Nestpy

-----------------
New Physics:
* Carried over from v2.3 beta: A new binomial random number generator (C++ default library), e- EE models, beta model with new yields and fluctuations, non-beta-ER (XELDA).
* New beta model is default regardless of E-field, but old one is still accessible
* ER model (betas and gammas weighted) is its own function, callable
* Pb-206 ion coming off wall from alpha decay has correct Ly and Qy versus field (Thomas-Imel box model for recomb)
* The electron extraction efficiency model now includes “optimistic” high e- EE Aprile/PandaX fits (activatable with EPS_GAS negative)

Code Quality and/or Miscellaneous Bug Fixes:
* C++11 -> 17 default, README updated with all new versioning requirements, but old gcc and cmake versions requested to allow backwards-compatibility with nestpy. std::clamp still doesn’t work, so similar function written by hand
* 1.1 -> 1.08 for increasing Qy to match new Zurich W-value measurement, but with new more logical variable names both deep in code and in detector file for user, and with one factor universal in NEST.cpp; general variable renaming for greater clarity
* Numerous cosmetic and aesthetic changes to code, including unused variable removal, while spacing and tabbing made Google clang-format (with shell script for that now included with NEST), if/else Mac dangle warning addressed
* Kr83m yields same but code overhauled to allow min versus max time separation flexibility and easier data comparison, with bug squashed where wrong error message got replayed
* NEST is now 30% faster, cf. v2.2.4, at least when using gcc 7+, despite the new binomial fluctuation function!

nestpy Specific:
* Bindings to energy spectra generators from TestSpectra.cpp, including: tritium and C14 beta sources; AmBe, DD, Cf252 neutron sources; Spin-Independent WIMP Generators.

-----------------
(Pre-releases, see version 1.4.7 for distributions)
<br>
Sync with [NEST v2.2.1](https://github.com/NESTCollaboration/nest/releases/tag/v2.2.1)

-----------------
Sync with [NEST v2.3.0beta](https://github.com/NESTCollaboration/nest/releases/tag/v2.3.0beta)

* Beta model updated to work with LUX C-14 (medium E-fields) in addition to XENON1T Rn-220 calibration, < 100 V/cm (Greg Rischbieter, Matthew Szydagis, Vetri Velan, and Quentin Riffard of LZ)
* Non-beta (L-shell and M-shell) ER data now matched, from the XELDA experiment, with weighting of beta and gamma models (Sophia Farrell, Rice/XENON and Greg R. UAlbany/LZ)
* Work function of 11.5 eV (EXO-200, Zurich) can now be more accurately reproduced with the "remove quanta" flag set to false, matching the EXO and Baudis (new) data sets (Kirsten McMichael, RPI/nEXO and Matthew Szydagis, UAlbany/LZ)
* Systematic error taken into account for 3D XYZ position reconstruction with S2, per LUX Mercury paper (Claudio Silva, Coimbra/LZ).
* Made the default S1 calculation mode hybrid instead of full, for faster simulation of large-energy (e.g., MeV-scale) events
* Tweaked the default detector example (LUX Run03) to work with the latest models and reproduce D-D and 3H bands perfectly still, while also adding units to this LUX detector header file (Szydagis)
* Binomial random number generator now uses default C++ library (Robert James, LZ). Code is a bit slower now (only by ~10%) but is more precise, and matches python, Flamedisx
* The e- ext eff from liquid to gas is now based on both E-field (in liquid) and temperature, not field alone, combining dozen data sets spanning decades, to address concern raised by Sergey Pereverzev, LLNL. So, it's NOT just PIXeY or LLNL, etc.

-----------------
Sync with [NEST v2.2.3](https://github.com/NESTCollaboration/nest/releases/tag/v2.2.3)

* Replaced useTiming variable by an enum and made the GetS1 result a class member. Separated the S1 and S2 calculation modes.
* Made GetS1 return a ref to avoid vector copy
* Made the GetS2 results a private member returned by reference, while also making GetS1 and GetS2 results "const"
* Removed useless, unused variables that caused a lot of memory allocation/deallocation; result of all this and the above: +~1-5% faster
* Updated the parametric S1 calc to account for the truncated-Gaussian SPE and DPE distributions, making it more consistent with "full"
* Changed hybrid-mode transition to be 100 keV, ~500 photon hits in modern TPCs, instead of hits directly, creating a smooth transition
* Efficiency adjustment in the S1 parametric mode that further makes the parametric and full modes (previously useTiming -1,0) closer.
* Changes driven by Quentin Riffard (LZ/LBNL) & Greg Rischbieter (LZ/UA), with ideas from Matthew (UA) & Luke Kreczko (Bristol)

-----------------
Sync with [NEST v2.2.2](https://github.com/NESTCollaboration/nest/releases/tag/v2.2.2)

Code Quality and/or Misc Bug Fixes:
* Added default density argument for LXe case, forcing an argument re-ordering (Sophia)
* Moved position of "delete detector" in execNEST to solve python problem (Albert Baker, Greg R.)
* Approx eff func for single phe made simpler, for FlameDisx (Robert James, Sophia, Matthew)
* More robust rule used for when to approximate binomial as Gaussian (Sophia, Greg R.)
* Warn that you are in a region of too-low time between S1a and S1b for Kr83m only 1x (Sophia)
* Bad-order if-statements simplified with a min within a max for <0, >1 checks (Luke K., Matthew)
New Physics:
* Liquid Ar model for ER fits all the data better now, in both energy and dE/dx bases (Kate K.)

Code Quality and/or Miscellaneous Bug Fixes:
* Deleted unused redundant line in GetS1 that re-calculated the drift time (Quentin Riffard, LBNL/LZ)
* Only print most error and warning messages if verbosity on (Quentin Riffard, LBNL/LZ)
* Updated TravisCI link in README and added note about OSX builds (Chris Tunnell, Rice/XENON)
* Use of abs value func standardized, lines broken up, multi-line string for cerr (Matthew at behest of Luke Kreczko, Bristol/LZ)
New Physics:
* Liquid Xe model for NR is now better behaved at few hundred keV and few hundred in S1: no odd increase in band width caused by Nex/Ni zeroing out and kinking the recombination probability. Mean yields model unchanged, nor recombination fluctuations / skewness. (Matthew and Greg R., UAlbany/LZ)

-----------------
Sync with [NEST v2.2.1patch2](https://github.com/NESTCollaboration/nest/releases/tag/v2.2.1patch2)

Code Quality and/or Misc Bug Fixes:
* Added default density argument for LXe case, forcing an argument re-ordering (Sophia)
* Moved position of "delete detector" in execNEST to solve python problem (Albert Baker, Greg R.)
* Approx eff func for single phe made simpler, for FlameDisx (Robert James, Sophia, Matthew)
* More robust rule used for when to approximate binomial as Gaussian (Sophia, Greg R.)
* Warn that you are in a region of too-low time between S1a and S1b for Kr83m only 1x (Sophia)
* Bad-order if-statements simplified with a min within a max for <0, >1 checks (Luke K., Matthew)
New Physics:
* Liquid Ar model for ER fits all the data better now, in both energy and dE/dx bases (Kate K.)