Update app.R

app.R

@@ -18,6 +18,7 @@ library(shinydashboard)
 library(DT)               # For data tables
 library(ggplot2)          # For basic plotting
 library(fastrerandomize)  # Our rerandomization package
+library(parallel)         # For detecting CPU cores
 
 # For production apps, ensure fastrerandomize is installed:
 # install.packages("devtools")
@@ -55,9 +56,22 @@ baseR_hotellingT2 <- function(X, W) {
 baseR_generate_randomizations <- function(n_units, n_treated, X, accept_prob, random_type,
                                           max_draws, batch_size) {
   
+  # For safety, check if exact enumerations will explode:
+  # If random_type == "exact", we do combn(n_units, n_treated), which might be huge
+  if (random_type == "exact") {
+    n_comb_total <- choose(n_units, n_treated)
+    if (n_comb_total > 1e6) {
+      warning(
+        sprintf("Exact randomization is requested, but that is %s combinations. 
+                 This may be infeasible in terms of memory/time. 
+                 Consider Monte Carlo instead.", format(n_comb_total, big.mark=",")), 
+        immediate. = TRUE
+      )
+    }
+  }
+  
   if (random_type == "exact") {
     # -------------- EXACT RANDOMIZATIONS --------------
-    # All combinations
     cidx <- combn(n_units, n_treated)
     # Build assignment matrix
     n_comb <- ncol(cidx)
@@ -81,10 +95,8 @@ baseR_generate_randomizations <- function(n_units, n_treated, X, accept_prob, ra
   } else {
     # -------------- MONTE CARLO RANDOMIZATIONS --------------
     # We'll sample max_draws permutations
-    # Start with a base assignment vector
     base_assign <- c(rep(1, n_treated), rep(0, n_units - n_treated))
     
-    # shuffle in R
     # We'll store T^2's in chunks to reduce memory overhead
     batch_count <- ceiling(max_draws / batch_size)
     all_assign <- list()
@@ -92,7 +104,6 @@ baseR_generate_randomizations <- function(n_units, n_treated, X, accept_prob, ra
     
     cur_draw <- 0
     for (b in seq_len(batch_count)) {
-      # how many draws in this batch
       ndraws_here <- min(batch_size, max_draws - cur_draw)
       cur_draw <- cur_draw + ndraws_here
       
@@ -255,7 +266,11 @@ ui <- dashboardPage(
               ),
               
               br(),
-              plotOutput("balance_hist", height = "250px")
+              plotOutput("balance_hist", height = "250px"),
+              
+              # Hardware info note
+              br(),
+              uiOutput("hardware_info")
           )
         )
       ),
@@ -482,6 +497,17 @@ server <- function(input, output, session) {
       theme_minimal(base_size = 14)
   })
   
+  # Hardware info (CPU cores, GPU note)
+  output$hardware_info <- renderUI({
+    num_cores <- detectCores(logical = TRUE)
+    # Basic note about GPU (this can be expanded if you have specialized checks)
+    HTML(paste(
+      "<strong>System Hardware Info:</strong><br/>",
+      "Number of CPU cores detected:", num_cores, "<br/>",
+      "With additional CPU or GPU, greater speedups can be expected."
+    ))
+  })
+  
   # -------------------------------------------------------
   # 3. Randomization Test
   # -------------------------------------------------------
@@ -658,8 +684,8 @@ server <- function(input, output, session) {
   })
   
   # A simple plot for the randomization distribution (for demonstration).
-  # In the minimal example, we do not store the entire distribution in 'randomization_test',
-  # so we simply show the observed effect with a placeholder.
+  # In this minimal example, we do not store the entire distribution in 'randomization_test',
+  # so we simply show the observed effect as a point.
   output$test_plot <- renderPlot({
     rt <- RandTestResult()
     if (is.null(rt)) {
@@ -685,4 +711,3 @@ server <- function(input, output, session) {
 # Run the Application
 # ---------------------------------------------------------
 shinyApp(ui = ui, server = server)
-