[tor-commits] [metrics-tasks/master] Add total excluded advertised bandwidth (#1854).

[karsten at torproject.org]

commit d279b3a6c37f98330f3ba6f1a5e75351d3204f13
Author: Karsten Loesing <karsten.loesing at gmx.net>
Date:   Mon Nov 26 16:38:55 2012 -0500

    Add total excluded advertised bandwidth (#1854).
---
 task-1854/plot-entropy.R |   24 +++++++++++++++---------
 task-1854/pyextract.py   |   29 +++++++++++++++++++----------
 2 files changed, 34 insertions(+), 19 deletions(-)

diff --git a/task-1854/plot-entropy.R b/task-1854/plot-entropy.R
index 34635e3..95b8b18 100644
--- a/task-1854/plot-entropy.R
+++ b/task-1854/plot-entropy.R
@@ -3,39 +3,45 @@ library(reshape)
 library(scales)
 
 e <- read.csv("extracted.csv", header = FALSE,
-  col.names = c("validafter", "min_adv_bw", "relays", "linf", "graph"),
-  stringsAsFactor = FALSE)
+  col.names = c("validafter", "min_adv_bw", "relays", "linf",
+  "excl_adv_bw", "graph"), stringsAsFactor = FALSE)
 
 l <- e[e$graph == 'last', ]
-l <- data.frame(x = l$min_adv_bw, relays = l$relays, linf = l$linf)
+l <- data.frame(x = l$min_adv_bw, relays = l$relays,
+  excladvbw = l$excl_adv_bw, linf = l$linf)
 l <- melt(l, "x")
 ggplot(l, aes(x = x, y = value)) +
 geom_line() +
 facet_grid(variable ~ ., scales = "free_y") +
 scale_x_log10(name = "\nAdvertised bandwidth cutoff in B/s (log scale)") +
-scale_y_continuous(name = "") +
+scale_y_continuous(name = paste("Number of relays, excluded advertised",
+  "bandwidth, or linf\n")) +
 opts(title = paste("Consensus with valid-after time ", max(e$validafter),
   "\n", sep = ""))
 
 l <- e[e$graph == 'last', ]
-l <- data.frame(x = l$min_adv_bw, relays = l$relays, linf = l$linf)
+l <- data.frame(x = l$min_adv_bw, relays = l$relays,
+  excladvbw = l$excl_adv_bw, linf = l$linf)
 l <- melt(l, "x")
 ggplot(l, aes(x = x, y = value)) +
 geom_line() +
 facet_grid(variable ~ ., scales = "free_y") +
 scale_x_log10(name = "\nAdvertised bandwidth cutoff in B/s (log scale)") +
-scale_y_log10(name = "Number of relays or linf (log scale)") +
+scale_y_log10(name = paste("Number of relays, excluded advertised",
+  "bandwidth, or linf (log scale)\n")) +
 opts(title = paste("Consensus with valid-after time ", max(e$validafter),
   "\n", sep = ""))
 
 l <- e[e$graph == 'last' & e$min_adv_bw >= 10000 & e$min_adv_bw <= 100000, ]
-l <- data.frame(x = l$min_adv_bw, relays = l$relays, linf = l$linf)
+l <- data.frame(x = l$min_adv_bw, relays = l$relays,
+  excladvbw = l$excl_adv_bw, linf = l$linf)
 l <- melt(l, "x")
 ggplot(l, aes(x = x, y = value)) +
 geom_line() +
 facet_grid(variable ~ ., scales = "free_y") +
-scale_x_log10(name = "\nAdvertised bandwidth cutoff in B/s (log scale)") +
-scale_y_continuous(name = "") +
+scale_x_log10(name = "\nAdvertised bandwidth cutoff in B/s (zoomed, log scale)") +
+scale_y_continuous(name = paste("Number of relays, excluded advertised",
+  "bandwidth, or linf\n")) +
 opts(title = paste("Consensus with valid-after time ", max(e$validafter),
   "\n", sep = ""))
 
diff --git a/task-1854/pyextract.py b/task-1854/pyextract.py
index 4220a2a..33614e2 100644
--- a/task-1854/pyextract.py
+++ b/task-1854/pyextract.py
@@ -3,25 +3,34 @@ import sys
 
 def main():
     out_file = open('extracted.csv', 'w')
-    last_validafter = ''
-    last_lines = []
+    prev_validafter, max_validafter = '', ''
+    max_lines = []
+    prev_relays, prev_min_adv_bw = 0, 0
     for line in open('entropy.csv'):
         parts = line.strip().split(',')
         validafter = parts[0]
         min_adv_bw = int(parts[1])
-        relays = parts[2]
+        relays = int(parts[2])
         linf = parts[3]
-        if last_validafter != validafter:
-            last_lines = []
+        if validafter != prev_validafter:
+            prev_relays, prev_min_adv_bw, excluded_adv_bw = 0, 0, 0
             next_cutoffs = [0, 10000, 20000, 30000, 40000, 50000, 75000,
                     100000, 100000000000000000000]
+        excluded_adv_bw += (prev_relays - relays) * prev_min_adv_bw
+        extended_line = "%s,%d" % (line.strip(), excluded_adv_bw, )
+        if validafter > max_validafter:
+            max_validafter = validafter
+            max_lines = []
+        if validafter == max_validafter:
+            max_lines.append(extended_line)
         while min_adv_bw >= next_cutoffs[0]:
-            out_file.write("%s,%d,%s,%s,history\n" % (validafter,
-                    next_cutoffs[0], relays, linf, ))
+            out_file.write("%s,%d,%d,%s,%d,history\n" % (validafter,
+                    next_cutoffs[0], relays, linf, excluded_adv_bw, ))
             next_cutoffs.pop(0)
-        last_lines.append(line.strip())
-        last_validafter = validafter
-    for line in last_lines:
+        prev_validafter = validafter
+        prev_relays = relays
+        prev_min_adv_bw = min_adv_bw
+    for line in max_lines:
         out_file.write(line + ",last\n")
 
 if __name__ == '__main__':