creation_analysis.md

   1 ---
   2 jupyter:
   3   jupytext:
   4     formats: ipynb,md
   5     text_representation:
   6       extension: .md
   7       format_name: markdown
   8       format_version: '1.3'
   9       jupytext_version: 1.14.5
  10   kernelspec:
  11     display_name: Python 3 (ipykernel)
  12     language: python
  13     name: python3
  14 ---
  15
  16 ```python
  17 import pandas as pd
  18 import matplotlib as mpl
  19 import matplotlib.pyplot as plt
  20 %matplotlib inline
  21 import pstats
  22 from pstats import SortKey
  23
  24 from riddle_definitions import *
  25 import random
  26 ```
  27
  28 ```python
  29 distances = [edit_distance(random.choice(dictionary),
  30                            random.choice(dictionary))
  31              for _ in range(10000)]
  32 distances = pd.Series(distances)
  33 distances.describe()
  34 ```
  35
  36 ```python
  37 distances[distances <= 3].count() / distances.count()
  38 ```
  39
  40 ```python
  41 metrics_original = pd.read_csv('metrics_original.csv')
  42 metrics_original
  43 ```
  44
  45 ```python
  46 metrics_related = pd.read_csv('metrics_related.csv')
  47 metrics_related
  48 ```
  49
  50 ```python
  51 metrics_lazy = pd.read_csv('metrics_lazy.csv')
  52 metrics_lazy
  53 ```
  54
  55 ```python
  56 metrics_original.describe()
  57 ```
  58
  59 ```python
  60 metrics_related.describe()
  61 ```
  62
  63 ```python
  64 metrics_lazy.describe()
  65 ```
  66
  67 ```python
  68 fig, ax = plt.subplots(1, 1)
  69
  70 metrics_original.cpu_time.plot.hist(bins=20, ax=ax, alpha=0.5)
  71 metrics_related.cpu_time.plot.hist(bins=20, ax=ax, alpha=0.5)
  72 metrics_lazy.cpu_time.plot.hist(bins=20, ax=ax, alpha=0.5);
  73 ```
  74
  75 ```python
  76 ax = metrics_original.cpu_time.plot.hist(bins=20, alpha=0.5)
  77 metrics_related.cpu_time.plot.hist(bins=20, ax=ax, alpha=0.5)
  78 metrics_lazy.cpu_time.plot.hist(bins=20, ax=ax, alpha=0.5);
  79 ```
  80
  81 ```python
  82 ax = metrics_original.cpu_time.plot.kde(xlim=(0, 10))
  83 metrics_related.cpu_time.plot.kde(ax=ax)
  84 metrics_lazy.cpu_time.plot.kde(ax=ax);
  85 ```
  86
  87 ```python
  88 fig, ax = plt.subplots()
  89 ax.hist([metrics_original.cpu_time], bins=20)
  90 ax.legend(['Original', 'Related', 'Related 2'])
  91 ax.set_title('Times taken to generate a riddle')
  92 ax.set_xlabel('Time (s)')
  93 plt.savefig('original_time_histogram.png')
  94 ```
  95
  96 ```python
  97 fig, ax = plt.subplots()
  98 ax.hist([metrics_original.cpu_time,
  99          metrics_related.cpu_time], bins=20)
 100 ax.legend(['Original', 'Related'])
 101 ax.set_title('Times taken to generate a riddle')
 102 ax.set_xlabel('Time (s)')
 103 plt.savefig('original_related_time_histogram.png')
 104 ```
 105
 106 ```python
 107 fig, ax = plt.subplots()
 108 ax.hist([metrics_original.cpu_time,
 109          metrics_related.cpu_time,
 110          metrics_lazy.cpu_time], bins=20)
 111 ax.legend(['Original', 'Related', 'Lazy'])
 112 ax.set_title('Times taken to generate a riddle')
 113 ax.set_xlabel('Time (s)')
 114 plt.savefig('original_related_lazy_time_histogram.png')
 115 ```
 116
 117 ```python
 118 fig, ax = plt.subplots()
 119 ax.hist([metrics_related.cpu_time,
 120          metrics_lazy.cpu_time], bins=20,
 121         color=['#ff7f0e', '#2ca02c'])
 122 ax.legend(['Related', 'Lazy'])
 123 ax.set_title('Times taken to generate a riddle')
 124 ax.set_xlabel('Time (s)')
 125 plt.savefig('related_lazy_time_histogram.png')
 126 ```
 127
 128 ```python
 129 metrics_original['generated_per_line'] = metrics_original.generated_lines / metrics_original.riddle_lines
 130 metrics_original['time_per_line'] = metrics_original.cpu_time / metrics_original.riddle_lines
 131 metrics_original
 132 ```
 133
 134 ```python
 135 metrics_related['generated_per_line'] = metrics_related.generated_lines / metrics_related.riddle_lines
 136 metrics_related['time_per_line'] = metrics_related.cpu_time / metrics_related.riddle_lines
 137 metrics_related
 138 ```
 139
 140 ```python
 141 metrics_lazy['generated_per_line'] = metrics_lazy.generated_lines / metrics_lazy.riddle_lines
 142 metrics_lazy['time_per_line'] = metrics_lazy.cpu_time / metrics_lazy.riddle_lines
 143 metrics_lazy
 144 ```
 145
 146 ```python
 147 fig, ax = plt.subplots()
 148 ax.hist([metrics_original.time_per_line], bins=20)
 149 ax.legend(['Original']);
 150 ```
 151
 152 ```python
 153 fig, ax = plt.subplots()
 154 ax.hist([metrics_original.time_per_line,
 155          metrics_related.time_per_line], bins=20)
 156 ax.legend(['Original', 'Related']);
 157 ```
 158
 159 ```python
 160 fig, ax = plt.subplots()
 161 ax.hist([metrics_original.time_per_line,
 162          metrics_related.time_per_line,
 163          metrics_lazy.time_per_line], bins=20)
 164 ax.legend(['Original', 'Related', 'Lazy'])
 165 ```
 166
 167 ```python
 168 plt.rcParams['axes.prop_cycle'].by_key()['color']
 169 ```
 170
 171 ```python
 172 for bars, column in zip(*ax.get_legend_handles_labels()):
 173   color = bars[0].get_facecolor()
 174   print(column, color)
 175 ```
 176
 177 ```python
 178 fig, ax = plt.subplots()
 179 ax.hist([metrics_related.time_per_line,
 180          metrics_lazy.time_per_line], bins=20,
 181        color=['#ff7f0e', '#2ca02c'])
 182 ax.legend(['Related', 'Lazy']);
 183 ```
 184
 185 ```python
 186 ax = metrics_original.time_per_line.plot.kde(xlim=(0, 2))
 187 metrics_related.time_per_line.plot.kde(ax=ax)
 188 metrics_lazy.time_per_line.plot.kde(ax=ax);
 189 ```
 190
 191 ```python
 192 metrics_original.describe()
 193 ```
 194
 195 ```python
 196 metrics_related.describe()
 197 ```
 198
 199 ```python
 200 metrics_lazy.describe()
 201 ```
 202
 203 ```python
 204 metrics_original.time_per_line.mean() / metrics_related.time_per_line.mean()
 205 ```
 206
 207 ```python
 208 metrics_original.time_per_line.median() / metrics_related.time_per_line.median()
 209 ```
 210
 211 ```python
 212 metrics_related.time_per_line.mean() / metrics_lazy.time_per_line.mean()
 213 ```
 214
 215 ```python
 216 metrics_related.time_per_line.median() / metrics_lazy.time_per_line.median()
 217 ```
 218
 219 ```python
 220 metrics_original.time_per_line.mean() / metrics_lazy.time_per_line.mean()
 221 ```
 222
 223 ```python
 224 metrics_original.time_per_line.median() / metrics_lazy.time_per_line.median()
 225 ```
 226
 227 ```python
 228 metrics_original.wall_time.mean() / metrics_related.wall_time.mean()
 229 ```
 230
 231 ```python
 232 metrics_related.wall_time.mean() / metrics_lazy.wall_time.mean()
 233 ```
 234
 235 ```python
 236 metrics_original.wall_time.mean() / metrics_lazy.wall_time.mean()
 237 ```
 238
 239 ```python
 240 stats = pstats.Stats('filtered.stats')
 241 stats.strip_dirs().sort_stats(SortKey.TIME).print_stats(10)
 242 ```
 243
 244 ```python
 245 stats2 = pstats.Stats('lazy.stats')
 246 stats2.strip_dirs().sort_stats(SortKey.TIME).print_stats(10)
 247 ```
 248
 249 ```python
 250
 251 ```