bumbl 1.0.4
- bumbl is now licenced under an MIT license
- fixed link in documentation for brkpt()
bumbl 1.0.3
- In autoplot.bumbldf()the first argument has been
re-named fromxtoobjectfor consistency with
otherautoplot()methods.
- bumblnow uses the- lifecyclepackage to
manage function lifecycle
bumbl 1.0.2
- Added a new dataset, test_df, containing simulated data
used for package testing.
- Added a tau_optim_maxitargument tobumbl()that is passed tooptim(). Mostly for
getting consistent convergence failures in tests, but possibly useful to
end user.
bumbl 1.0.1
- Updated package citation and author list
bumbl 1.0.0
- First “stable” release
- fixed bug with duplicate column names
- added note in documentation clarifying that bumbl()assumes a switchpoint, and doesn’t test for one.
bumbl 0.1.0
- bumbl was peer-reviewed
through rOpenSci
- changed bumbl()to useglm()instead oflm()to fit models. Uses log-link gaussian error by
default.
- familyargument in- bumbl()now passed
directly to- glm()unless- "negbin", in which
case- MASS::glm.nb()is used.
- Added ...argument tobumbl()to pass
additional arguments (e.g. offset) toglm()orglm.nb()
- Added a keep.modelargument tobumbl()which outputs a list-column with the models fit to each colony.
- added change in mass to bombusdata
- removed bumbl_plot()and instead added aplot()method for base R plots and anautoplot()method forggplot2
- now requires tidyr version 1.0.0 or greater
- optimization of switchpoint now done with optim().
Removedtauargument tobumbl()andbrkpt()as it is no longer applicable (may be added again
in the future, but with different behavior).
bumbl 0.0.1
- bumbl()no longer errors when a single colony produces
an error
- added a vignette
- updated bombusdata to include cumulative floral
resources
- fixed major bug (issue #12) that cause bumbl()to
return the incorrect colony IDs
- added bubml_plot()for checking the output ofbumbl()
bumbl 0.0.0.9001
- Added a NEWS.mdfile to track changes to the
package.
- This is the first minimally functioning version of bumbl. More
features to come. Feedback welcome.