignChisqConstraintTool(),
 
 
 
-
-
-with reconstruction.bind(from_file=False):
+# note that the PVs reconstructed by TrackBeamLineVertexFinderSoA do not store the list of associated tracks. Use PatPV3DFuture instead
+from RecoConf.hlt1_tracking import make_reco_pvs, make_PatPV3DFuture_pvs
+with reconstruction.bind(from_file=False), make_reco_pvs.bind( make_pvs_from_velo_tracks=make_PatPV3DFuture_pvs):
     #make tracks and PVs
     # TODO: in principle the PVs should be an optional input for the alignment
-    # TODO: currently, the reconstructed PVs do not store the list of associated tracks. This means that the PVs can not be used by the alignment
     hlt2_tracks = make_hlt2_tracks(light_reco=False)
     best_tracks = hlt2_tracks["Best"]['v1']
     pvs = make_pvs()