# forestecology 0.1.0: Refactoring package

- Switched CI from travis to GitHub actions
- Refactored spatial cross-validation in
`run_cv()`

to use `purrr::map_dfr()`

using the `fit_one_fold()`

function
- Transition modeling, prediction, and plotting functions from generics to S3 methods for the
`comp_bayes_lm`

class.
- Aligned inputs and outputs of modeling and prediction functions with S3 modeling conventions and tidy data principles. Namely,
- The
`comp_bayes_lm()`

modeling function takes in a data frame at a level of observation equivalent to that which the model is actually fit to: each row is a unique focal observation/tree rather than focal-competitor observation pairs/trees. The function outputs a model object with several associated modeling S3 methods.
`predict.comp_bayes_lm()`

takes in a model object as its first argument and input data as its second argument. The output, an unnamed vector, has length equal to the input data.

- Added argument checks.

# forestecology 0.1.0.9003: “Bad first draft” of package paper

- Completed “bad first draft” of paper on package itself, including Michigan Big Woods & SCBI running examples
- Further refactoring of alpha-version of
`forestecology`

package code

# forestecology 0.1.0.9002: Created Michigan Big Woods & SCBI data modeling examples

- Got Smithsonian Conservation Biology Institute (loaded as CSV’s directly from SCBI GitHub) example model working
- Got Michigan Big Woods data (data from University of Michigan Deep Blue Data repository pre-loaded in package) example model working
- Go toy example model working in README
- Second pass at clean-up of package

# forestecology 0.1.0.9001: Replicated Allen & Kim (2020) PLOS One results

# forestecology 0.1.0.9000: First version

- Launched alpha-version of
`forestecology`

package