tidytext 0.4.2

• Added alt text to figures in vignettes and README (#233)

• Update vignette for quanteda::dfm() v4 (#242)

tidytext 0.4.1

• Fixed bug for FREX stm tidier (#228)

tidytext 0.4.0

• hunspell is now a suggested dependency, thanks to @MichaelChirico (#221)
• Added stm() tidiers for high FREX and lift words (#223)
• Removed tweet-specific tokenizers because of changes in upstream dependencies (#227)

tidytext 0.3.4

• Updated the tidy method for a quanteda dfm because of the upcoming release of Matrix (#218)

tidytext 0.3.3

• scale_x/y_reordered() now uses a function labels as its main input (#200)
• Fixed how to_lower is passed to underlying tokenization function for character shingles (#208)
• Added support for tidying STM models that use content, thanks to @jonathanvoelkle (#209)

tidytext 0.3.2

• Update testing for rlang change + testthat 3e

tidytext 0.3.1

• Check for installation of stopwords more gracefully
• Update tidiers and casters for new version of quanteda

tidytext 0.3.0

• Use vdiffr conditionally
• Bug fix/breaking change for collapse argument to unnest_functions(). This argument now takes either NULL (do not collapse text across rows for tokenizing) or a character vector of variables (use said variables to collapse text across rows for tokenizing). This fixes a long-standing bug and provides more consistent behavior, but does change results for many situations (such as n-gram tokenization).

tidytext 0.2.6

• Move one vignette to pkgdown site, because of dependency removal
• Move all CI from Travis to GH actions

tidytext 0.2.5

• reorder_within() now handles multiple variables, thanks to @tmastny (#170)
• Move stopwords to Suggests so tidytext can be installed on older versions of R
• Pass to_lower argument to other tokenizing functions, for more consistent behavior (#175)
• Add glance() method for stm's estimated regressions, thanks to @vincentarelbundock (#176)