-
Notifications
You must be signed in to change notification settings - Fork 0
/
text-collection.html
233 lines (232 loc) · 5.82 KB
/
text-collection.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
<!DOCTYPE HTML>
<html>
<head>
<meta charset="utf-8"/>
<meta name="viewport" content="width=device-width, initial-scale = 1, maximum-scale=1, user-scalable=no"/>
<!-- TODO: Meta / OpenGraph / Facebook -->
<!-- TODO: read Pandoc's meta --><title>tudocomp</title>
<link rel="stylesheet" type="text/css" href="font-awesome-4.6.3/css/font-awesome.min.css"/>
<link rel="stylesheet" type="text/css" href="fonts/fonts.css"/>
<link rel="stylesheet" type="text/css" href="style.css"/>
<link rel="icon" type="image/svg+xml" href="tudocomp-icon.svg">
<!-- Chrome currently doesn't support SVG favicons -->
<link rel="icon" type="image/png" href="tudocomp-icon.png">
</head>
<body>
<header>
<h1>– The TU Dortmund Compression Framework</h1>
</header>
<section id="content">
<div id="page"><h1 id="text-collection">Text Collection</h1>
<p>For testing and benchmarking purposes, we have compiled a text collection that is described on this page.</p>
<p><em>tudocomp</em> provides a utility to automatically download or generate this collection. For convenience, it also creates several prefixes of the included texts (200 MiB, 100 MiB 50 MiB, ...).</p>
<p>The collection includes the <a href="http://acube.di.unipi.it/tagme-dataset"><code>tagme</code></a> and <a href="http://acube.di.unipi.it/hashtag-datasets"><code>hashtag</code></a> datasets from the <a href="http://acube.di.unipi.it/datasets">Lab of Advanced Algorithms and Applications</a>, as well as the <a href="pizzachili.dcc.uchile.cl">Pizza&Chili Corpus</a> (identified by the prefixes <code>pc</code> and <code>pcr</code>).</p>
<p>The following collections are generated by <em>tudocomp</em> itself:</p>
<ul>
<li><code>commoncrawl</code> - composition of a random subset of <a href="http://commoncrawl.org">common crawl</a>, containing only the plain text (i.e., excluding header and markup) of websites in ASCII format.</li>
<li><code>wiki-all-vital</code> - plain text composition of about 10,000 of the <a href="https://en.wikipedia.org/wiki/Wikipedia:Vital_articles/Expanded">most vital Wikipedia articles</a>.</li>
</ul>
<p>The following table contains information relevant for text compression on each of the included texts:</p>
<table>
<thead>
<tr class="header">
<th>Text</th>
<th>σ</th>
<th>max lcp</th>
<th>avg lcp</th>
<th>bwt runs</th>
<th>z</th>
<th>max LZ77 factor</th>
<th>H<sub>0</sub></th>
<th>H<sub>1</sub></th>
<th>H<sub>2</sub></th>
<th>H<sub>3</sub></th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td><code>hashtag</code></td>
<td>179</td>
<td>54,075</td>
<td>84</td>
<td>63,014K</td>
<td>13,721K</td>
<td>54,056</td>
<td>4.59</td>
<td>3.06</td>
<td>2.69</td>
<td>2.46</td>
</tr>
<tr class="even">
<td><code>pc-dblp.xml</code></td>
<td>97</td>
<td>1,084</td>
<td>44</td>
<td>29,585K</td>
<td>7,035K</td>
<td>1,060</td>
<td>5.26</td>
<td>3.48</td>
<td>2.17</td>
<td>1.43</td>
</tr>
<tr class="odd">
<td><code>pc-dna</code></td>
<td>17</td>
<td>97,979</td>
<td>60</td>
<td>128,863K</td>
<td>13,970K</td>
<td>97,966</td>
<td>1.97</td>
<td>1.93</td>
<td>1.92</td>
<td>1.92</td>
</tr>
<tr class="even">
<td><code>pc-english</code></td>
<td>226</td>
<td>987,770</td>
<td>9390</td>
<td>72,032K</td>
<td>13,971K</td>
<td>987,766</td>
<td>4.52</td>
<td>3.62</td>
<td>2.95</td>
<td>2.42</td>
</tr>
<tr class="odd">
<td><code>pc-proteins</code></td>
<td>26</td>
<td>45,704</td>
<td>278</td>
<td>108,459K</td>
<td>20,875K</td>
<td>45,703</td>
<td>4.20</td>
<td>4.18</td>
<td>4.16</td>
<td>4.07</td>
</tr>
<tr class="even">
<td><code>pcr-cere</code></td>
<td>6</td>
<td>175,655</td>
<td>3541</td>
<td>10,422K</td>
<td>1,447K</td>
<td>175,643</td>
<td>2.19</td>
<td>1.81</td>
<td>1.81</td>
<td>1.80</td>
</tr>
<tr class="odd">
<td><code>pcr-einstein.en</code></td>
<td>125</td>
<td>935,920</td>
<td>45983</td>
<td>153K</td>
<td>496K</td>
<td>906,995</td>
<td>4.92</td>
<td>3.66</td>
<td>2.61</td>
<td>1.63</td>
</tr>
<tr class="even">
<td><code>pcr-kernel</code></td>
<td>161</td>
<td>2,755,550</td>
<td>149872</td>
<td>2,718K</td>
<td>775K</td>
<td>2,755,550</td>
<td>5.38</td>
<td>4.03</td>
<td>2.93</td>
<td>2.05</td>
</tr>
<tr class="odd">
<td><code>pcr-para</code></td>
<td>6</td>
<td>72,544</td>
<td>2268</td>
<td>13,576K</td>
<td>1,927K</td>
<td>70,680</td>
<td>2.12</td>
<td>1.88</td>
<td>1.88</td>
<td>1.87</td>
</tr>
<tr class="even">
<td><code>pc-sources</code></td>
<td>231</td>
<td>307,871</td>
<td>373</td>
<td>47,651K</td>
<td>11,542K</td>
<td>307,871</td>
<td>5.47</td>
<td>4.08</td>
<td>3.10</td>
<td>2.34</td>
</tr>
<tr class="odd">
<td><code>tagme</code></td>
<td>206</td>
<td>1,281</td>
<td>26</td>
<td>65,195K</td>
<td>13,841K</td>
<td>1,279</td>
<td>4.90</td>
<td>3.77</td>
<td>3.20</td>
<td>2.60</td>
</tr>
<tr class="even">
<td><code>wiki-all-vital</code></td>
<td>205</td>
<td>8,607</td>
<td>15</td>
<td>80,609K</td>
<td>16,274K</td>
<td>8,607</td>
<td>4.56</td>
<td>3.62</td>
<td>3.03</td>
<td>2.45</td>
</tr>
<tr class="odd">
<td><code>commoncrawl</code></td>
<td>115</td>
<td>246,266</td>
<td>1727</td>
<td>45,899K</td>
<td>10,791K</td>
<td>246,266</td>
<td>5.37</td>
<td>4.30</td>
<td>3.55</td>
<td>2.78</td>
</tr>
</tbody>
</table>
<p><em>Legend</em>:</p>
<ul>
<li><em>σ</em> is the alphabet size, i.e., the amount of distinct symbols occuring in the text.</li>
<li><em>max lcp</em> is the maximum value in the text's longest common prefix array.</li>
<li><em>avg lcp</em> is the average of all longest common prefix array values.</li>
<li><em>bwt runs</em> is the number of runs in the Burrows-Wheeler transform of the text (i.e., how often <code>BWT[i] != BWT[i-1]</code> occurs).</li>
<li><em>z</em> is the number of LZ77 factors (n window) for the text.</li>
<li><em>max LZ77 factor</em> is the length of the longest LZ77 factor (no window).</li>
<li><em>H<sub>k</sub></em> is the k-th order entropy of the text.</li>
</ul></div>
</section>
<footer>
</footer>
</body>
</html>