multiverse 0.6.2
Bug Fixes
- Fix for #119. Fixes for compatibility with evaluate 1.0.0
Minor Changes
- Added functions to allow export of the multiverse results as JSON to
be compatible with Milliways. EMARs are
not yet supported however, as that requires changes to knitr
- Added support for naming variables when extracting multiple
variables using
extract_variables
- Added support for extraction of objects using
extract_variables
- Changes to printing behavior: (1) we do not allow a user to print
variables declared in a multiverse code block from an R code block (it
will throw an error); (2) we do not show the output of multiverse code
blocks when knitting (as it can create hundreds or thousands of
outputs)
multiverse 0.6.1
Bug Fixes
- Fix for #107. Parallel execution of the multiverse (i.e. when cores
>= 2 as an argument to
execute_multiverse
) resulted in
creation of new environments and execution in these new environments,
which are different from the ones that are stored in the multiverse
object and returned by expand
. Fixed implementation so that
all the bindings from the environments were copied to the stored
environments, and the resultant variable bindings can be accessed
through expand
(thanks, @markromanmiller, #108).
- Throw errors when empty options are declared (eg. trailing commas)
inside
branch
statements
Minor Changes
- Improved support for parallel processing using
future
as backend. Users can set up parallel execution workflows on both unix
and non-unix systems using future
. Multiverses can be
executed in parallel on a local machine or concurrently on a
cluster.
- Support for displaying progress of execution using
txtProgressBar
. Use argument progress = TRUE
for the function execute_multiverse()
.
.errors
column added to the output of
expand
which lists the errors encountered during execution
of the multiverse.
- Updates to the documentation and examples, to improve walkthrough of
the steps involved in creating a multiverse analysis.
multiverse 0.5.0
Major changes: