excluder 0.5.1

NEW FEATURES

• Added new qualtrics_fetch2 dataset that includes labels as column names.

PACKAGE DEVELOPMENT

• The {tidyselect} package has deprecated using the .data pronoun in tidy selections. So .data is no longer used in select() or other selection functions.
• We used the usethis::use_tidy_description() to tidy up DESCRIPTION file.

BUG FIXES

• Fixed bug to now reassign column names when rename = TRUE.

DOCUMENTATION UPDATES

• We have switched from the {devtools} to the {remotes} version of install_github() in the README to reduce dependencies.
• Because the {iptools} package has been removed from CRAN, the links to {iptools} functions now refer to the GitHub site.

excluder 0.5.0

PACKAGE DEVELOPMENT

• The {iptools} package has been replaced with the {ipaddress} package.

DOCUMENTATION UPDATES

• The Getting started vignette has been renamed excluder.
• The data sets have been added to the References page.

excluder 0.4.0

NEW FEATURES

• The use_labels() function extracts the Qualtrics question labels included when importing with qualtRics::fetch_survey() and renames the column names with these labels.
• The qualtrics_fetch data set offers an example data set that is imported using the qualtRics::fetch_survey() function.
• The rename_columns() function renames the columns to match standard Qualtrics column names. This is primarily a utility function for the mark_*() functions but may be useful to users in some circumstances.

MINOR IMPROVEMENTS

• For excluding based on screen resolution, previously you had to provide a minimum width or height. This works fine if all screens are in the same orientation, but if screens differ in their orientation, this isn't sufficient. Now there is an overall minimum argument res_min that will apply to both width and height. That is, now you can test if at least one of the dimensions has a resolution greater than res_min. (#9)

DOCUMENTATION UPDATES

• To use several of the exclusion criteria, you must collect non-anonymized data and/or computer metadata. The documentation now points to the Qualtrics documentation about non-anonymized data and computer metadata.

PACKAGE DEVELOPMENT

• The package now has a logo!
• The number of CRAN downloads is now indicated by a badge.
• Tests for helper functions have been improved.

excluder 0.3.3

BUG FIXES

• Updating {dplyr} to 1.0.8 caught problem with using across() inside is.na(). Instead, it now uses if_all() with is.na() as argument. Thanks to @romainfrancois for pull request #7.
• remove_label_rows() now properly converts numeric data in Status and Finished columns.

Release v0.3.2

MINOR IMPROVEMENTS

• The remove_label_rows() function can now rename columns to match the default column names used in all of the verb function arguments.
• The mark_ip_() function now checks for (1) internet connectivity, (2) whether the IP address data can be downloaded from https://www.iwik.org/ipcountry/, and (3) if the country code is valid. The function fails gracefully if any of these are not met.

DOCUMENTATION UPDATES

• The README and vignette have clarified that multiple check_*() functions should not be piped.
• The *_ip() function documentation has clarified the internet connectivity requirements.

PACKAGE DEVELOPMENT

• The paper for the {excluder} package has now been published in Journal of Open Source Software. The JOSS badge has been added to the README.
• Tests for *_ip() functions are now skipped on CRAN to avoid timeout delays.
• Clean up package in preparation for submission to CRAN.
• Accepted on CRAN!

Release v0.3.1

MINOR IMPROVEMENTS

• The messages now use {cli} to generate messages that more clearly outputs numbers and text.
• All functions now print the output data frames by default but can all be turned off with print = FALSE.
• *_ip() functions now include include_na as an argument (like *_duplicates() and *_location() functions), so users can decide whether to include NA values in the data that meet exclusion criteria.
• There are now five new utility functions that help simplify the primary verb functions. keep_exclusion_column() allows users to keep the exclusion column in the output from check_*() functions and moves the column the first column in the output. mark_rows() does the bulk of the work creating new columns for exclusion criteria and marking rows that meet the criteria. print_data() controls whether the output is printed to the console. print_exclusion() generates the message about how many rows were excluded by the exclude_*() functions. validate_columns() validates the number, names, and type of columns that are inputted as arguments in the verb functions.
• The NEWS.md file is now based on the rOpenSci template.

BUG FIXES

• The unite_exclusions() function now properly removes multiple separators when multiple exclusion criteria are used.
• The mark_duplicates() function now properly counts and includes the correct number of NAs for both IP addresses and locations and properly prints data.

PACKAGE DEVELOPMENT

• The {excluder} package has now been approved by and transferred to rOpenSci. The package was peer reviewed by Joseph O'Brien (@jmobrien) and Julia Silge (@juliasilge), who are now listed as reviewers in the DESCRIPTION file.

Release v0.3.0

• The primary functionality of the package has moved from the check_*() functions to the mark_*() functions. Thus, check_*() and exclude_*() now first call mark_*() to mark the rows with exclusion criteria, then filter the excluded rows. The documentation for check_*() and exclude_*() now inherit the arguments from mark_*(). This change has been updated in the README and Getting Started vignette.
• collapse_exclusions() has been renamed unite_exclusions() to match {tidyverse} terminology. collapse_exclusions() is now deprecated and will be removed later. unite_exlusions() also switched from using NA to "" for rows with no exclusions. Combined columns now no longer have leftover separators.
• The mark_durations() function now marks fast and slow durations separately.
• exclude__*() functions now have print = FALSE and quiet = TRUE set as default argument values.
• Calls to rbind() have been replaced with bind_cols() and dplyr::pull() has been replaced with [[]].
• Calls to all_of() and any_of() now refer to {tidyselect} rather than {dplyr}.
• if() statements are now more robust by using identical() rather than == and && instead of &.
• The {stringer} package is now imported instead of suggested.
• The *_ip() functions and documentation have been updated to fix a bug/typo to clarify that they mark, check, and exclude rows with IP addresses outside of the specified country.
• Package links are replaced with external URLs.
• All mark, check, and exclude functions for a particular exclusion type have been combined into a single R file. So now each exclusion type has its own R file. Similarly, data file scripts have been combined into a single file.

Release v0.2.2

• Lifecycle has been updated to Stable, and the repo status of Active has been added.
• There is now a Getting Started Excluding Data vignette that gives an overview of package usage.
• Users can now specify the separator used between exclusion types in the collapse_exclusions() function. They can also opt to not remove the excluded columns.
• A bug in the collapse_exclusions() function was fixed, so now a subset of exclusion columns can be collapsed into a single column.
• A bug in the remove_label_rows() function was fixed, so now the Finished column is converted to a logical vector.
• A codemeta.json file was created.
• URLs have been replaced or tweaked to address CRAN package check warnings.
• Functions are now organized by topic in the Reference page of the website.

Release v0.2.1

• The argument name for the data frame switched from .data to x to avoid confusion with the {rlang} use of .data.
• Instead of using quo() and sym() to create new names for columns used as arguments, .data[[var]] is now used.
• The dupe_count column was removed from check_duplicates() output. Tests were adjusted to account for the new number of columns.
• check_duplicates() now specifies the number of NA columns.
• URLs have been replaced or tweaked to address CRAN package check warnings.
• Links function reference pages have been added to the README.

Release v0.2.0

• The deidentify() function was added, which removes IP address, location, and computer information columns.