GitLab

Add a custom `logging` formatter for all child loggers in the
TriggerMenuMT package. Due to the deep module hierarchy, logger names
often become too long and make log file reading difficult. The custom
formatter now only shows the first and last module name, e.g.:

Py:TriggerMenuMT.HLTMenuConfig.UnconventionalTracking.FullScanLRTConfiguration
becomes
Py:TriggerMenuMT:FullScanLRTConfiguration

Also remove all custom logger names and use the module `__name__`
consistently.

Savanna Shaw authored Apr 21, 2021 and Edward Moyse committed Apr 21, 2021

A couple of fixes for cosmic trigger MC tests (ATR-23207):
- Adding BG3 (EMPTY) to the bunch group set and setting ForceBunchGroupPattern=True for the CTP simulation when running in cosmics mode so that the EMPTY L1 triggers are active.
- Updating the MC art tests to include the doCosmics=True option so that the everything (including the HLT reconstruction) is configured for cosmics.

Py2 compatibility is no longer needed.

Joerg Stelzer authored Oct 02, 2020 and Adam Edward Barton committed Oct 02, 2020

The script generates the necessary input files and then runs the menu
comparison code

Joerg Stelzer authored Aug 10, 2020 and Vakhtang Tsulaia committed Aug 10, 2020

* Also unify input and output
* Fix float values in some parameters (L1Topo expects int here)

 * Threshold values and ptMinToTopo stored internally in MeV as unsigned int
 * Access methods in MeV, GeV (float), and counts (using energyResolution)
 * Change "emscale" and "jetscale" to "resolutionMeV" in the menu and C++ sw
 * Added eta-dependend pTMinToTopo to jFEX jets
 * Fix EM*V thresholds in the menu such that the eta-ranges and thrvalues match Run 2
 * Some cleanup of ThresholdDefLegacy.py and ItemDef.py to reduce dependence on menu version
 * Added XS definition in menu and access software
 * Added L1Calo specific base class for L1Thresholds (many commonalities between L1Calo thresholds that don't apply to L1Muon thresholds)
 * Change RangeValue from struct to class with proper access functions to member variables
 * Restructur of upload() member function to avoid calling virtual methods in constructor
 * FIX BUG: access to TopoAlgs by output is working now

filter/map in py3 return iterators, not lists.

* filter behaves differently in python 3 (can be called only once)
* dictionaries key order follows insertion
* json dump separators explicitly specified to create uniform behavior
As a result L1 menus generated in python 2 and python 3 are now
identical

Fixes ATR-21075

Joerg Stelzer authored Jan 28, 2020 and Adam Edward Barton committed Jan 28, 2020

Prior all DataStructure-derived classes were holding a copy of the
ptree that they were constructed from. Now they hold either a
reference (ptree*) or own the original (shared_ptr<ptree>).

Basically the objects L1Menu, HLTMenu, and L1-/HLTPrescaleSet will hold
the ptree created from the json files. All sub structures will only have
a pointer to the sub-tree which is part of the big ptree.

This avoids copying ptree's around and inflating the size of the menu.
One has to be carefull when creating the sub-structure and only pass
references to sub-trees of the original tree to the constructors.

 - cmp
 - no destructuring for lambda arguments
 - xrange -> range
 - Fix relative imports.
 - generators -> list
 - iteritems
 - fix use of filter on strings
 - can't use -1 for level in __import__
 - duplicate enum value
 - except statment syntax.
 - func_name
 - izip
 - integer literals starting with 0
 - exec can't set local variables
 - sorting
 - integer vs float division

Updates to L1Item.cxx L1Menu.cxx to reflect current JSON structure. Short term hack to L1/Base/Items.py to duplicate L1 item name information

Joerg Stelzer authored Oct 08, 2019 and Adam Edward Barton committed Oct 08, 2019

Apply suggestion to Trigger/TrigConfiguration/TrigConfInterfaces/TrigConfInterfaces/ITrigConfigSvc.h