diff --git a/CHANGELOG b/CHANGELOG index 85777fb1a02554f5316a476914405f60d318523e..d8c9186dee06429c74c35f9d4804a4ef4bec0912 100644 --- a/CHANGELOG +++ b/CHANGELOG @@ -11,4 +11,21 @@ https://keepachangelog.com/en/0.3.0/ - Added functionality to use a custom system_generator. This function allows the user to provide their own generator that yields system dictionaries. Can be useful for some random sampling, or interfacing with other codes - merged the autoresolution branch that contains: - Many updates in terms of functionality of the grid - - Changed the grid_variable structure and grid code generation + - improved verbose output to be in boxes, use ANSI colours + - and report total maximum memory usage throughout the run + - added postcode, topcode and bottomcode options to place extra user-defined code in the gridcode + - removed the cleanup of caches (e.g. for the Moe interpolators) when the object is nested: the caches need to stay until the process shuts down (they're not big) + - resolution parameter deprecated, now uses the samplerfunc only + - the samplerfunc grid parameter replaces spacingfunc: this returns a list of sampled points + - grid spacing updated with left, right, center locators + - const_dt spacing function now allows you to set a timestep and obtain a list of masses that fill the timesteps (also in log time) + - gaussian_zoom spacing function allows you to zoom in a particular region of parameter space + - changed the grid_variable structure and grid code generation + - ensemble functions put into ensemble.py + - new load_ensemble and write_ensemble functions to put new functionality into single access functions + - automatically gzip or bzip2 compress ensembles based on file extension + - added msgpack support based on file extension + - JSON decoding done with simplejson because it's faster + - dictionary utility functions put into dicts.py + - some new functions to parse recursively dict keys and/or values that are faster than David's, but probably not as flexible. When these are all you need, they're more efficient. + - added a few new Jupyter notebooks