NEWS.md

In jedwards24/edwards: James Edwards personal package

edwards 0.3.4

• dir_contents(): fixes to level and directory columns. Reorder output columns.
• dir_contents() and dir_files:
  • The directory column is now a fs:path relative path.
  • Add "call" attribute to output.
  • Restrict to single dir input (would not have worked with multiple before anyway).
• dir_files() now gives modification time instead of just date.
• latest_file() now handles ext arguments starting in ".".
• Change factor_to_numeric() convert logical strings consistent with as.logical(). Output is now always numeric.
• count2() and vcount():
  • Add pname argument which defaults to p.
  • Add countp() and countv() as alternative names for count2() and vcount().
• Minor docs fixes.

edwards 0.3.3

• Add dir column to output for dir_files() which gives directory path for each file.
• var_summary(): add str_missing column to output which counts matches to supplied strings.
• var_summary(): remove name attributes from elements in columns of output.
• var_summary() now uses dplyr::n_distinct() instead of length(unique() to count unique values in columns. This should only change counts for data frame columns. Previously the columns in the data frame column were compared for uniqueness, whereas now the rows are compared. For example,

x <- tibble::tibble(a = c(1:3, 1), b = c(1, 1, 1, 1))
length(unique(x))
dplyr::n_distinct(x)

edwards 0.3.2

Deprecated functions

• Renamed count_string() to count_pattern().
• Deprecate prop_ci(). Replaced by response package https://github.com/jedwards24/response.
• Deprecate glmnet_to_table(). See jemodel package https://github.com/jedwards24/jemodel.
• Deprecate ilogit(). Moved to jemodel package.
• Deprecate count_matches2(). Absorbed into rewritten count_matches().

File tools

• Add new columns to dir_contents() output.
• Add recurse argument to dir_contents() and dir_size().
• Add shorten argument to dir_contents().
• Add dir_count_files() and dir_count_dirs().
• Rewrite latest_file() to simplify (breaking).

Changes to existing functions

• Refactor count_matches() (tidier and faster). Absorb count_matches2() into count_matches().
• Refactor find_similar() code and add tests.
• Rewrite min_n() and max_n(). Now length stable due to handling NAs differently (breaking).

New functions

• Add vcount(). Like count() but for vectors.
• Add na_if_string() and na_if_all(). Used to convert elements to NA.

edwards 0.3.0

Remove Modelling Functions (Breaking Change)

Removed all deprecated modelling functions (ones starting rang_, glmnet_, or roc_). All are available in the jemodel package. A simplified version of glmnet_to_table() has been kept in the package because I have used this most often out of these functions.

Remove Deprecated Functions (Breaking Change)

Removed previously deprecated functions: count_at(), count_nas2(), and bin_numeric(). They are replaced respectively by count_over(), var_summary(), and bin_integer().

CMD Check and Tests

• Handle NSE variables so that R CMD check passes.
• Use testthat 3rd edition and ammend tests to handle messages appropriately.

edwards 0.2.0

Modelling Functions

I will be moving functions for modelling to a new package jemodel. To start this:

• Deleted modelling development code (moved to jemodel).
• Deprecate all glmnet, ranger, and roc functions (everything in glmnet.R, ranger.R, and roc.R).

Other

• Add NEWS file.
• Add package-edwards.R for imports.

jedwards24/edwards documentation built on Sept. 2, 2023, 8:16 a.m.