- update to Hector 3.2.0 (https://github.com/JGCRI/hector/releases/tag/v3.2.0)
- update to Hector 2.5.0 (https://github.com/JGCRI/hector/releases/tag/rcmip-tier1)
- fix install errors with newer setuptools
- fix for pybind issue on ARM
- update to Hector 2.4.0 (revamped multi-biome system)
- update to Hector 2.3.1
- fix
__hector_version__
field - update to Hector 2.3.0 (https://github.com/JGCRI/hector/releases/tag/v2.3.0)
- update to Hector 2.2.2 (https://github.com/JGCRI/hector/releases/tag/v2.2.2)
- includes new output variables from Hector 2.2.1
- fix installation
- update to Hector 2.2.0 (https://github.com/JGCRI/hector/releases/tag/v2.2.0)
- add a new parameter "VOLCANIC_SCALE"
- switch to new versioning scheme: Hector version + ascending number denoting changes in Pyhector
- update to Hector 2.1.0 (https://github.com/JGCRI/hector/releases/tag/v2.1.0)
- switch to new versioning scheme: Hector version + ascending letters denoting changes in Pyhector
- C++ bindings are using pybind11 now (making the hector-wrapper obsolete)
- drop offical support for 2.7 and 3.4 (not tested anymore)
- new in Pyhector:
shutdown
,run(until=None)
,__hector_version__
,clear_observables
,reset
- updated to Hector 2.0.1 (license clarification)
- updated to Hector 2.0.0
- directly works with Hector, no fork needed anymore
- see the Hector changelog for details (https://github.com/JGCRI/hector/releases)
- include more forcings as output variables
- updated underlying Hector version to include the fix for halo carbon double counting (JGCRI/hector#201)
- overhauled docs to include tables for configuration dicts
- fixed start_date bug when not setting observables
- explicitly state C++11 in
setup.py
- enable spinup output to be readable
- config dictionary can also take a Pandas series instead of list of tuples for time series
- add function to export scenarios as CSV files (in Hector format)
- add API docs using Sphinx and Readthedocs
- return
parameters
only when requested inrun
function - allow different configuration objects to be used
- default config object uses Python numbers or booleans instead of
strings, units can be included as tuples like
(35.0, 'pptv')
and time series as lists of tuples like'N2ON_emissions': [(1765, 11), (2000, 8), (2300, 8)]
- first PyPI beta release