[tor-commits] [onionperf/master] Remove existing Tor control log visualizations.
karsten at torproject.org
karsten at torproject.org
Tue May 26 08:25:27 UTC 2020
commit 36c9120ddc52c10ed7272ffbafa43850008e83bf
Author: Karsten Loesing <karsten.loesing at gmx.net>
Date: Tue May 26 10:02:45 2020 +0200
Remove existing Tor control log visualizations.
These visualizations are not really useful to us, so we can as well
remove them and all remaining pylab code with it.
As part of this change, output files are being renamed from
tgen.onionperf.viz.yyyy-mm-dd.{csv,pdf} to just
onionperf.viz.yyyy-mm-dd.{csv,pdf}.
Implements #34214.
---
onionperf/onionperf | 20 +-----
onionperf/visualization.py | 150 ++-------------------------------------------
2 files changed, 8 insertions(+), 162 deletions(-)
diff --git a/onionperf/onionperf b/onionperf/onionperf
index ee28344..1edd5a7 100755
--- a/onionperf/onionperf
+++ b/onionperf/onionperf
@@ -334,13 +334,6 @@ files generated by this script will be written""",
action="store", dest="prefix",
default=None)
- visualize_parser.add_argument('-f', '--format',
- help="""A comma-separated LIST of color/line format strings to cycle to
- matplotlib's plot command (see matplotlib.pyplot.plot)""",
- metavar="LIST", type=str,
- action="store", dest="lineformats",
- default=util.LINEFORMATS)
-
# get args and call the command handler for the chosen mode
if len(sys.argv) == 1:
main_parser.print_help()
@@ -429,25 +422,16 @@ def analyze(args):
logging.error("Given paths were an unrecognized mix of file and directory paths, nothing will be analyzed")
def visualize(args):
- from onionperf.visualization import TGenVisualization, TorVisualization
+ from onionperf.visualization import TGenVisualization
from onionperf.analysis import Analysis
- lflist = args.lineformats.strip().split(",")
- lfcycle = cycle(lflist)
-
tgen_viz = TGenVisualization()
- tor_viz = TorVisualization()
-
for (path, label) in args.datasets:
- nextformat = next(lfcycle)
-
anal = Analysis.load(filename=path)
if anal is not None:
- tgen_viz.add_dataset(anal, label, nextformat)
- tor_viz.add_dataset(anal, label, nextformat)
+ tgen_viz.add_dataset(anal, label)
tgen_viz.plot_all(args.prefix)
- tor_viz.plot_all(args.prefix)
def type_nonnegative_integer(value):
i = int(value)
diff --git a/onionperf/visualization.py b/onionperf/visualization.py
index 51ea59e..4547da1 100644
--- a/onionperf/visualization.py
+++ b/onionperf/visualization.py
@@ -6,138 +6,25 @@
import matplotlib; matplotlib.use('Agg') # for systems without X11
from matplotlib.backends.backend_pdf import PdfPages
-import pylab, numpy, time
+import time
from abc import abstractmethod, ABCMeta
import matplotlib.pyplot as plt
import pandas as pd
import seaborn as sns
import datetime
-'''
-pylab.rcParams.update({
- 'backend': 'PDF',
- 'font.size': 16,
- 'figure.max_num_figures' : 50,
- 'figure.figsize': (6, 4.5),
- 'figure.dpi': 100.0,
- 'figure.subplot.left': 0.10,
- 'figure.subplot.right': 0.95,
- 'figure.subplot.bottom': 0.13,
- 'figure.subplot.top': 0.92,
- 'grid.color': '0.1',
- 'axes.grid' : True,
- 'axes.titlesize' : 'small',
- 'axes.labelsize' : 'small',
- 'axes.formatter.limits': (-4, 4),
- 'xtick.labelsize' : 'small',
- 'ytick.labelsize' : 'small',
- 'lines.linewidth' : 2.0,
- 'lines.markeredgewidth' : 0.5,
- 'lines.markersize' : 10,
- 'legend.fontsize' : 'x-small',
- 'legend.fancybox' : False,
- 'legend.shadow' : False,
- 'legend.ncol' : 1.0,
- 'legend.borderaxespad' : 0.5,
- 'legend.numpoints' : 1,
- 'legend.handletextpad' : 0.5,
- 'legend.handlelength' : 1.6,
- 'legend.labelspacing' : .75,
- 'legend.markerscale' : 1.0,
- 'ps.useafm' : True,
- 'pdf.use14corefonts' : True,
- 'text.usetex' : True,
-})
-'''
-
class Visualization(object, metaclass=ABCMeta):
def __init__(self):
self.datasets = []
- def add_dataset(self, analysis, label, lineformat):
- self.datasets.append((analysis, label, lineformat))
+ def add_dataset(self, analysis, label):
+ self.datasets.append((analysis, label))
@abstractmethod
def plot_all(self, output_prefix):
pass
-class TorVisualization(Visualization):
-
- def plot_all(self, output_prefix, relays_only=False):
- self.relays_only = relays_only
- if len(self.datasets) > 0:
- prefix = output_prefix + '.' if output_prefix is not None else ''
- ts = time.strftime("%Y-%m-%d_%H:%M:%S")
- self.page = PdfPages("{0}tor.onionperf.viz.{1}.pdf".format(prefix, ts))
- self.__plot_bytes(direction="bytes_read")
- self.__plot_bytes(direction="bytes_written")
- self.page.close()
-
- def __plot_bytes(self, direction="bytes_written"):
- mafig = pylab.figure()
- allcdffig = pylab.figure()
- eachcdffig = pylab.figure()
-
- for (anal, label, lineformat) in self.datasets:
- tput = {}
- pertput = []
- for node in anal.get_nodes():
- if self.relays_only and 'relay' not in node and 'thority' not in node: continue
- d = anal.get_tor_bandwidth_summary(node, direction)
- if d is None: continue
- for tstr in d:
- mib = d[tstr] / 1048576.0
- t = int(tstr)
- if t not in tput: tput[t] = 0
- tput[t] += mib
- pertput.append(mib)
-
- pylab.figure(mafig.number)
- x = sorted(tput.keys())
- y = [tput[t] for t in x]
- y_ma = movingaverage(y, 60)
- pylab.scatter(x, y, s=0.1)
- pylab.plot(x, y_ma, lineformat, label=label)
-
- pylab.figure(allcdffig.number)
- x, y = getcdf(y)
- pylab.plot(x, y, lineformat, label=label)
-
- pylab.figure(eachcdffig.number)
- x, y = getcdf(pertput)
- pylab.plot(x, y, lineformat, label=label)
-
- pylab.figure(mafig.number)
- pylab.xlabel("Tick (s)")
- pylab.ylabel("Throughput (MiB/s)")
- pylab.xlim(xmin=0.0)
- pylab.ylim(ymin=0.0)
- pylab.title("60 second moving average throughput, {0}, all relays".format("write" if direction == "bytes_written" else "read"))
- pylab.legend(loc="lower right")
- self.page.savefig()
- pylab.close()
- del(mafig)
-
- pylab.figure(allcdffig.number)
- pylab.xlabel("Throughput (MiB/s)")
- pylab.ylabel("Cumulative Fraction")
- pylab.title("1 second throughput, {0}, all relays".format("write" if direction == "bytes_written" else "read"))
- pylab.legend(loc="lower right")
- self.page.savefig()
- pylab.close()
- del(allcdffig)
-
- pylab.figure(eachcdffig.number)
- # pylab.xscale('log')
- pylab.xlabel("Throughput (MiB/s)")
- pylab.ylabel("Cumulative Fraction")
- pylab.title("1 second throughput, {0}, each relay".format("write" if direction == "bytes_written" else "read"))
- pylab.legend(loc="lower right")
- self.page.savefig()
- pylab.close()
- del(eachcdffig)
-
class TGenVisualization(Visualization):
def plot_all(self, output_prefix):
@@ -145,9 +32,9 @@ class TGenVisualization(Visualization):
prefix = output_prefix + '.' if output_prefix is not None else ''
ts = time.strftime("%Y-%m-%d_%H:%M:%S")
self.__extract_data_frame()
- self.data.to_csv("{0}tgen.onionperf.viz.{1}.csv".format(prefix, ts))
+ self.data.to_csv("{0}onionperf.viz.{1}.csv".format(prefix, ts))
sns.set_context("paper")
- self.page = PdfPages("{0}tgen.onionperf.viz.{1}.pdf".format(prefix, ts))
+ self.page = PdfPages("{0}onionperf.viz.{1}.pdf".format(prefix, ts))
self.__plot_firstbyte_ecdf()
self.__plot_firstbyte_time()
self.__plot_lastbyte_ecdf()
@@ -161,7 +48,7 @@ class TGenVisualization(Visualization):
def __extract_data_frame(self):
transfers = []
- for (analysis, label, lineformat) in self.datasets:
+ for (analysis, label) in self.datasets:
for client in analysis.get_nodes():
tgen_transfers = analysis.get_tgen_transfers(client)
for transfer_id, transfer_data in tgen_transfers.items():
@@ -321,28 +208,3 @@ class TGenVisualization(Visualization):
sns.despine()
self.page.savefig()
plt.close()
-
-# helper - compute the window_size moving average over the data in interval
-def movingaverage(interval, window_size):
- window = numpy.ones(int(window_size)) / float(window_size)
- return numpy.convolve(interval, window, 'same')
-
-# # helper - cumulative fraction for y axis
-def cf(d): return pylab.arange(1.0, float(len(d)) + 1.0) / float(len(d))
-
-# # helper - return step-based CDF x and y values
-# # only show to the 99th percentile by default
-def getcdf(data, shownpercentile=0.99, maxpoints=10000.0):
- data.sort()
- frac = cf(data)
- k = len(data) / maxpoints
- x, y, lasty = [], [], 0.0
- for i in range(int(round(len(data) * shownpercentile))):
- if i % k > 1.0: continue
- assert not numpy.isnan(data[i])
- x.append(data[i])
- y.append(lasty)
- x.append(data[i])
- y.append(frac[i])
- lasty = frac[i]
- return x, y
More information about the tor-commits
mailing list