From e500868b3d2d5ef5596175d338b493c1d6a867dc Mon Sep 17 00:00:00 2001
From: Audrey Yeo <comptesaudrey@gmail.com>
Date: Tue, 7 Jan 2025 10:50:31 +0100
Subject: [PATCH] some first workings

---
 R/plotOc.R        | 25 +++++++++++++++++++----
 examples/plotOc.R | 45 +++++++++++++++++++++++++++++++++++++----
 man/plotOc.Rd     | 51 ++++++++++++++++++++++++++++++++++++++++-------
 3 files changed, 106 insertions(+), 15 deletions(-)

diff --git a/R/plotOc.R b/R/plotOc.R
index c4765cdd..9806db9a 100644
--- a/R/plotOc.R
+++ b/R/plotOc.R
@@ -1,9 +1,10 @@
 #' Display the operating characteristics using an oc object
 #'
-#' Reads results from \code{\link{ocPredprob}}, \code{\link{ocPostprob}}
+#' Reads results from [ocPredprob()]
 #' etc. and displays a bar plot of the operating characteristics
 #'
-#' @param z returned oc value
+#' @typed oc : list
+#' returned oc parameters
 #' @return nothing, only plots as side effect
 #'
 #' @importFrom graphics barplot title
@@ -11,9 +12,25 @@
 #' @example examples/plotOc.R
 #' @export
 #' @keywords graphics
-plotOc <- function(z) {
+plotOc <- function(oc) {
+
+  if (wiggle == FALSE) {
+    data <-  table(oc$Decision, oc$SampleSize) / oc$params$sim
+  } else {
+
+  }
+
+ggplot(oc, aes(x = name, y = value)) +
+  geom_bar(stat = "identity") +
+  ggtitle("Percentage of trials that Go and Stop per look") +
+  ylabs("Percentage %") +
+  xlabs("Looks and sample size")
+
+
+
   ## plot function for oc.predprob or oc.postprob, or the dist versions of them
-  graphics::barplot(table(z$Decision, z$SampleSize) / z$params$sim, beside = TRUE)
+  graphics::barplot(table(oc$Decision, oc$SampleSize) / oc$params$sim, beside = TRUE)
+
 
   ## get the parameter
   parDat <- lapply(z$params, deparse)
diff --git a/examples/plotOc.R b/examples/plotOc.R
index 930ca464..647979d1 100644
--- a/examples/plotOc.R
+++ b/examples/plotOc.R
@@ -1,9 +1,46 @@
 # get operating character result from oc.postprob
 
-res1 <- ocPostprob(
+oc <- ocPostprob(
   nnE = c(10, 20, 30), truep = 0.4, p0 = 0.2,
-  p1 = 0.3, tL = 0.6, tU = 0.8, parE = c(1, 1), sim = 50000
+  p1 = 0.3, tL = 0.6, tU = 0.8, parE = c(1, 1), sim = 100, wiggle = FALSE
 )
-res1$oc
+oc$oc
+# plotOc(oc)
 
-plotOc(res1)
+
+oc <- ocPostprob(
+  nnE = c(10, 20, 30), truep = 0.4, p0 = 0.2,
+  p1 = 0.3, tL = 0.6, tU = 0.8, parE = c(1, 1), sim = 100, wiggle = TRUE
+)
+oc$oc
+
+data <- data.frame(Decision = as.factor(oc$Decision),
+           Looks = oc$SampleSize)
+
+table(data$Decision, data$Looks)
+
+data %>% group_by(Decision, Looks) %>% summarise(n = n())
+
+data %>% group_by(SampleSize) %>% summarise(n = n())
+
+ggplot(dat, aes(x = SampleSize, y = Decision)) +
+  geom_bar(stat = "identity")
+
+data %>% group_by(Looks, Decision) %>% summarise(tot = n())
+
+
+table(oc$Decision, oc$SampleSize) / oc$params$sim
+table(oc$Decision, oc$SampleSize)
+
+oc$Decision <- ifelse(oc$Decision == TRUE, "GO",
+                      ifelse(oc$Decision == FALSE, "Stop", "Grey"))
+
+oc$Decision
+
+data.frame(looks = c(res1$wiggled_nnrE))
+
+table(oc$Decision, oc$SampleSize)
+
+tt = data.frame(Decision = res$Decision,
+                look = res$SampleSize)
+tt %>% group_by(look) %>% summarise(success = per())
diff --git a/man/plotOc.Rd b/man/plotOc.Rd
index 16389213..76965376 100644
--- a/man/plotOc.Rd
+++ b/man/plotOc.Rd
@@ -4,27 +4,64 @@
 \alias{plotOc}
 \title{Display the operating characteristics using an oc object}
 \usage{
-plotOc(z)
+plotOc(oc)
 }
 \arguments{
-\item{z}{returned oc value}
+\item{oc}{(\code{list}):\cr returned oc parameters}
 }
 \value{
 nothing, only plots as side effect
 }
 \description{
-Reads results from \code{\link{ocPredprob}}, \code{\link{ocPostprob}}
+Reads results from \code{\link[=ocPredprob]{ocPredprob()}}
 etc. and displays a bar plot of the operating characteristics
 }
 \examples{
 # get operating character result from oc.postprob
 
-res1 <- ocPostprob(
+oc <- ocPostprob(
   nnE = c(10, 20, 30), truep = 0.4, p0 = 0.2,
-  p1 = 0.3, tL = 0.6, tU = 0.8, parE = c(1, 1), sim = 50000
+  p1 = 0.3, tL = 0.6, tU = 0.8, parE = c(1, 1), sim = 100, wiggle = FALSE
 )
-res1$oc
+oc$oc
+# plotOc(oc)
 
-plotOc(res1)
+
+oc <- ocPostprob(
+  nnE = c(10, 20, 30), truep = 0.4, p0 = 0.2,
+  p1 = 0.3, tL = 0.6, tU = 0.8, parE = c(1, 1), sim = 100, wiggle = TRUE
+)
+oc$oc
+
+data <- data.frame(Decision = as.factor(oc$Decision),
+           Looks = oc$SampleSize)
+
+table(data$Decision, data$Looks)
+
+data \%>\% group_by(Decision, Looks) \%>\% summarise(n = n())
+
+data \%>\% group_by(SampleSize) \%>\% summarise(n = n())
+
+ggplot(dat, aes(x = SampleSize, y = Decision)) +
+  geom_bar(stat = "identity")
+
+data \%>\% group_by(Looks, Decision) \%>\% summarise(tot = n())
+
+
+table(oc$Decision, oc$SampleSize) / oc$params$sim
+table(oc$Decision, oc$SampleSize)
+
+oc$Decision <- ifelse(oc$Decision == TRUE, "GO",
+                      ifelse(oc$Decision == FALSE, "Stop", "Grey"))
+
+oc$Decision
+
+data.frame(looks = c(res1$wiggled_nnrE))
+
+table(oc$Decision, oc$SampleSize)
+
+tt = data.frame(Decision = res$Decision,
+                look = res$SampleSize)
+tt \%>\% group_by(look) \%>\% summarise(success = per())
 }
 \keyword{graphics}