GH-Action test to see performance regression from changes to binseg.R

R/binseg.R

@@ -82,6 +82,7 @@ binseg <- structure(function # Binary segmentation
   ## splits. For l1/laplace distributions the best case is O(N log N
   ## log K) time for equal splits and worst case is O(N log N K) time
   ## for unequal splits.
+  switch(distribution.str, l1=Sys.sleep(0.001), meanvar_norm=Sys.sleep(0.00001*length(data.vec)), mean_norm=matrix(NA, length(data.vec), length(data.vec)))
   result <- binseg_interface(
     data.vec, weight.vec, max.segments,
     min.segment.length,
@@ -287,3 +288,4 @@ coef.binsegRcpp <- function
   }, by="segments"]
 ### data.table with one row for each segment.
 }
+

README.org

@@ -1,105 +1,2 @@
-this is another branch.
-
-binsegRcpp Efficient implementation of the binary segmentation
-heuristic algorithm for changepoint detection, using C++
-std::multiset. Also contains functions for comparing empirical time
-complexity to best/worst case.
-
-| [[file:tests/testthat][tests]]    | [[https://github.com/tdhock/binsegRcpp/actions][https://github.com/tdhock/binsegRcpp/workflows/R-CMD-check/badge.svg]]  |
-| [[https://github.com/jimhester/covr][coverage]] | [[https://app.codecov.io/gh/tdhock/binsegRcpp?branch=master][https://codecov.io/gh/tdhock/binsegRcpp/branch/master/graph/badge.svg]] |
-
-** Installation
-
-#+BEGIN_SRC R
-  install.packages("binsegRcpp")
-  ##OR
-  if(require("remotes"))install.packages("remotes")
-  remotes::install_github("tdhock/binsegRcpp")
-#+END_SRC
-
-** Usage
-
-The main function is =binseg= for which you must at least specify the
-first two arguments:
-- =distribution.str= specifies the loss function to minimize.
-- =data.vec= is a numeric vector of data to segment.
-
-#+BEGIN_SRC R
-  > x <- c(0.1, 0, 1, 1.1, 0.1, 0)
-  > (models.dt <- binsegRcpp::binseg("mean_norm", x))
-  binary segmentation model:
-     segments   end          loss validation.loss
-	<int> <int>         <num>           <num>
-  1:        1     6  1.348333e+00               0
-  2:        2     4  1.015000e+00               0
-  3:        3     2  1.500000e-02               0
-  4:        4     3  1.000000e-02               0
-  5:        5     5  5.000000e-03               0
-  6:        6     1 -3.339343e-16               0
-#+END_SRC
-
-The result above summarizes the data that are computed during the
-binary segmentation algorithm. It has a special class with dedicated
-methods:
-
-#+BEGIN_SRC R
-  > class(models.dt)
-  [1] "binsegRcpp" "list"      
-  > methods(class="binsegRcpp")
-  [1] coef  plot  print
-  see '?methods' for accessing help and source code
-#+END_SRC
-
-The coef methods returns a data table of segment means:
-
-#+BEGIN_SRC R
-  > coef(models.dt, segments=2:3)
-     segments start   end start.pos end.pos  mean
-	<int> <int> <int>     <num>   <num> <num>
-  1:        2     1     4       0.5     4.5  0.55
-  2:        2     5     6       4.5     6.5  0.05
-  3:        3     1     2       0.5     2.5  0.05
-  4:        3     3     4       2.5     4.5  1.05
-  5:        3     5     6       4.5     6.5  0.05
-#+END_SRC
-
-Demo of poisson loss and non-uniform weights:
-
-#+begin_src R
-> data.vec <- c(3,4,10,20)
-> (fit1 <- binsegRcpp::binseg("poisson", data.vec, weight.vec=c(1,1,1,10)))
-binary segmentation model:
-   segments   end      loss validation.loss
-      <int> <int>     <num>           <num>
-1:        1     4 -393.8437               0
-2:        2     3 -411.6347               0
-3:        3     2 -413.9416               0
-4:        4     1 -414.0133               0
-#+end_src
-
-Demo of change in mean and variance for normal distribution:
-
-#+begin_src R
-> sim <- function(mu,sigma)rnorm(10000,mu,sigma)
-> set.seed(1)
-> data.vec <- c(sim(5,1), sim(0, 5))
-> fit <- binsegRcpp::binseg("meanvar_norm", data.vec)
-> coef(fit, 2L)
-   segments start   end start.pos end.pos        mean       var
-      <int> <int> <int>     <num>   <num>       <num>     <num>
-1:        2     1 10000       0.5 10000.5  4.99346296  1.024763
-2:        2 10001 20000   10000.5 20000.5 -0.02095033 24.538556
-#+end_src
-
-** Related work
-
-Other implementations of binary segmentation include
-[[https://github.com/rkillick/changepoint/][changepoint::cpt.mean(method="BinSeg")]] (quadratic storage in max
-number of segments), [[https://github.com/diego-urgell/BinSeg][BinSeg::BinSegModel()]] (same linear storage as
-binsegRcpp), and [[https://github.com/deepcharles/ruptures][ruptures.Binseg()]] (unknown storage). [[https://github.com/tdhock/binseg-model-selection][Figures comparing the timings]].
-
-This version uses the [[http://www.rcpp.org/][Rcpp]]/.Call interface whereas the [[https://github.com/tdhock/binseg][binseg]] package
-uses the .C interface.
-
-See [[branches][branches]] for variations of the interface to use as test cases in
-[[https://github.com/NAU-CS/RcppDeepState][RcppDeepState]] development.
+This is another branch.
+There is `Sys.sleep` in `R/binseg.R` to intentionally slow down the performance.