History

6.10.1 (2018-08-13)

• Add per PMT contribution to tight coincidence level (#717)
• Push git-tag before actual commit (#714)
• clean map name and RunDB example (#715, #716)

6.10.0 (2018-07-19)

• Clustering and desaturation upgrade for high-E (#712)
• Enabling lone-hit simulation in FAX (#708)
• notebooks to update RunDB and fix NN file version issue (#713)
• Some minor improvement: Lower tight coincidence to 2-fold at pax level (#705), speed improvement for map interpolation (#676, #672), update S1 pulse shape parameter (#673, #665), show reconstructed position in WF plot (#677)
• FIX related to Maps/inputs: 4th FDC map (#670), PMT after pulse (#671),maps renaming (#654), add missing tensorflow files (#679), new NN files after SR1 (#712)
• Code style and dependencies: Issue template (#707), solfware dependencies (#701, #683), increase deletion timeout (#667)

6.9.0 (2018-01-18)

• Clustering change: Tuning gap size parameters, using index_of_maximum instead of left_central (#642), configuration of tight coincidence (#644)
• Maps changes: Alpha map using time depending 3d FDC for SR1 (#643), per-PMT S2 LCE map update (#646), True LCE map (#658), Update time-dependent 3D FDC maps for SR1 (#662), Add SR0 and update SR1 corrected LCE maps (#661), Update S2 xy map (#664)
• Tensor Flow NN: 3D FDC for TFNN in SR0 (#656), Time-dependent 3D FDC for Tensor Flow Neural Net (#655), tensorflow based nn of SR0 (#651), tensorflow based nn (#650), New TF Json with bad PMT List (#659)
• Bug FIX: Don’t include inactive PMTs in S1 pattern fit (#663), Bugfix for AFT calculation (#657), * Enhencement: Refactor S1 AFT probability calculation (#660), Add SR-dependent configs from processing (#652), More reasonable electron lifetime default (#647), Add sr1 pmt after pulse configs from beginning/middle/end of the runs (#645)

6.8.3 (2017-09-26)

• saving per channel contribution to tight coincidence (#641)

6.8.2 (2017-09-26)

• New SR0 Net for SR0 reprocessing (#612)
• Tuning clustering for SR1 (#615)
• New maps for SR1 and SR0 (#625, #626 #628, #631)
• Time dependent 3-D FDC (#632)
• FAX update: single electron shape (#616), New PMT after pulse model for SR1 (#635), Update single-PE and S1 waveform model (#639)
• No FDC in pax (#620)
• Delete old maps (#627, #621, #623, #624)
• Bug fixes (#618, #638, #633)
• Code cleaning and style changes (#614, #619, #630, #634)

6.8.1 (2017-09-26)

• New 3D data-driven FDC map (#611)
• Fix de-saturation bug for very rare event (#601)
• SR1 PMT gains set to default in fax config (#606)
• S1 optical photon hit patterns with improved binning (#603)
• New S1 area fraction top map (#598)
• Prepare for supernova trigger (#568)
• Don’t keep LED data in buffer (#600)
• Add random initial pickle to acquisition monitor file (#608)

6.8.0 (2017-08-07)

• Two step gap clustering (#596)
• Add data-driven SPE distributions for each channel (fax) (#595)
• InterpolatingMap bug fix (#582)

6.6.6 (2017-06-28)

• add S2 correction in task list (#573)
• S1 area fraction top plugin (#572)
• Update package versions in environment creation (#578)
• New per-PMT LCE maps for full LXe+GXe volume (#577, #588)
• Fax update for better photo-ionisation model and memory leak fix (#574, #581, #589)
• Classification update (#576, #584)
• Disabling pulse-based desaturation for S1 (#580)
• New Position reconstruction (#591, #592)

6.6.5 (2017-05-04)

• S2 map update for SR1 with new gains (#570)
• S1 LCE map update for SR1 with new gains (#571)
• Re-enable reduction of processed data (#569)
• Coherent noise and maximum S2 afterpulse delay (#562)

6.6.4 (2017-04-24)

• New Neural network training results with PMT 2 off in SR1 (#560)
• Update XENON1T_MV.ini file for gain calibration (#561)
• Field Distortion Correction and LCE maps for SR1 (#563)
• New S2 correction map for SR1 (#565)
• Add threshold on saturation correction reference region to avoid biased correction for S1 (#566)
• Fix event builder error related to UTC timestamp (#567)

6.6.3 (2017-03-30)

• Do not compute S1 AFT for large S1 (#558)
• Don’t delete area<0 hits in fixed range hitfinder mode (#557)

6.6.2 (2017-03-15)

• Fix peak.hits_fraction_top calculation (#555)

  • Bug found in S1AreaFractionTopProbability calculation (#553)

6.6.1 (2017-03-11)

• Updated field correction and S1 LCE map (#551)
• Fix typo in S1 Area fraction top computation (#550)

6.6.0 (2017-03-08)

• Updates to data files:

  • S2(x,y) relative light yield map (#535)
  • Neural net (#531)

• S1 area fraction top probability computed (#533)

• Fix bug in peak.area_midpoint computation

• Simulator: Use realistic SPE spectrum in fax (#528)

• Simulator: Fix bug in self-trigger emulation, AmBe noise file (#537)

6.5.1 (2017-02-28)

• Minor bugfixes (#526, #527, #529)

6.5.0 (2017-02-21)

• Change S1/S2 pairing loop order (#522)
• Minor speed improvements (#520)
• Add S1 hit-based pattern fit variable (#525)
• Simulator: extend afterpules to just after the initial signal (#521)
• Simulator: effective S1 shape model (#524)
• Travis builds fixed (#523)

6.4.2 (2017-02-19)

• Update author list
• Fix drift velocity fetched in the runs database (no change in pax)

6.4.1 (2017-02-18)

• Fix classification minus sign error
• Fix waveform simulator internal bug

6.4.0 (2017-02-17)

• New classification based on new properties. Order S1s by coincidence. (#510)
• Cluster based on central regions of hits (#504)
• Simulator: aft truth info, update for S2, afterpulse gains, fix dead PMT double counting (#516)
• Simulator: Update SE gain, double PE emission (#507)
• Update (r,z) correction map (#512)
• Update S2(x,y) correction map, top and bottom correction map (#515)
• Fix saturated PMT accounting config typo (#513)

6.3.3 (2017-02-14)

• Fix for rare edge cases (#499)
• ROOT-based event display (#503)
• Fax update: afterpulses, gains, S2 AFT (#502)
• Add run number to plot_dir filename (#500)

6.3.2 (2017-02-13)

• Fix (r,z) correction bug (#498)
• Neural network update (#495)

6.3.1 (2017-02-09)

• Fix crash due to edge case due to negative samples in hits (#490)
• Remove AWS support (#489)

6.3.0 (2017-02-08)

• Hitfinder upgrade to accomodate long-tailed pe-pulses (#479)
• r-z position correction from field simulation (#482)
• Pulse-based desaturation correction (#480, #484)
• New S1(x,y,z) correction map (#488) and QE rescaling fix
• S2(x,y) correction (#487)
• Realistic photoelectron pulse model (#478)
• Peak property cleanup (#485)
• Pickle output for plotting (#481)
• Plugin to apply software prescale retroactively (#486)

6.2.1 (2017-01-11)

• Fax: first realistic noise and ZLE thresholds (#471)
• Fax: first realistic PMT after-pulse configuration (#475)

6.2.0 (2016-12-20)

• Low-energy classification update, new peak properties (#467)
• Corrected error in peak.mean_area_over_noise computation
• Fix timezone offset for event time in plots (#466)
• Pattern goodness of fit is now a likelihood ratio (#465)
• Event builder robustness to db connectivity issues (#461)
• Fixed-range hit finder (#458)

6.1.1 (2016-11-24)

• Fax update - PMT and S2 afterpulses (#452)
• Include factorial in posrec likelihood (#459)

6.1.0 (2016-11-11)

• Improved clustering (#450)
• Make processed data smaller (#456)
• Write fax output to its own ROOT file (#453)
• Neural network update (#457)
• Fix for acquisition monitor pulse rescue logic (#454)
• Bug related to decimal types in datastructure (#447)
• Assorted event builder improvements after major changes

6.0.2 (2016-10-31)

• More lenient timeouts for batch queue processing

6.0.1 (2016-10-31)

• Fixed memory leak

6.0.0 (2016-10-26)

• Remote/distributed multiprocessing (#439)
• New S1 overall LCE maps (#436) and S2 per-pmt LCE maps (#431)
• Amazon Dynamo support (#438)
• Saturation correction computed, not added to total for XENON1T (#437)
• Matched waveform simulator S2 model (#428)
• Event display improvements (#440, #427)
• Event builder speed improvements (#434)
• Fax naming changes (#443)

5.7.0 (2016-10-07)

• Eventbuilder improvements

• Trigger base logic improvements (#407, 418)
• do_not_trigger option (#419)
• Fix for online dead time calculation (#416)
• Trigger monitor improvements (#412, #424, #429)
• Save acquisition monitor pulses (#414)

• pulses_per_channel and n_pulses attributes added to Event class (#425, #422)
• Fax saves GEANT4 id to ROOT file, more noise options (#426)

5.6.5 (2016-09-21)

• Screwed up release, fixed HISTORY.rst

5.6.4 (2016-09-21)

• Mistaken release

5.6.3 (2016-09-21)

• MV processing support further

5.6.2 (2016-09-01)

• MV processing support

5.6.1 (2016-09-01)

• Muon veto processing support using run database information

5.6.0 (2016-08-31)

• Deal correctly with gains almost zero but see hit (#415)
• LCE map from Kr83m (#413)
• Raw data from MV had suffix (#411)

5.5.1 (2016-08-15)

• Fixes for trigger (#409)
• Adjust drift velocity (#410)

5.5.0 (2016-08-08)

• Muon veto integration (#406)

5.4.0 (2016-08-06)

• New raw data format
• Reconstruction fix (#401)
• Event builder changes (#404)

5.3.6 (2016-07-25)

• Small DAQ fixes.

5.3.5 (2016-07-25)

• Trying to get a DOI

5.3.4 (2016-07-25)

• Small tweak to description in setup.py (which I actually made because I want to get a DOI for pax).

5.3.3 (2016-07-25)

• Small DAQ fixes.

5.3.2 (2016-07-20)

• Only TPC errors can stop DAQ.

5.3.1 (2016-07-19)

• Fix MV bug in ini (#397).

5.3.0 (2016-07-19)

• Trigger changes for deadtime and acquisition monitor - no backword compatbility for new data (#395)
• Muon veto DAQ changes (#396)

5.2.1 (2016-07-06)

• Fix release

5.2.0 (2016-07-06)

• Store acquisition monitor info (#392)

5.1.0 (2016-06-27)

• S1 relative light yield map update (#382)
• Makefile now writes to stable branch (#379)
• Geant4 interface simulates from below cathode if desired (#389)
• Trigger changes with error handling (#386)
• Trigger prevent invalid event ranges (#388)
• Equalized gains (#387)

5.0.2 (2016-06-20)

• Unintentional release.

5.0.1 (2016-06-20)

• Minor DAQ fixes including error handling (#384)

5.0.0 (2016-06-08)

• Ready for XENON1T data
• New clustering (#372)
• Extended trigger window (#369)

4.11.0 (2016-06-06)

• Run database interface (#366)
• Revive PMTs that were masked (#371)

4.10.2 (2016-06-01)

• ROOT fix (#370)
• Configurable low-level info (#368)
• Event builder fixes (scattered commits)

4.10.1 (2016-05-30)

• Add PMT information (#364)
• Event builder changes (#365 plus other commits), including processing related changes.

4.10.0 (2016-05-20)

• Initial pax tuning for XENON1T #361

4.9.3 (2016-05-12)

• Temporarily downgrade scipy due to issues with latest build on some systems
• Event builder: split collections handling, save-all-pulses / mega event / timed trigger mode
• Lowered threshold in XENON1T-LED config until we can specify optimal threshold (#357)
• Waveform simulator bugfix (#354), LED signal simulation (#355)

4.9.2 (2016-05-03)

• Poisson likelihood statistic for position reconstruction, confidence contour improvement (#342)
• Event builder: parallel queries, delete-as-we-go, optimized queries, better config / run_doc handling
• Lock-based race condition prevention for ROOT class compilation (see #351)
• Fix wrong numbers in connector map (#349)

4.9.1 (2016-04-25)

• Neural net uses correct QEs
• Small changes for event builder
• Split S2 afterpulse models so independent for XENON100 and XENON1T

4.9.0 (2016-04-18)

• XENON1T: gains to 1 in LED mode, amplifiers and positions in pmts config dictionary (#339)
• XENON100 S2(x,y) map, XENON100 S2 simulation bugfix (#334)
• Event builder fixes, cax integration
• Pax version no longer append to output filename (0f26ac0)
• Multiprocessing and ROOT fix (#337)
• Waveform simulator afterpulses fix (#341)

4.8.0 (2016-03-29)

• New event builder version (#336)

4.7.0 (2016-03-21)

• Geant4 input to waveform simulator
• Tuning classification for XENON1T gas-mode zero-field.

4.6.1 (2016-03-07)

• Screwed up release, fixing...

4.6.0 (2016-03-07)

• Confidence levels on position reconstruction
• Saturation correction bug
• Several small bug fixes
• Minor event builder changes

4.5.0 (2016-02-26)

• .cpp classes now included within the ROOT output file (#323)
• Area corrections stored separately in datastructure (#322)
• Waveform simulator refactor, PMT afterpulses support (#321)
• Small event builder changes (#316, several loose commits)

4.4.1 (2016-02-05)

• Weird outlier bug fixes found in bulk processing

4.4.0 (2016-02-02)

• New event builder iteration (#297)
• Configuration bugs fixed

4.3.2 (2016-01-31)

• Small argument fixes for default configuration.

4.3.1 (2016-01-28)

• Nasty multiprocessing bug fix

4.3.0 (2016-01-25)

• Parallelization refactor (#298)
• Store meta data in ROOT output (#303)
• z coordinate system now negative in liquid (#302)
• Neural net reconstruction (#296)

4.2.0 (2016-01-11)

• Fixes for ROOT output: memory leak (#282), LED output (#283), long int fields (#289)
• Event builder changes (#278)
• 3D position reconstruction for S1s (#277)
• Hits and Pulses for S1s saved by default (#283)
• Raw data filename format changed, progress bar fix (#289)

4.1.2 (2015-11-30)

• Docs fixes
• TableWriter bug
• Saturation bug #274

4.1.0 (2015-11-17)

• ROOT class output
• Signal processing speedup (#245)
• S1 3d pattern simulation & goodness of fit computation (#237)
• Modifications for working with other TPCs (#247)
• Improvements to / fixes for noisy channel hit rejection
• Assorted bug fixes (#241, #244) and documentation fixes

4.0.1 (2015-10-17)

• Memory leak fixed
• Corrections to position reconstruction (#244)
• Documentation fixes

4.0.0 (2015-10-02)

• Add/remove several peak properties (#223, #214, #203), such as the peak’s hits-only sum waveform.
• Clustering changes: separate plugins, better goodness of split, faster (#223, #213)
• Python 2 support (#217)
• Paxer options to switch input and output type (#212)
• Position reconstruction before classification (#223)
• Fast PatternFitter for position reconstruction (#233)
• Irregular correction map support, XENON100 S1(x,y,z) correction (#219)
• S1 vs S2 classification fix (#221)
• Several bugfixes and documentation improvements (e.g. #230)

3.3.0 (2015-08-03)

• Natural break declustering (#187)
• Improvements to chi2gamma accuracy and speed (#193, #196)
• Non-continuous events in ZippedBSON format (#192)
• XED writing (#177)
• Refactor plugin base and timing code (#190)
• S2 LCE in waveform simulator (#185)
• Cleanup plugin folders and names (#202)
• Minor improvements to logging (#155, #86) and plotting (#98, #144, #200)
• Documentation improvements

3.2.0 (2015-07-06)

• Multithreading of paxer (see –help)
• Clustering bug fixed (#186)
• Contribution section for non-XENON TPCs in examples.
• Chi2 algorithm now runs by default (and has energy cutoff for speed)
• Event builder pretrigger merged into pax
• Units now statically defined
• Various docs improvements

3.1.2 (2015-06-07)

• Update requirements.txt
  • Require new numba version since use new features
  • Pymongo3 required for all our Mongo setups

3.1.1 (2015-06-07)

• Fixed merge issue with minor release (mea culpa)

3.1.0 (2015-06-07)

• Simplified hit finder (#167)
• ZLE in waveform simulator
• BSON output
• Cleanup of Travis building
• Various bug fixes

3.0.0 (2015-04-20)

• Reprocessing capability, switch HDF5 backend (#116)
• Better clustering algorithms: MeanShift, GapSize (#124)
• Hitfinder: faster, new noise definition, work on raw ADC data (#126)
• Bad channel rejection -> suspicious channel testing (#126)
• ROOT output, including tests for Travis (#127)
• Speed and feature improvements to folder-based IO (XED, Avro, ...) (#131)
• Datastructure update (#139)
• Bugfixes, plotting and comment improvements

2.1.0 (2015-02-15)

• Avro raw data output
• Neural net reconstruction
• And lots of meaningless commits to get Travis continuous integration and Coveralls code coverage to work! (And ROOT, which will be in 2.2)

2.0.1 (2015-01-26)

• Travis CI continuous integration is enabled
• Minor bug fixes
  • Values missing from output if they were always default
  • Memory leak after many events due to logger

2.0.0 (2015-01-15)

• Changes to core
  • Cleanup of datastructure (see #80 and #81)
  • Most of core wrapped in Processor class
  • Plugins shut down at end of Processor.run(), not just on destruction
• New signal processing chain
  • BaselineExcursionMethod, finds single-photon peaks in each channel (Default)
  • FindBigPeaks, a traditional sum-waveform peakfinder
  • Supporting peak classification and property computation plugins
• Chi-square gamma x,y position reconstruction
• Waveform simulator enhancements
  • Wrapped in Simulator class, loaded along with processor
  • Performance improvements
  • Basic zero-length encoding emulation
• WritePandas: write our data as DataFrames to containers supported by pandas
• 2D channel waveforms plot
• Support for arbitrary external detectors / extra channel groups
• More tests

1.4.0 (2014-11-21)

• DSP
  • Peak width fields added to datastructure
  • newDSP: Interpolated peak width computations
  • DSP plugins cleaned up and reorganized (except old peak finder)
  • Frequency bandpass filtering support
• Updated docs, comments, logging
• Plots
  • 3D channel waveforms plot
  • Event summary plot
• Music output (fun side project)
• Separate directory for example data files
• Configurations for XAMS, Bern test setup
• –input and –output override settings for most plugins
• WaveformSimulator: improved defaults
• Stable DAQ injector
• Various bug fixes and cleanups that polish

1.3.0 (2014-10-17)

• Plugin directory moved (fix bug in previous release)
• Bulk processing enhancements
  • Scripts for parallelization
  • XED: read in entire datasets, not just single files
• More command line arguments: input, plotting
• Configuration enhancements
  • Module-level settings
  • Multiple inheritance
• DAQInjector
  • New run-database format
  • Repeat single events
  • Create shard index
  • Further debugging and maturing
• Implement run database interface
• WaveformSimulator (Fax) cleanup:
  • Several truth file & instruction file formats
  • Better-motivated settings
  • ER/NR S1s
• Cut overhanging pulses
• Several PosSimple improvements
• Interpolating detector maps (for position-dependent signal corrections)
• Plot 2D hit patterns

1.2.0 (2014-10-02)

• DAQ injector - can inject data into DAQs.
• Nested configurations - better handling of configurations and allows for nesting

1.1.0 (2014-08-29)

• HDF5 output - will be, for now, default output format.
  • We now have a binary output format for peaks and event.
  • Should also be easily extendible to ROOT output, which is blocked until ROOT solves some Py3.4 bugs.
  • Allows bulk comparison with high statistics for things like trigger efficiency
• Bug fixes relating to difference between XENON100 and XENON1T formats (occurences extending past event windows).
• Starting work on a new SimpleDSP processor
• Waveform generator
• General bug fixes and cleanup

1.0.0 (2014-08-16)

• Completely refactored event datastructure

• Moved from Python dictionaries to an event class, seen in pax.datastructure
• Ported all modules with pax to the new structure
• Should open up I/O and C++ binding opportunities
• Now there are Event, Peak, Waveform, and ReconstructedPosition classes
• All of this is based on an extensively modified fork of ‘micromodels’.

• Input control (See Issue #26)

• Can now run pax with single events
• Run paxit –help to see how one can process events

• Binaries of paxit installed when pax is installed
• Improved testing

• Started testing plugins (this will start including other plugins later in the release)
• Extensively testing the event class

• Peak finder now nearly identical to Xerawdp: better than 99.9% agreement on >20000 peaks tested

• Simulation of the Xerawdp convolution bug (filtered waveform mutilation around pulse edges)
• Small bugfixes (empty isolation test regions, strange behaviour when max of filtered waveform is negative)
• Xerawdp XML file interpretation is off-by one (min_width=10 means: width must be 11 or higher)

• Integration of a waveform simulator (FaX) which can simulate S1s, S2s, and white noise

• Script to convert from MC/NEST root files to FaX instructions
• Simplified but much faster simulation mode used for peaks >1000 pe

• Plotting improvement: largest S1 & S2 in separate subplot
• Numerous bug fixes:

• Pickler I/O
• Remove dead code (clustering)

0.2.1 (2014-08-14)

• paxit binaries installed by default to allow working out of source

0.2.0 (2014-08-04)

• Define static event class data structure
• Transforms now specified in ini file
• Can launch small web server for viewing plots
• Major changes to the peak finding to better match Xerawdp. Agreement is currently at the 95% level.

• Two important bugfixes for determining included channels : XED channel mask parsing, 0->1 start
• Filter impulse response now identical to Xerawdp
• Different summed waveforms for s1 and s2 peakfinding

• Transforms have start and stop methods

0.1.0 (2014-07-18)

• First release of software framework
• Functional but not complete digital signal processing * Sum waveform for top, bottom, veto * Filtering with raised cosine filter * Peak finding of S1 and S2
• Basic inputs * MongoDB (used online for DAQ) * XED (XENON100 format)
• Basic outputs * ROOT * Pickle * Plots
• Demo reconstruction algorithm of charge-weighted sum