Update online() post processing and class vignette

NAMESPACE

@@ -23,6 +23,7 @@ export(make_knots)
 export(online)
 export(oracle)
 export(penalty)
+export(post_process_model)
 export(splines2_basis)
 export(tidy)
 import(Rcpp)

NEWS.md

@@ -8,6 +8,7 @@ profoc 1.3.0
 * A new article on the usage of the `conline` C++ class was added.
 * Various functions are now exported to R to allow easier interaction with the `conline` C++ class. These functions are: `init_experts_list()`, `make_basis_mats` and `make_hat_mats`
 * The code of `online()` was simplified a bit by utilizing the new `init_experts_list()` function.
+* Function `post_process_model()` was improved and is now exposed to be used in conjunction with the `conline` C++ class.
 * Move aggregation of timings from cppclock.R to clock.h. This make it faster, easier to maintain and simplifies the code (which will be used in python in the future as well).
 
 profoc 1.2.1

R/misc.R

@@ -105,6 +105,17 @@ val_or_def <- function(val, def) {
     }
 }
 
+
+#' Post Process Data from conline Class
+#'
+#' This function works in conjunction with the conline class.
+#' After the main learning task, it takes the output of the
+#' conline class and returns an object suitable for, visualization,
+#' further, and deployment.
+#' analysis.
+#' @param model_instance  An instance of conline.
+#' @param names A named list with dimnames of `y` and `experts`.
+#' @export
 post_process_model <- function(model_instance, names) {
     # Generate output
     model <- list(

R/online.R

@@ -457,7 +457,7 @@ online <- function(y, experts, tau,
 
     model <- post_process_model(model_instance, names)
 
-    model_instance$teardown()
+    model_instance$get_times()
     rm(model_instance)
 
     return(model)

R/online_update.R

@@ -74,7 +74,7 @@ update.online <- function(object,
 
     object <- post_process_model(model_instance, names = names)
 
-    model_instance$teardown()
+    model_instance$get_times()
     rm(model_instance)
 
     return(object)

inst/include/conline.h

@@ -25,10 +25,15 @@ class conline
     std::string method = "bewa";
 
     std::map<std::string, arma::colvec> params;
-    std::map<std::string, arma::colvec> params_basis_pr;
-    std::map<std::string, arma::colvec> params_basis_mv;
-    std::map<std::string, arma::colvec> params_hat_pr;
-    std::map<std::string, arma::colvec> params_hat_mv;
+    std::map<std::string, arma::colvec> params_basis_pr = {{"",
+                                                            arma::ones(0)}};
+    std::map<std::string, arma::colvec> params_basis_mv = {{"",
+                                                            arma::ones(0)}};
+    std::map<std::string, arma::colvec> params_hat_pr = {{"",
+                                                          arma::ones(0)}};
+    std::map<std::string, arma::colvec> params_hat_mv = {{"",
+                                                          arma::ones(0)}};
+
     double forget_past_performance = 0.0;
     bool allow_quantile_crossing = false;
     bool trace = true;
@@ -104,8 +109,7 @@ class conline
         Rcpp::List &object,
         arma::mat &new_y,
         arma::field<arma::cube> &new_experts);
-    Rcpp::List output();
-    void teardown()
+    void get_times()
     {
         clock.stop();
     };

src/conline_exports.cpp

@@ -60,5 +60,5 @@ RCPP_MODULE(conlineEx)
       .method("getP", &conline::getP)
       .method("getK", &conline::getK)
       .method("getX", &conline::getX)
-      .method("teardown", &conline::teardown);
+      .method("get_times", &conline::get_times);
 }

vignettes/class.Rmd

@@ -155,5 +155,31 @@ The learning process fills the class objects. So we can inspect them using the `
 head(model$weights[[T]][, , 1])
 ```
 
-However, we can also use the post processing functions from `online()` to access the results. This will create output that is identical to the output of `online()`.
+However, we can also use the post processing function of `online()` to access the results. This will create output that is identical to the output of `online()`:
 
+```{r}
+names <- list(y = dimnames(y))
+names$experts <- list(
+  1:T,
+  paste("Marginal", 1:D),
+  tau,
+  paste("Expert", 1:N)
+)
+
+output <- post_process_model(model, names)
+```
+
+We can now use `output` in `update()`, `plot()` and others. 
+
+At this point, we do not need to keep the model in memory anymore. So we can delete it:
+
+```{r}
+rm(model)
+```
+
+## Summary
+
+The C++ class `conline` allows you to gain fine grained control over the learning process. However, it is not as convenient as the `online()` wrapper. So you should only use it if you need to alter the default behavior. However, mixing up helper
+functions from `online()` and the C++ class is possible. So you can compute your 
+combinations using the class interface while still being able to use `update()`, 
+`plot()` and others afterward.