Release Notes¶
main¶
Feb. 20, 2024 don la dieu (nega at icecube.umd.edu)¶
IceTray Release v1.9.2
Delete mailinglist files (#3357)
Add upgrade rules to ruff (#3316)
Remove all tab characters used for indentation from python files (#3300)
Dec. 14, 2023 don la dieu (nega at icecube.umd.edu)¶
IceTray Release v1.9.0
Remove Uber Header (I3.h) (#3151)
fix E703 useless-semicolon (#3266)
python ruff cleanup fix EXE001 (#3289)
Sep. 28, 2023 don la dieu (nega at icecube.umd.edu)¶
IceTray Release v1.8.0
replace deprecated unittest asserts with modern equivalents (#3168)
June 22, 2023 don la dieu (nega at icecube.umd.edu)¶
IceTray Release v1.7.0
Make I3Tray.I3Tray and icecube.icetray.I3Tray the same thing. (#3102)
Remove all references to load_pybindings (#3086)
Dec. 12, 2022 don la dieu (nega AT icecube.umd.edu)¶
IceTray Release v1.5.0
Documentation cleanups
Jul. 27, 2022 don la dieu (nega AT icecube.umd.edu)¶
IceTray Release v1.4.0
Drop support for Python 2
Dec. 21, 2021 Kevin Meagher (kmeagher AT icecube.wisc.edu)¶
IceTray Release V01-03-00
Use rpdf instead of ipdf in scripts/Resplit.py
Documentation improvements
Dec. 20, 2019 Alex Olivas (aolivas@umd.edu)¶
Combo Release V00-00-00
March 14th, 2017 Marcel Zoll (marcel.zoll@fysik.su.se)¶
Release V00-03-03
bug with possibly no impact: frame indexes have been switched for cluster assignments in decisionMaker; tests execute smoothly - probably that part of code never has been triggered or is symmetrical
Massive python3 fixup, should get most tests to pass with python3 – KJ Meagher
break out FrameRegister from Helper code –Marcel Zoll
use ‘linefit’ which implements the ‘improvedLineFit’ now – David Schultz
use internal serialization library
Rearrange Code to fully enclose an if/else-statement; ticket #1790 (August 9, 2016) – Marcel Zoll
May 2, 2016 Alex Olivas¶
Release V00-03-02
cleanups towards c++11 compatibility – David Schultz
fix an I3Position comparison (fixed #1539) – Marcel Zoll
April 18, 2015 Meike de With (meike.de.with@desy.de)¶
Release V00-03-01
Use log_fatal if module cannot continue running (in this case, if nFrac == 0 in PartialCOG, to avoid divide-by-zero later on)
– Marcel Zoll
April 8, 2015 Meike de With (meike.de.with@desy.de)¶
Release V00-03-00
Get rid of the internal implementation of the which_split-function now that it is implemented in phys-services
Fixed bug (int devision always yielding zero) in AfterPulse Discard and Tester, which effectively rendered this module useless
Added a py-module ‘DiscardEmptySplits’ removing EmptySplits (splits which have no pulses present in the desired RecoPulseSeriesMap)
Quick fix of malicious behaviour in the reclustering routine, it is working now but might need some more checks
Add parameter “RecombineRecoMaps” to DecisionMaker: keys of I3RecoPulseSeriesMap/Masks in the frames found at these keys will also be recombined, which is useful if already some cleanings etc. have been performed on the SplitPulses!
Fix (documentation and other) compile warnings, general code cleanup
Add option “Offset” to “CreateTimeWindow” so that a additional time can be added to the nominal time-window defined by the first and last pulse in the passed RecoPulseSeriesMap/Mask
Make the createTimeWindow routine more robust, now it will also give a meaningful result (-offset, offset) when there are no pulses
Correct limit comparison in AfterpulseDiscard
Add more documentation
– Marcel Zoll
Replace make_shared by boost::make_shared to fix name collisions
– Christopher Weaver
Recognize I3_TESTDATA
– Alex Olivas
Corrected the check for zero pulse length in the createTimeWindow routine
– Jan Kunnen
December 11, 2014 Meike de With (meike.de.with@desy.de)¶
Release V00-02-00
Remove obsolete cxx-test
Fix python tests
Fix bug where Hypoframe got a null object written if the masks to be united were derived more than once (if you were affected, subsequent modules would have failed, so you would have noticed)
Making the code faster and more efficient
Add Refit-option (default=False) to ReducingLikelihoodTester. Refit = True reproduces the old behaviour (the fit is redone), Refit = False means that just the LLH values for the fit done before are calculated
– Marcel Zoll
Replaced make_shared by boost::make_shared
– Don la Dieu
October 9, 2014 Meike de With (meike.de.with@desy.de)¶
Release V00-01-00
First release! This release was made specifically for the 2012 muon L3 processing metaproject. The tests do not pass yet because the testdata is not available, but it was checked locally that they pass if the test file is used. As soon as this is solved, the project will be added to icerec.
trunk 2014/06/25 mzoll <marcel.zoll@fysik.su.se>¶
added script to just split data, for example to create a testcase: /resources/scripts/split_data.py
Deprecate the FrameSentinel because the functionality can be easier achieved by a python module: tray.Add(lambda f: return f.Has(FlagName))
Rename EmptyFrameDiscard to LowNChDiscard in order to reflect better the intended use of the Module; which is to discard NoiseClusters
Removed al
Change in UI for TrackLikelihoodTester: changed Parameter names:
Afterpulse and NoiseClusterDiscard have now a options ‘Discard’ for direct discard of frames
Adpated to new lilliput.services interface (I prepend from my previous sins)
removed paranoid checks for object existence in places that they have to exist
startet implementation work on helper functions to stich together I3TriggerHierarchies with DecisionMaker
All Discard and DecisionMaker, which in general identify discardable frames, now have the option ‘Discard’ to do this on the fly; Also written objects of these to the frame have the name of teh Module now (e.g. <I3Bool>(GetName()))
Broke out namespace CoincSuite from FrameCombiner into /lib/CoincSuiteHelpers.h/cxx
Cleaned up the code in some places.
Started with Unit Tests //TODO finish these
Two Bugs found in I3FakeCalculator: functions TrackClosestPoint (lines where always intersecting) and TrackSurfaceIntersection (norm was calulated wrong); affected are all Testers which used functions which might have had deminished selection power in the past //TODO investigate
Added ReducingLikelihoodTester, which is what TrackLikelihoodTester actually should have done; It’s selectiveness should be much stronger.
Depricate TrackLikelihoodTester to the benefit of using ReducingLikelihoodTester
Depricate CausausalConnectedTester to the benefit of cogCausalConnectedTester
Depricate CylinderPulsesTester to the benefit of TrackSystemTester
Changed the interface to specify TimeResiduals to pair<double, double> aka dataclasses.make_pair(double, double)
2013/10/03 mzoll <marcel.zoll@fysik.su.se>¶
In AfterpUlseTester and Discard include the specified threshold values in the concideration (< -> <=)
Added new Tester: TrackSystemTester : TODO Documentation
API change for AfterpulsesDiscard: The <I3Int> tag in the afterpulse event is now named after the module name instead of stataid “Afterpulses”!
TODO move functions in I3FakeCalculator into I3Calculator.cxx
TODO Unit-tests: write the following:
Test for every Tester: if it executes and writes a positive result for a special constructed case
The TesterModule and FrameRegister base-functions have to be tested
2013/10/03 mzoll <marcel.zoll@fysik.su.se>¶
found a newly inroduced bug (after code-rewrite) in PartialCOG which prohibited the correct summup (a while loop did break immediatly); the accuricy cogCausalConnected was impacted
modified eventHeaders are now a integral part of the code
2013/09/19 mzoll <marcel.zoll@fysik.su.se>¶
Massive Code Refurbishment with new functionallity
removed redundant code
provided a nicer interface (API) through TesterModule and function pointers; should also run much faster;
/python and /resources are tydied up; lost links an loose ties
OMKeyHash and HitSorting are now the new and nicer libraries from HiveSplitter
checked for some functions if we can use copies instead of references and vise versa; where it makes sense
TODO check if FrameRegister can work with maps instead of vectors to make access faster; NOTE It could however it would not be neccessarily nicer
new Test emerged from AfterpulseDiscard: AfterpulseTest, it does basically the same thing:
Frameregister has been fitted with some more convenience functions that should soon allow to make the subordinated Frame-vectors private;
an experimental parameter to HypoFrameCreater that now will limit the time difference between frames so that Hypoframes are not created from everything, but only objects that make sense
Nosie and AfterpulseDiscard have been fitted with an option to directly discard according frames instead of going through a seperate module; DANGER ReducedCount might have to be readjusted
2013/07/27 mzoll <marcel.zoll@fysik.su.se>¶
Serious code rework
Corrected inherited bug in AfterpulsesDiscard, in the computation of the charged averaged time, which was rather a time/tot_charge
Corrected bug in MutualCylinderPulsesTester, which was just testing the criteria one-sided way instead of two-sided.
2013/06/20 mzoll <marcel.zoll@fysik.su.se>¶
more documentation
more parameter correctness
2013/02/20 mzoll <marcel.zoll@fysik.su.se>¶
many more modules;
DiscardSplits is now doing the right recombinations; reduced count was right all along
AfterPulsesDIscard added after input from Chris Weaver, very appreciated, thanks
2013/02/20 mzoll <marcel.zoll@fysik.su.se>¶
almost there for a final release:
even more modular now
removed old code fragments
TODO Documentation
TODO finalize recombinations of frames: -> DiscardSplits.
TODO complete resources
TODO find sensible default parameters
TODO make histogramming and weighting a little bit more flexible and not so hardcoded: -> histogramming.py
TODO …
