scripts/lttng-baremetal-tests/generate-plots.py

   1 # Copyright (C) 2017 - Francis Deslauriers <francis.deslauriers@efficios.com>
   2 #
   3 # This program is free software: you can redistribute it and/or modify
   4 # it under the terms of the GNU General Public License as published by
   5 # the Free Software Foundation, either version 3 of the License, or
   6 # (at your option) any later version.
   7 #
   8 # This program is distributed in the hope that it will be useful,
   9 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  10 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  11 # GNU General Public License for more details.
  12 #
  13 # You should have received a copy of the GNU General Public License
  14 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
  15
  16
  17 import os, sys
  18 import numpy as np
  19 import pandas as pd
  20
  21 #Set Matplotlib to use the PNG non interactive backend
  22 import matplotlib as mpl
  23 mpl.use('Agg')
  24
  25 import matplotlib.pyplot as plt
  26 from matplotlib.ticker import MaxNLocator
  27 from cycler import cycler
  28 from collections import OrderedDict
  29
  30 def rename_cols(df):
  31     new_cols = {'baseline_1thr_peritermean': 'basel_1thr',
  32             'baseline_2thr_peritermean': 'basel_2thr',
  33             'baseline_4thr_peritermean': 'basel_4thr',
  34             'baseline_8thr_peritermean': 'basel_8thr',
  35             'baseline_16thr_peritermean': 'basel_16thr',
  36             'lttng_1thr_peritermean': 'lttng_1thr',
  37             'lttng_2thr_peritermean': 'lttng_2thr',
  38             'lttng_4thr_peritermean': 'lttng_4thr',
  39             'lttng_8thr_peritermean': 'lttng_8thr',
  40             'lttng_16thr_peritermean': 'lttng_16thr',
  41             'baseline_1thr_periterstdev': 'basel_1thr_stdev',
  42             'baseline_2thr_periterstdev': 'basel_2thr_stdev',
  43             'baseline_4thr_periterstdev': 'basel_4thr_stdev',
  44             'baseline_8thr_periterstdev': 'basel_8thr_stdev',
  45             'baseline_16thr_periterstdev': 'basel_16thr_stdev',
  46             'lttng_1thr_periterstdev': 'lttng_1thr_stdev',
  47             'lttng_2thr_periterstdev': 'lttng_2thr_stdev',
  48             'lttng_4thr_periterstdev': 'lttng_4thr_stdev',
  49             'lttng_8thr_periterstdev': 'lttng_8thr_stdev',
  50             'lttng_16thr_periterstdev': 'lttng_16thr_stdev'
  51             }
  52     df.rename(columns=new_cols, inplace=True)
  53     return df
  54
  55 def convert_us_to_ns(df):
  56     cols = [col for col in df.columns if 'periter' in col]
  57     df[cols] = df[cols].apply(lambda x: x*1000)
  58     return df
  59
  60 def create_plot(df, graph_type):
  61     # We map all test configurations and their
  62     # respective color
  63     conf_to_color = OrderedDict([
  64         ('basel_1thr','lightcoral'),
  65         ('lttng_1thr','red'),
  66         ('basel_2thr','gray'),
  67         ('lttng_2thr','black'),
  68         ('basel_4thr','chartreuse'),
  69         ('lttng_4thr','forestgreen'),
  70         ('basel_8thr','deepskyblue'),
  71         ('lttng_8thr','mediumblue'),
  72         ('basel_16thr','orange'),
  73         ('lttng_16thr','saddlebrown')])
  74
  75     # We create a list for each of the subplots
  76     baseline = [x for x in conf_to_color.keys() if 'basel' in x]
  77     lttng = [x for x in conf_to_color.keys() if 'lttng' in x]
  78     one_thr = [x for x in conf_to_color.keys() if '_1thr' in x]
  79     two_thr = [x for x in conf_to_color.keys() if '_2thr' in x]
  80     four_thr = [x for x in conf_to_color.keys() if '_4thr' in x]
  81     eight_thr = [x for x in conf_to_color.keys() if '_8thr' in x]
  82     sixteen_thr = [x for x in conf_to_color.keys() if '_16thr' in x]
  83
  84     plots = [baseline, lttng, one_thr, two_thr, four_thr, eight_thr, sixteen_thr]
  85
  86     title='Meantime per syscalls for {} testcase'.format(graph_type)
  87
  88     # Create a axe object for each sub-plots
  89     f, arrax = plt.subplots(len(plots), sharex=True, figsize=(16, 25))
  90     f.suptitle(title, fontsize=20)
  91
  92     for (ax, data_cols)  in zip(arrax, plots):
  93         curr_df = df[data_cols]
  94
  95         stdev_cols = ['{}_stdev'.format(x) for x in data_cols]
  96         # Extract the color for each configuration
  97         colors = [conf_to_color[x] for x in data_cols]
  98
  99         # set the color cycler for this plot
 100         ax.set_prop_cycle(cycler('color', colors))
 101
 102         # Plot each line and its errorbars
 103         for (data, stdev) in  zip(data_cols, stdev_cols):
 104             ax.errorbar(x=df.index.values, y=df[data], yerr=df[stdev], marker='o')
 105
 106         ax.set_ylim(0)
 107         ax.grid()
 108         ax.set_xlabel('Jenkins Build ID')
 109         ax.set_ylabel('Meantime per syscall [us]')
 110
 111         ax.xaxis.set_major_locator(MaxNLocator(integer=True))
 112
 113         ax.legend(prop={'family': 'monospace'},
 114                     labels=curr_df.columns.values, bbox_to_anchor=(1.2,1))
 115
 116     plt.subplots_adjust(top=0.95)
 117     plt.savefig('{}.png'.format(graph_type), bbox_inches='tight')
 118
 119 # Writes a file that contains commit id of all configurations shown in the
 120 # plots
 121 def create_metadata_file(res_dir):
 122     list_ = []
 123     for dirname, dirnames, res_files in os.walk('./'+res_dir):
 124         if len(dirnames) > 0:
 125             continue
 126         metadata = pd.read_csv(os.path.join(dirname, 'metadata.csv'))
 127         list_.append(metadata)
 128
 129     df = pd.concat(list_)
 130     df.index=df.build_id
 131     df.sort_index(inplace=True)
 132     df.to_csv('metadata.csv', index=False)
 133
 134 #Iterates over a result directory and creates the plots for the different
 135 #testcases
 136 def create_plots(res_dir):
 137     df = pd.DataFrame()
 138     metadata_df = pd.DataFrame()
 139     list_ = []
 140     for dirname, dirnames, res_files in os.walk('./'+res_dir):
 141         if len(dirnames) > 0:
 142             continue
 143         metadata = pd.read_csv(os.path.join(dirname, 'metadata.csv'))
 144
 145         for res in res_files:
 146             if res in 'metadata.csv':
 147                 continue
 148             tmp = pd.read_csv(os.path.join(dirname, res))
 149             #Use the build id as the index for the dataframe for filtering
 150             tmp.index = metadata.build_id
 151             #Add the testcase name to the row for later filtering
 152             tmp['testcase'] = res.split('.')[0]
 153             list_.append(tmp)
 154
 155         df = pd.concat(list_)
 156         df = convert_us_to_ns(df)
 157         df = rename_cols(df)
 158         df.sort_index(inplace=True)
 159
 160     #Go over the entire dataframe by testcase and create a plot for each type
 161     for testcase in df.testcase.unique():
 162         df_testcase  = df.loc[df['testcase'] == testcase]
 163         create_plot(df=df_testcase, graph_type=testcase)
 164
 165 def main():
 166     res_path = sys.argv[1]
 167     create_plots(os.path.join(res_path))
 168     create_metadata_file(os.path.join(res_path))
 169
 170 if __name__ == '__main__':
 171     main()