V 0.2.0

monkeylearn 0.2.0

• New functions monkey_classify() and monkey_extract() that:
      * Accept as input both a vector and a dataframe and named column
      * Always return a tibble explicitly relating each input to its classification, allowing for the removal of the MD5 hash
      * Have an unnest flag to unnest the output (turn 1 row per input into 1 row per output)
      * Have a .keep_all flag to retain other columns if input is a dataframe
      * Coerce NULL values and empty vectors returned from MonkeyLearn to NAs
      * Include inputs that could not be processed as NAs in the output
      * Message the first 20 indices of inputs that are not sent to the API (these now include NA and NULL values as well as empty strings)
     * Message the currently processing batch

• Bug fixes and improvements to monkeylearn_classify() and monkeylearn_extract() * monkeylearn_classify() can now accept params

• Fix to messaging when unable to connect to MonkeyLearn API

• Default texts per request is set to 200 now (the recommended number), rather than 20

• Addition of message suggesting that users switch to newer functions

• Implementation of ratelimitr. Creation and documentation of two environment variables allowing smarter rate handling when querying the MonkeyLearn API.

• Creation of pkgdown website

• Programmatic test coverage to re-use common tests for multiple circumstances.

• Use of a cowsay monkey when verbose=TRUE