- All
profile*()
functions now throw a warning when using threshold arguments (@Tilmon #146). This is to remind users that the default thresholds are usually the best, and avoid changes by mistake.
- Each workflow now uses default
*threshold
s (#141).
profile*()
are now sensitive totiltIndicatorAfter_options
(#135).
- The
profile*()
functions gain the argumentisic
and deprecate the argumentisic_tilt
, to catch up with changes introduced in tiltIndicatorAfter#112 (#130).
- Change license to GPLv3 (#116).
- Ensure chunks is numeric (#106).
- Remove needless toy files from the toy input/ directory (#103). Those files were toy output included by mistake.
- In the yaml setup change
NULL
to the empty string (#102). This allows the use of the GUI that pops up when clicking "Knit with Params" in RStudio.
- New helpers to handle the cache (#98).
- Polish README and extend Get started (#87).
- Reexport
rmarkdown::render()
(#88).
cache_dir
is now a parameter (#84). By default it'srappdirs::user_cache_dir("tiltWorkflows")
.
- Allow to run chunks in different order (#78).
- In all workflows the parameter
chunks
now defaults toNULL
, which automatically sets a reasonable number of chunks (#74).
-
Add installation instructions from r-universe and point to [posit.cloud])(https://posit.cloud/) (@AnneSchoenauer #67).
-
Import packages that the user needs to render workflows (#68).
- In each workflow, each section now has some explanation.
- Explain indicator and results sections (@Tilmon #65).
- Document that knitting the workflow runs the code and creates the results (@Tilmon #59).
- Show results where readers expect them: Under Results (@Tilmon #60).
- Remove
read_csv_falling_back_to()
without deprecation (#45). This function was internal, and the package is still not exposed to users.
- Default to expecting input/ and output/ in the working directory (#41)
- New
use_toy_input()
helps users get toy input/ .csv files (#42).
use_workflow()
with notemplate
now throws a helpful error message that makesworkflows()
unnecessary (#39). Thus,workflow()
becomes internal and is retired from the public API without deprecation since this package is still experimental and has no users.
- The workflow files now output a
github_document
(#30). This makes it easy for the user to share the output on GitHub.
use_workflow()
now automatically opens the workflow file in interactive sessions (#21).
- Workflows are now split by indicator, and run by chunks (#20). The split by indicator allows the user more granular control over what they want to run. And the computations by chunk run in parallel and caching intermediate results, so the process completes faster, saves memory, and allows resuming after interruptions.
- Parametrize the workflow tilt-profiles.Rmd (#14). This helps users pass their own data via a graphical user interface, without changing the code.
- The workflow "tilt profiles" is now reproducible with toy data (#6).
- New article: Use parameters (#12).
- Depend on tiltToyData (#8). This makes toy datasets available to the user
automatically when they call
library(tiltWorkflows)
.
- Depend on tiltIndicatorAfter (#7). This makes tiltWorkflows more user friendly since it's the only package that users need to install and use -- everything else happens behind the hoods.
- New
use_workflow()
helps users create a workflow file from a template (#5).
- New article "Get started" (#4 @kalashsinghal).