From 1bd9cb6c0f4c0882bccb6a39b618655c080bbc63 Mon Sep 17 00:00:00 2001
From: Rodrigo Arias Mallo <rodrigo.arias@bsc.es>
Date: Fri, 16 Oct 2020 15:55:52 +0200
Subject: [PATCH] Move the plot script to R

---
 garlic/fig/nbody/test.nix         | 65 ---------------------------
 garlic/fig/nbody/test/default.nix | 62 ++++++++++++++++++++++++++
 garlic/fig/nbody/test/plot.R      | 74 +++++++++++++++++++++++++++++++
 overlay.nix                       |  2 +-
 4 files changed, 137 insertions(+), 66 deletions(-)
 delete mode 100644 garlic/fig/nbody/test.nix
 create mode 100644 garlic/fig/nbody/test/default.nix
 create mode 100644 garlic/fig/nbody/test/plot.R

diff --git a/garlic/fig/nbody/test.nix b/garlic/fig/nbody/test.nix
deleted file mode 100644
index 37810f9..0000000
--- a/garlic/fig/nbody/test.nix
+++ /dev/null
@@ -1,65 +0,0 @@
-{
-  stdenv
-, gnuplot
-, jq
-, experiments
-, garlicTools
-, getExpResult
-, writeText
-}:
-
-with garlicTools;
-with stdenv.lib;
-
-let
-  experiment = builtins.elemAt experiments 0;
-  expResult = getExpResult {
-    garlicTemp = "/tmp/garlic-temp";
-    inherit experiment;
-  };
-    #set xrange [16:1024]
-  plotScript = writeText "plot.plg" ''
-    set terminal png size 800,800
-    set output 'out.png'
-    set xrange [*:*]
-
-    set nokey
-    set logscale x 2
-    set logscale y 2
-    set grid
-
-    set xlabel "blocksize"
-    set ylabel "time (s)"
-
-    plot filename using 1:2 with points
-  '';
-
-in stdenv.mkDerivation {
-  name = "plot";
-  phases = [ "installPhase" ];
-  buildInputs = [ jq gnuplot ];
-  preferLocalBuild = true;
-  dontPatchShebangs = true;
-  installPhase = ''
-    mkdir $out
-    for unit in ${expResult}/*/*; do
-      name=$(basename $unit)
-      log="$unit/stdout.log"
-      conf="$unit/garlic_config.json"
-      bs=$(jq .blocksize $conf)
-      awk "/^time /{print $bs, \$2}" $log >> $out/data.csv
-    done
-    gnuplot -e "filename='$out/data.csv'" ${plotScript}
-    cp out.png $out/out.png
-  '';
-  #installPhase = ''
-  #  mkdir $out
-  #  for unit in ${expResult}/*/*; do
-  #    name=$(basename $unit)
-  #    log="$unit/stdout.log"
-  #    bs=$(jq .blocksize $log)
-  #    awk "/^time /{print $bs, \$2}" $log >> $out/data.csv
-  #  done
-  #'';
-    #gnuplot -e "filename='$out/data.csv'" ${plotScript}
-}
diff --git a/garlic/fig/nbody/test/default.nix b/garlic/fig/nbody/test/default.nix
new file mode 100644
index 0000000..b23aba1
--- /dev/null
+++ b/garlic/fig/nbody/test/default.nix
@@ -0,0 +1,62 @@
+{
+  stdenv
+, gnuplot
+, jq
+, experiments
+, garlicTools
+, getExpResult
+, writeText
+, rWrapper
+, rPackages
+}:
+
+with garlicTools;
+with stdenv.lib;
+
+let
+  experiment = builtins.elemAt experiments 0;
+  expResult = getExpResult {
+    garlicTemp = "/tmp/garlic-temp";
+    trebuchetStage = experiment;
+    experimentStage = getExperimentStage experiment;
+  };
+
+  customR = rWrapper.override {
+    packages = with rPackages; [ tidyverse ];
+  };
+
+  plotScript = ./plot.R;
+
+in stdenv.mkDerivation {
+  name = "plot";
+  buildInputs = [ jq gnuplot customR ];
+  preferLocalBuild = true;
+  dontPatchShebangs = true;
+
+  inherit expResult;
+
+  src = ./.;
+
+  buildPhase = ''
+    echo "using results ${expResult}"
+
+    substituteAllInPlace plot.R
+
+    for unit in ${expResult}/*/*; do
+      name=$(basename $unit)
+      log="$unit/stdout.log"
+      conf="$unit/garlic_config.json"
+      bs=$(jq .blocksize $conf)
+      awk "/^time /{print $bs, \$2}" $log >> data.csv
+    done
+
+    Rscript plot.R
+  '';
+
+  installPhase = ''
+    mkdir $out
+    ln -s ${expResult} $out/result
+    cp *.png $out/
+    cp data.csv $out/
+  '';
+}
diff --git a/garlic/fig/nbody/test/plot.R b/garlic/fig/nbody/test/plot.R
new file mode 100644
index 0000000..1a6da21
--- /dev/null
+++ b/garlic/fig/nbody/test/plot.R
@@ -0,0 +1,74 @@
+library(ggplot2)
+library(dplyr)
+library(scales)
+
+# Load the dataset
+df=read.table("data.csv", col.names=c("blocksize", "time"))
+
+bs_unique = unique(df$blocksize)
+nbs=length(bs_unique)
+
+# Normalize the time by the median
+D=group_by(df, blocksize) %>% mutate(tnorm = time / median(time) - 1)
+
+ppi=300
+h=5
+w=5
+png("box.png", width=w*ppi, height=h*ppi, res=ppi)
+
+# Create the plot with the normalized time vs blocksize
+p = ggplot(D, aes(x=blocksize, y=tnorm)) +
+
+	# Labels
+	labs(x="Blocksize", y="Normalized time",
+              title="Nbody granularity",
+              subtitle="@expResult@") +
+
+	# Center the title
+	#theme(plot.title = element_text(hjust = 0.5)) +
+
+	# Black and white mode (useful for printing)
+	#theme_bw() +
+
+	# Draw boxplots
+	geom_boxplot(aes(group=blocksize)) +
+
+	# Use log2 scale in x
+	scale_x_continuous(trans=log2_trans(),
+			   breaks=bs_unique) +
+
+	scale_y_continuous(breaks = scales::pretty_breaks(n = 10)) +
+
+	# Add the maximum allowed error lines
+	geom_hline(yintercept=c(-0.01, 0.01),
+		linetype="dashed", color="red")
+
+# Render the plot
+print(p)
+
+# Save the png image
+dev.off()
+
+D=group_by(df, blocksize) %>% mutate(tnorm = time / median(time) - 1)
+
+png("scatter.png", width=w*ppi, height=h*ppi, res=ppi)
+
+# Create the plot with the normalized time vs blocksize
+p = ggplot(D, aes(x=blocksize, y=time)) +
+
+	labs(x="Blocksize", y="Time (s)",
+              title="Nbody granularity",
+              subtitle="@expResult@") +
+
+	geom_point(
+		   #position=position_jitter(width=0.2, heigh=0)
+		   shape=21, size=1.5) +
+	scale_x_continuous(trans=log2_trans(),
+			   breaks=bs_unique) +
+	scale_y_continuous(trans=log2_trans())
+
+# Render the plot
+print(p)
+
+# Save the png image
+dev.off()
diff --git a/overlay.nix b/overlay.nix
index e4007c9..843e3cb 100644
--- a/overlay.nix
+++ b/overlay.nix
@@ -295,7 +295,7 @@ let
       # Figures generated from the experiments
       fig = {
         nbody = {
-          test = callPackage ./garlic/fig/nbody/test.nix {
+          test = callPackage ./garlic/fig/nbody/test/default.nix {
             experiments = [
               self.bsc.garlic.exp.nbody.tampi
             ];