Add testing functions (expect_linear/loglinear/quadratic_time) + testthat files

[Anirban166]

• Added simple testing functions for the three main complexity classes (linear, log-linear, quadratic) following whats written here:

  "testing functions, using syntax similar to testthat package. For example expect_linear_time(fun(N)) should fail with an error if fun(N) is not linear in N, i.e. O(N)."

• Added testthat files such as one for constrained changepoint model algorithms (more functions will be added later) with initial test cases of PeakSegOptimal::PeakSegPDPA and PeakSegDP::cDPA made with the help of the above functions (syntax similar to testthat) and placed the optional packages in suggests as mentioned here

@tdhock are these changes okay? Also are the testing functions in the form you expected them to be? please review once

[Anirban166]

• Added expect_time_complexity, which will take care of all expected complexity testing cases, example:

  expect_time_complexity("quadratic", PeakSegDP::cDPA(rpois(data.sizes, 1), rep(1, length(rpois(data.sizes, 1))), 3L), data.sizes = 10^seq(1, 4, by = 0.5))
  [1] quadratic

• Updated expect_x_time functions (x = complexity class) to call expect_time_complexity internally, with the only requirement as a parameter being the ones passed onto asymptoticTimings (no need to pass it, and complexity class will be defined by x) Example usage:

  expect_quadratic_time(PeakSegDP::cDPA(rpois(data.sizes, 1), rep(1, length(rpois(data.sizes, 1))), 3L), data.sizes = 10^seq(1, 4, by = 0.5))
  [1] quadratic