Add praise and encourage arguments to graded() helpers

[gadenbuie]

This PR adds praise and encourage arguments that follow the same pattern as hint.

I also added give_praise() and give_encouragement() which are small wrappers around an internal S3 generic give_random_phrase(). This follows the same pattern as give_code_feedback() and if we do this again we might want to think about one more layer of abstraction (but I'm hoping that we don't need to do this again for a while).

grader <- grade_this({
  if (.result <= 2) fail("Too low.")
  
  fail_if(.result == 3, "Not three.")
  fail_if_equal(4, "Not four.")
  
  if (.result == 5) pass("Five works!")
  pass_if(.result == 6, "Good six!")
  pass_if_equal(7, "Great seven!")
})

gradethis_setup(
  pass.praise = TRUE,
  fail.hint = TRUE,
  fail.encourage = TRUE
)

# Fails with encouragement
grader(mock_this_exercise(2))
#> <gradethis_graded: [Incorrect] Too low. But no need to fret, try it again.>
grader(mock_this_exercise(3))
#> <gradethis_graded: [Incorrect] Not three. Try it again. Perseverence is the key to success. Try it again; next time's the charm!>
# Fails with encouragement and a hint...
grader(mock_this_exercise(4, 5))
#> <gradethis_graded: [Incorrect] Not four. I expected `5` where you wrote `4`. That's okay: you learn more from mistakes than successes. Let's do it one more time.>

# Passes with praise...
grader(mock_this_exercise(5, 5))
#> <gradethis_graded: [Correct] Resplendent! Five works!>
grader(mock_this_exercise(6, 5))
#> <gradethis_graded: [Correct] I couldn't have done it better myself. :) Good six!>
grader(mock_this_exercise(7, 5))
#> <gradethis_graded: [Correct] Terrific! Great seven!>

Similarly, give_praise() and give_encouragement() can be used explicitly, although the global option is likely to be the most common usage pattern.

gradethis_setup(
  pass.praise = FALSE,
  fail.hint = FALSE,
  fail.encourage = FALSE
)

grader <- grader %>% give_praise() %>% give_encouragement()
# Fails with encouragement and a hint...
grader(mock_this_exercise(4, 5))
#> <gradethis_graded: [Incorrect] Not four. Try it again. I have a good feeling about this.>

# Passes with praise...
grader(mock_this_exercise(5, 5))
#> <gradethis_graded: [Correct] You should be proud. Five works!>

[schloerke]

This is similar behavior as give_code_feedback(). Exciting that the pattern is working!