Changelog - to keep track of all relevant changes

please update the categories "new features" and "bugfixes" before a pull request merge!

version 1.2.0
new features:
- major change in internal looping. Now, the radio signal from each shower is calculated and signal arrival times 
  (from different shower, ray racing solutions and receiver positions) that cluster together are simualted as one event
- merge hdf5 utility has multithreading option (and was refactored -> function names changed)
- distance cut can sum up all sourounding shower energies to properly account for possible interference
- noise temperature can be specified per station and channel
- trigger thresholds can be specified per channel
- bandwidth can be specified per station and channel 
- specifying the detector simulation became easier (check out the updated examples)
- memory consumption was optimized to stay <4GB per core
- random realization of showers are saved so that triggered events can be resimulated using the same random realization
- added option for noiseless channels in a "with noise" simulation
- add option to generate events on the fly and pass them directly to the simulation part (no need to save input hdf5 files anymore)
- added uncertainties to CTW cross sections
- 

bugfixes:
- Fixed issue with merge hdf5 utility so that "event_group_ids" are properly unique




version 1.1.2 - 
new features:
- Veff utility can now handle extended bins
- New tutorial and example for the webinar
- trace start time for the electric field models adjusted such that global time of pulse position corresponds to propagation time
- Simplified treatment of reference angles and polarization for the ARZ module
- Proposal 6.1.1 supported
- Safeguard for events at more than 20 degrees from the Cherenkov angle when using the ARZ models
- Antenna model now needs to be fully specified in the detector description (previously `_InfFirn` was automatically appended to the antenna name for antennas below the surface)

bugfixes:
- Fixed primary particle code bug when using Proposal
- Fixed geometry cut when generating secondary particles
- Incorrect Veff of secondary interactions fixed

version 1.1.1 - 2020/03/23
new features
- New version for the ARZ model available (ARZ2020) 
- a list with event ids can be passed to the main simulation class. All events not in this list will not be simulated. 
  This is useful for a quick resimulation of certain events. 
- Alvarez???? Askaryan models now place the trace into the center of the trace (instead of 50ns into the trace) 
- New data set array 'vertex_times' contains the time elapsed from the first interaction to the current interaction
- new utility to split output hdf5 files into smaller chucks (to be able to resimulate events on a cluster)
- Greenland added to proposal config
- new unforced event generator module
- improved Earth model (PREM), path from interaction vertex through Earth is calculated (before interaction was assumed
  to happen at the surface)
- detector description is saved to nur output files
- new handling of random numbers. Each module has its own random generator instance. Global seed can be controlled via 
  config file setting.

bugfixes:
- ARZxxxx and Alvarez2009 Askaryan modules now use the same (random) shower per event. 
- fixes zenith distribution of event generator back to cos(zenith)
- ray tracing precision was increased to 1e-9
- saveguard against too many ray tracing solutions added

version 1.1.0 - 2020/01/15
new features:
- specifying the bandwidth of the detector is no longer required. Now, the bandwidth is calculated automatically from the
  detector_simulation that the user defines by examining the NuRadioReco modules that alter the bandwidth (i.e. bandpass filters
  and hardware respone incoroprator) using the new module registry feature of NuRadioReco. The noise RMS is then calculated
  from the noise temperature and the bandwidth of the detector.
  Alternatively, the user can specify the RMS noise value manually which is useful to simulate an existing detector where
  the noise RMS is known.
- production includes testing of NuRadioReco files
- NuRadioMC simulations with noise and at Moore's Bay site are tested in our CI test suite
- Event generation using PROPOSAL lepton propagation available
- fft and units utilities are removed. These should be taken from NuRadioReco from now on.
  IMPORTANT: Note that in the fft the frequency spectrum is now given per frequency instead of per bin


bugfixes:
	- fixed that Veff utility script didn't calculate effective volumes for individual triggers
	- fixed that the multiple_triggers key was not created for multiple stations
	- Angular distribution now accounts for projected area
	- 

version 1.0.0 - 2019/08/30 - first python 3 release
