NEWS

vader 0.2.1.9001

Renova release, based on CRAN vader 0.2.1 (published 2020-09-07).

NEWS.md: the upstream entries were reformatted as Markdown (# vader 0.2.1 headings, - bullets, numbered list for 0.0.2) so R can parse them; no entry text was changed. No R code, data or DESCRIPTION field was changed.
DESCRIPTION: Version set to 0.2.1.9001; "Compatibility updates by Renova." appended to Description.
The stale MD5 file was removed. LICENSE.md (license text), README.md, NEWS.md and .Rbuildignore were added.

Tibbles do not work in the vader_df function. Worked around this by unlisting tibbles within the vader_df function.
neu_set = F previously had a bug where booster words were still being counted. This has been fixed.
vader_df previously returned numeric results as factors. This has been fixed so vader_df returns numeric results.
Addd error testing and warning messages.
get_vader and vader_df now clean text of quotation marks, which can otherwise cause errors. Users can set rm_qm to F if they want to include quotation marks. (There should be no need for this unless the lexicon or special dictionaries are updated with new words/emoticons that include quotation marks.)

Changed get_vader function so that it returns 1 index item with a character string of individual word scores (instead of multiple index items for each individual word score)
Changed get_vader function so that it returns a named vector of NAs if the text is blank (rather than a single NA)
Simplified vader_df function, changed to lapply rather than for-loop

Fixed bug where text documents with single words would match with idioms containing multiple words.

Contains a new function that calls getVader() on multiple text documents and returns a dataframe of results.

Minor bug fixes have been completed:

The previous version crashed when emoticons started with open parentheses/brackets, such as :(
The Vader lexicon had a handful of duplicate entries. The sentiment scores will return the average of any duplicate lexicon entries.