Major updates
- All prepare_TASK() are using the same format now (closes #4)
- All prepare_TASK() are using the new pipe (|>) and are a bit faster now
Minor updates
- Sync version number with admin, maker, monkeys and manual
- clean_names_analysis() to go from
TEST_DimensionName_DIRd
toTEST Dimension Name
. Should be useful for tables, plots, etc. - sync_server_local() and run_initial_setup() should work now for development protocols.
- Add digits parameter to create_number_series() so create_new_task(get_info_googledoc = TRUE) will output 3 digit trialid's
- All credentials parameters are now credentials_file instead of list_credentials
- Replace all readr instances for data.table to reduce dependencies and speed up things
- Simplify save_files
- Get rid of some unused files/functions
- Started replacing %>% by |> elsewhere
- Add output_formats parameter (closes #33)
- csv files with 'repeated' in the name will be ignored
- Improve run.R
Major updates
- Adapted to work with targets 1.3.0
Minor updates
- Improved some messages
- Delete snaps if not in protocol 999
- Improved input_files test
- Some small fixes to renv cache
- Updated DT version to avoid a false positive in Google malware detector
- Fixed a couple bugs in report_PROGRESS
- Add tasks
Minor updates
- Fixed issues with translated tasks (BNTen)
- Fixed group report
- Other minor fixes and clean ups
Major updates
- jsPsychHelpeR is now an R package
- We can check the canonical_protocol outputs with testthat now, making sure results don't change
- standardized_names now uses an output list instead of global variables.
- NEW get_dimensions_googledoc() function to help create new tasks
- input files can now be in a single ZIP file
- prepare_TEMPLATE() changed to use a named list for dimensions and items
- With admin/create_docker.R we can create a fully reproducible docker container for a specific pid
- Create DF_analysis_blinded to help with blinded analyses. See MacCoun, R., & Perlmutter, S. (2015). Blind analysis: Hide results to seek the truth. Nature, 526(7572), 187-189 (https://doi.org/10.1038/526187a), or Sarafoglou, A., Hoogeveen, S., & Wagenmakers, E. J. (2023). Comparing analysis blinding with preregistration in the many-analysts religion project. Advances in Methods and Practices in Psychological Science, 6(1), 25152459221128319. (https://doi.org/10.1177/25152459221128319)
Minor updates
- improvements in prepare_TEMPLATE:
- get dependencies to _targets_options.R
- add output_formats parameter to make easier to use csv2 (easier to open for Spanish speaking users)
- Use more informative messages
- Squashed lots of small usability bugs
- Create wide versions of DF_raw and DF_clean to be able to check DF before having all tasks
- NEW functions for some of the recurrent admin tasks
- Can use the canonical prepare_FUN for translated tasks where the logic is the same (e.g. BNTen.js can use prepare_BNT()). See
create_targets_file()
- When running initial setup, data can be downloaded, or a local folder with the original data can be passed to the function
- sensitive data is moved out of the data folder regardless of the origin (downloaded or local)
Major updates
- Works with jsPsychMaker 0.2.0
- Working with ~60 tasks
- Automatic setup with run_initial_setup()
- Should be fine when we have multiple responses per screen and multi-select
- First take to automatically create a codebook
- First take to create Snapshots of each task