v0.2.2

General

• drop support for Python 2.7.x and Python < 3.7

Big fixes

• fix bug in output writing for montecarlo due to change in behaviour in spotpy
  (e2f695b)
• fix mistake in routing parameter used for "educated guess" as initial conditions
  (5345b1d)

Documentation

• add documentation build with sphinx
  (#3)
• add API reference
  (89dc0b)

Tests

• add GitHub workflow to run tests
  (#4)

v0.2.1

Algorithm

• check for the duration of the warm-up period against the simulation period
  (to avoid the former being greater than the latter)

Bug fixes

• change test on existence of observed streamflow data in get_evaluation_array
  method (which was raising a ValueError since v0.2.0)

Documentation fixes

• update tutorial notebook that was not up-to-date since v0.2.0 and that
  contained errors
• update readme file to remove mention of objective functions that do not
  exist anymore since v0.2.0

v0.2.0

Algorithm

• redesign of run and run_all_steps to allow for the use of the
  version 0.2.0 of smartcpp where a method allsteps makes it possible
  to go through the whole simulation time series in C++

v0.1.4-2

• Corrects another error in parallel computing with observed output files (using unlimited dimension)

v0.1.4-1

• Corrects error in parallel computing with observed output files (still using compression)

v0.1.4

Functionality

• make gauged area optional (default to catchment area)
• allow choice on output files to write out
• indicate smartpy version in output files

Algorithm

• use more versatile way to convert Epoch timestamps to DateTime objects

Bug fixes

• use parallel IO to write NetCDF output files in parallel Monte Carlo simulations
• move netCDF compression to the end of simulation (for netCDF4 v1.4.2 compatibility)

v0.1.3

Algorithm

• add built-in Latin hypercube sampling functionality for Monte Carlo experiments

Functionality

• add a class Total to run existing sampling on an evaluation period
• allow use of netCDF files for outputs (for both single and Monte Carlo runs)
• allow use of netCDF files for sampling (for Monte Carlo runs)
• allow use of netCDF files for evaluation data
• add optional compression when writing CSV files

v0.1.2

Functionality

• allow use of smartpy without the need for evaluation data
• add possibility to set parameter values using a dictionary

v0.1.1

Bug fixes

• fix improper display of README on PyPI due to use of outdated setuptools,
  twine, and wheel when building the distribution

v0.1.0

• first release