jrising
commented
6 years ago I'm moving on to the optimization-under-uncertainty work, so I want to merge this into master. The big changes, aside from new paleo-specific material, are:
- The introduction of
loadpath, which is likedatapathbut is aware of the parent-dataset option in the configuration file. In general code,loadpathcan always be used in place ofdatapath, butdatapathstill has a role sinceloadpathcallsdatapathfor its actual paths. - A
src/lib/coding.jllibrary, meant to abstract away from the changes that are going on within Julia about handling of CSVs and missing values. - a
knownvariablefunction indatastore.jl, which allows functions elsewhere to just ask for data by a variable name, for some variable, even though how that data is stored can be quite different by dataset. - an
inputcache.jlsystem, where files that would otherwise be read from disk multiple times can be stored in a cache of their original file data. - a principle of configuration-reading, embodied in
readdatasetconfigandmergeconfigsfunctions, whereby parent dataset configs get merged into their children, except for any place they would overwrite their children's values.
Simulation and optimization under the paleoreconstructed record, and work to identify a set of scenarios for optimization under uncertainty.