siteseeker_normalizer 0.0.6
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +7 -0
- data/.gitignore +14 -0
- data/.rspec +1 -0
- data/COPYING +661 -0
- data/Gemfile +4 -0
- data/README.md +46 -0
- data/Rakefile +6 -0
- data/lib/siteseeker_normalizer.rb +5 -0
- data/lib/siteseeker_normalizer/client.rb +26 -0
- data/lib/siteseeker_normalizer/parse.rb +89 -0
- data/lib/siteseeker_normalizer/parse/category.rb +26 -0
- data/lib/siteseeker_normalizer/parse/entry.rb +44 -0
- data/lib/siteseeker_normalizer/version.rb +3 -0
- data/siteseeker_normalizer.gemspec +24 -0
- data/spec/fixtures/barn.html +1885 -0
- data/spec/fixtures/barnomsrg.html +347 -0
- data/spec/siteseeeker_normalizer_spec.rb +69 -0
- data/spec/spec_helper.rb +2 -0
- metadata +122 -0
|
@@ -0,0 +1,347 @@
|
|
|
1
|
+
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
|
|
2
|
+
<html>
|
|
3
|
+
<head>
|
|
4
|
+
<meta http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">
|
|
5
|
+
<meta name="ROBOTS" content="NOINDEX,NOFOLLOW" />
|
|
6
|
+
<title>Malmö stad: sök ”barnomsrg”</title>
|
|
7
|
+
|
|
8
|
+
<link rel="stylesheet" href="http://foo.appliance.siteseeker.se/search/bar/?p=ext&theme=modular&doc=style.css&i=sv" type="text/css" >
|
|
9
|
+
<link rel="stylesheet" href="http://foo.appliance.siteseeker.se/search/bar/?p=ext&theme=modular&doc=jquery.autocomplete.css&i=sv" type="text/css" >
|
|
10
|
+
<script type="text/javascript" src="http://foo.appliance.siteseeker.se/search/bar/?p=ext&theme=modular&doc=jquery-1.3.2.min.js&i=sv"></script>
|
|
11
|
+
<script type="text/javascript" src="http://foo.appliance.siteseeker.se/search/bar/?p=ext&theme=modular&doc=jquery.autocomplete.1.2.1-siteseeker.js&i=sv"></script>
|
|
12
|
+
<script type="text/javascript" src="http://foo.appliance.siteseeker.se/search/bar/?p=ext&theme=modular&doc=script.js&i=sv"></script>
|
|
13
|
+
<script type="text/javascript">var enableQueryCompletion = true; var searchFieldId = 'essi-queryfield'; var language = 'sv'; var qcUrl = 'http://foo.appliance.siteseeker.se/qc/bar/'; var searchUrl = 'http://foo.appliance.siteseeker.se/search/bar/';</script>
|
|
14
|
+
|
|
15
|
+
<meta name="keywords" content="">
|
|
16
|
+
<meta name="description" content="Malm� stads officiella webbplats. Official website of the city of Malm�. V�lkommen till Malm� stad! Welcome to the City of Malm� in Sweden!">
|
|
17
|
+
<meta http-equiv="imagetoolbar" content="no">
|
|
18
|
+
<link rel="schema.DC" href="http://purl.org/DC/elements/1.0">
|
|
19
|
+
</head>
|
|
20
|
+
<body>
|
|
21
|
+
<table cellspacing="5">
|
|
22
|
+
<tr>
|
|
23
|
+
<td valign="top" width="250">
|
|
24
|
+
<form id="essi-mainform" method="get"
|
|
25
|
+
action="http://foo.appliance.siteseeker.se/search/bar/">
|
|
26
|
+
|
|
27
|
+
<div id="essi-query-block">
|
|
28
|
+
<label for="essi-queryfield"><strong>Sök efter:</strong></label>
|
|
29
|
+
<input type="text" name="q" value="barnomsrg" tabindex="1" id="essi-queryfield" autocomplete="off" >
|
|
30
|
+
|
|
31
|
+
|
|
32
|
+
<input type="submit" name="x" tabindex="2" id="essi-search-button"
|
|
33
|
+
value="Hitta!" >
|
|
34
|
+
|
|
35
|
+
|
|
36
|
+
|
|
37
|
+
|
|
38
|
+
|
|
39
|
+
|
|
40
|
+
|
|
41
|
+
<a href="http://foo.appliance.siteseeker.se/search/bar/?q=barnomsrg&t=simple&ls=2&d=0&d1=01&d2=1&d3=1970&d4=26&d5=11&d6=2013&s=0&so=1&h=0&hn=20&hd=1&i=sv&p=helppopup&b=1&c=0&t=s&l=0&ll=-2&f=0&ff=0&oenc=UTF-8&ua=f111c901872e2ee9f3fea7552d9dde4d" tabindex="3" onclick="essOpenWindow('http://foo.appliance.siteseeker.se/search/bar/?q=barnomsrg&t=simple&ls=2&d=0&d1=01&d2=1&d3=1970&d4=26&d5=11&d6=2013&s=0&so=1&h=0&hn=20&hd=1&i=sv&p=helppopup&b=1&c=0&t=s&l=0&ll=-2&f=0&ff=0&oenc=UTF-8&ua=f111c901872e2ee9f3fea7552d9dde4d', 'siteseekerHelpPopupWindow', false, 500, 500); return false;"
|
|
42
|
+
id="essi-helplink" title="Sökhjälp (öppnas i nytt fönster)">Hjälp</a>
|
|
43
|
+
|
|
44
|
+
|
|
45
|
+
|
|
46
|
+
</div>
|
|
47
|
+
|
|
48
|
+
<div id="essi-options">
|
|
49
|
+
<input type="hidden" name="i" value="sv" >
|
|
50
|
+
<input type="hidden" name="s" value="1" >
|
|
51
|
+
<input type="hidden" name="so" value="1" >
|
|
52
|
+
<input type="hidden" name="ua" value="f111c901872e2ee9f3fea7552d9dde4d" >
|
|
53
|
+
<input type="hidden" name="charset" value="UTF-8" >
|
|
54
|
+
<input type="hidden" name="oenc" value="UTF-8" >
|
|
55
|
+
<input type="hidden" name="origin" value="" >
|
|
56
|
+
|
|
57
|
+
<!-- CATEGORY -->
|
|
58
|
+
|
|
59
|
+
<div class="ess-group" id="essi-opt-category">
|
|
60
|
+
<p><input type="radio" name="c" value="0" id="essi-catall" checked="checked" >
|
|
61
|
+
<strong><label for="essi-catall">Sök dokument på hela webbplatsen</label></strong></p>
|
|
62
|
+
<p><input type="radio" name="c" value="1" id="essi-catselect" >
|
|
63
|
+
<label for="essi-catselect">Avgränsa till:</label></p>
|
|
64
|
+
<div class="ess-option-block">
|
|
65
|
+
<div>
|
|
66
|
+
|
|
67
|
+
</div>
|
|
68
|
+
<div class="ess-cat-group-with-cats" id="essi-cg-medborgare">
|
|
69
|
+
<p class="ess-cat-group">Medborgare:</p>
|
|
70
|
+
|
|
71
|
+
<p>
|
|
72
|
+
<input type="checkbox" id="essi-cc-119" name="cc[]"
|
|
73
|
+
value="119" onclick="essEnableCats(this)" >
|
|
74
|
+
<label for="essi-cc-119">Kultur & nöje</label>
|
|
75
|
+
</p>
|
|
76
|
+
|
|
77
|
+
<p>
|
|
78
|
+
<input type="checkbox" id="essi-cc-116" name="cc[]"
|
|
79
|
+
value="116" onclick="essEnableCats(this)" >
|
|
80
|
+
<label for="essi-cc-116">Social- & familjefrågor</label>
|
|
81
|
+
</p>
|
|
82
|
+
|
|
83
|
+
<p>
|
|
84
|
+
<input type="checkbox" id="essi-cc-125" name="cc[]"
|
|
85
|
+
value="125" onclick="essEnableCats(this)" >
|
|
86
|
+
<label for="essi-cc-125">Biblioteken</label>
|
|
87
|
+
</p>
|
|
88
|
+
|
|
89
|
+
<p>
|
|
90
|
+
<input type="checkbox" id="essi-cc-121" name="cc[]"
|
|
91
|
+
value="121" onclick="essEnableCats(this)" >
|
|
92
|
+
<label for="essi-cc-121">Jobb & praktik</label>
|
|
93
|
+
</p>
|
|
94
|
+
|
|
95
|
+
<p>
|
|
96
|
+
<input type="checkbox" id="essi-cc-123" name="cc[]"
|
|
97
|
+
value="123" onclick="essEnableCats(this)" >
|
|
98
|
+
<label for="essi-cc-123">Förskola & utbildning</label>
|
|
99
|
+
</p>
|
|
100
|
+
|
|
101
|
+
<p>
|
|
102
|
+
<input type="checkbox" id="essi-cc-115" name="cc[]"
|
|
103
|
+
value="115" onclick="essEnableCats(this)" >
|
|
104
|
+
<label for="essi-cc-115">Stadsplanering & trafik</label>
|
|
105
|
+
</p>
|
|
106
|
+
|
|
107
|
+
<p>
|
|
108
|
+
<input type="checkbox" id="essi-cc-118" name="cc[]"
|
|
109
|
+
value="118" onclick="essEnableCats(this)" >
|
|
110
|
+
<label for="essi-cc-118">Miljö & hållbarhet</label>
|
|
111
|
+
</p>
|
|
112
|
+
|
|
113
|
+
<p>
|
|
114
|
+
<input type="checkbox" id="essi-cc-117" name="cc[]"
|
|
115
|
+
value="117" onclick="essEnableCats(this)" >
|
|
116
|
+
<label for="essi-cc-117">Omsorg, vård & stöd</label>
|
|
117
|
+
</p>
|
|
118
|
+
|
|
119
|
+
<p>
|
|
120
|
+
<input type="checkbox" id="essi-cc-124" name="cc[]"
|
|
121
|
+
value="124" onclick="essEnableCats(this)" >
|
|
122
|
+
<label for="essi-cc-124">Bo & bygga</label>
|
|
123
|
+
</p>
|
|
124
|
+
|
|
125
|
+
<p>
|
|
126
|
+
<input type="checkbox" id="essi-cc-122" name="cc[]"
|
|
127
|
+
value="122" onclick="essEnableCats(this)" >
|
|
128
|
+
<label for="essi-cc-122">Idrott & fritid</label>
|
|
129
|
+
</p>
|
|
130
|
+
|
|
131
|
+
</div>
|
|
132
|
+
<div class="ess-cat-group-with-cats" id="essi-cg-vriga_delgrenar">
|
|
133
|
+
<p class="ess-cat-group">Övriga delgrenar:</p>
|
|
134
|
+
|
|
135
|
+
<p>
|
|
136
|
+
<input type="checkbox" id="essi-cc-114" name="cc[]"
|
|
137
|
+
value="114" onclick="essEnableCats(this)" >
|
|
138
|
+
<label for="essi-cc-114">Företagare</label>
|
|
139
|
+
</p>
|
|
140
|
+
|
|
141
|
+
<p>
|
|
142
|
+
<input type="checkbox" id="essi-cc-112" name="cc[]"
|
|
143
|
+
value="112" onclick="essEnableCats(this)" >
|
|
144
|
+
<label for="essi-cc-112">Kommun & politik</label>
|
|
145
|
+
</p>
|
|
146
|
+
|
|
147
|
+
<p>
|
|
148
|
+
<input type="checkbox" id="essi-cc-113" name="cc[]"
|
|
149
|
+
value="113" onclick="essEnableCats(this)" >
|
|
150
|
+
<label for="essi-cc-113">Turist</label>
|
|
151
|
+
</p>
|
|
152
|
+
|
|
153
|
+
</div>
|
|
154
|
+
<div class="ess-cat-group-with-cats" id="essi-cg-kategori">
|
|
155
|
+
<p class="ess-cat-group">Kategori:</p>
|
|
156
|
+
|
|
157
|
+
<p>
|
|
158
|
+
<input type="checkbox" id="essi-cc-0" name="cc[]"
|
|
159
|
+
value="0" onclick="essEnableCats(this)" >
|
|
160
|
+
<label for="essi-cc-0">Övriga</label>
|
|
161
|
+
</p>
|
|
162
|
+
|
|
163
|
+
<p>
|
|
164
|
+
<input type="checkbox" id="essi-cc-110" name="cc[]"
|
|
165
|
+
value="110" onclick="essEnableCats(this)" >
|
|
166
|
+
<label for="essi-cc-110">Miljöbarometern</label>
|
|
167
|
+
</p>
|
|
168
|
+
|
|
169
|
+
|
|
170
|
+
</div>
|
|
171
|
+
</div>
|
|
172
|
+
<p class="ess-clear"></p>
|
|
173
|
+
</div>
|
|
174
|
+
|
|
175
|
+
|
|
176
|
+
<!-- FORMAT -->
|
|
177
|
+
|
|
178
|
+
<div class="ess-group" id="essi-opt-format">
|
|
179
|
+
<p><input type="radio" name="f" value="0" checked="checked" id="essi-filter-doctype-any" >
|
|
180
|
+
<strong><label for="essi-filter-doctype-any">Sök dokument av alla typer</label></strong></p>
|
|
181
|
+
|
|
182
|
+
|
|
183
|
+
<p><input type="radio" name="f" value="1" id="essi-filter-doctype" >
|
|
184
|
+
<label for="essi-filter-doctype">Endast detta format</label>:</p>
|
|
185
|
+
<p class="ess-option-block">
|
|
186
|
+
<select name="ff" onfocus="essEnableFilter('essi-filter-doctype')" title="Endast detta format"
|
|
187
|
+
id="essi-filter-doctype-sel" onclick="essEnableFilter('essi-filter-doctype')">
|
|
188
|
+
<option value="1">webbsidor</option>
|
|
189
|
+
<option value="3">PDF</option>
|
|
190
|
+
<option value="5">Word</option>
|
|
191
|
+
<option value="6">Excel</option>
|
|
192
|
+
<option value="7">PowerPoint</option>
|
|
193
|
+
|
|
194
|
+
</select>
|
|
195
|
+
</p>
|
|
196
|
+
|
|
197
|
+
|
|
198
|
+
<p><input type="radio" name="f" value="2" id="essi-filter-doctype-image" >
|
|
199
|
+
<label for="essi-filter-doctype-image">Bilder</label></p>
|
|
200
|
+
|
|
201
|
+
</div>
|
|
202
|
+
|
|
203
|
+
|
|
204
|
+
<!-- DATE -->
|
|
205
|
+
|
|
206
|
+
<div class="ess-group" id="essi-opt-date">
|
|
207
|
+
<p><input type="radio" name="da" value="0" checked="checked" id="essi-filter-date-any" >
|
|
208
|
+
<strong><label for="essi-filter-date-any">Sök dokument ändrade när som helst</label></strong></p>
|
|
209
|
+
<p><input type="radio" name="da" value="1" id="essi-filter-date-week" >
|
|
210
|
+
<label for="essi-filter-date-week">Senaste veckan</label></p>
|
|
211
|
+
<p><input type="radio" name="da" value="2" id="essi-filter-date-month" >
|
|
212
|
+
<label for="essi-filter-date-month">Senaste månaden</label></p>
|
|
213
|
+
<p><input type="radio" name="da" value="3" id="essi-filter-date-year" >
|
|
214
|
+
<label for="essi-filter-date-year">Senaste året</label>
|
|
215
|
+
<input type="hidden" name="d" value="1" >
|
|
216
|
+
</p>
|
|
217
|
+
</div>
|
|
218
|
+
|
|
219
|
+
|
|
220
|
+
<!-- LANGUAGE -->
|
|
221
|
+
|
|
222
|
+
<div class="ess-group" id="essi-opt-lang">
|
|
223
|
+
<p><input type="radio" name="l" value="0" checked="checked" id="essi-filter-lang-any" >
|
|
224
|
+
<strong><label for="essi-filter-lang-any">Sök dokument på alla språk</label></strong></p>
|
|
225
|
+
<p><input type="radio" name="l" value="1" id="essi-filter-lang" >
|
|
226
|
+
<label for="essi-filter-lang">Endast på</label>:
|
|
227
|
+
</p>
|
|
228
|
+
<p class="ess-option-block">
|
|
229
|
+
<select name="ll" onfocus="essEnableFilter('essi-filter-lang')" title="Endast på"
|
|
230
|
+
id="essi-filter-lang-sel" onclick="essEnableFilter('essi-filter-lang')">
|
|
231
|
+
<option value="-1">alla språk</option>
|
|
232
|
+
<option value="5">svenska</option>
|
|
233
|
+
<option value="0">danska</option>
|
|
234
|
+
<option value="1">tyska</option>
|
|
235
|
+
<option value="2">engelska</option>
|
|
236
|
+
<option value="3">franska</option>
|
|
237
|
+
<option value="6">spanska</option>
|
|
238
|
+
<option value="11">ryska</option>
|
|
239
|
+
<option value="14">polska</option>
|
|
240
|
+
<option value="15">kroatiska</option>
|
|
241
|
+
<option value="17">turkiska</option>
|
|
242
|
+
<option value="21">rumänska</option>
|
|
243
|
+
<option value="40">albanska</option>
|
|
244
|
+
|
|
245
|
+
</select>
|
|
246
|
+
</p>
|
|
247
|
+
</div>
|
|
248
|
+
|
|
249
|
+
|
|
250
|
+
<!-- STEMMING -->
|
|
251
|
+
|
|
252
|
+
|
|
253
|
+
<div class="ess-group" id="essi-search-button-2-div">
|
|
254
|
+
<input type="submit" name="x" tabindex="2" id="essi-search-button-2"
|
|
255
|
+
value="Hitta!" >
|
|
256
|
+
</div>
|
|
257
|
+
|
|
258
|
+
</div>
|
|
259
|
+
|
|
260
|
+
</form>
|
|
261
|
+
|
|
262
|
+
</td>
|
|
263
|
+
<td valign="top">
|
|
264
|
+
|
|
265
|
+
|
|
266
|
+
|
|
267
|
+
<div class="ess-result">
|
|
268
|
+
|
|
269
|
+
<h2 class="ess-topcell">
|
|
270
|
+
<strong class="ess-header">Resultat:</strong>
|
|
271
|
+
<strong><span id="essi-hitcount">Inga</span></strong> <span id="essi-hitname">träffar</span> på <strong>barnomsrg</strong><span id="essi-wholesite-prep"> inom <strong id="essi-wholesite">Hela Malmö stads webbplats</strong></span>
|
|
272
|
+
</h2>
|
|
273
|
+
|
|
274
|
+
|
|
275
|
+
|
|
276
|
+
|
|
277
|
+
</div>
|
|
278
|
+
|
|
279
|
+
|
|
280
|
+
|
|
281
|
+
|
|
282
|
+
|
|
283
|
+
|
|
284
|
+
|
|
285
|
+
|
|
286
|
+
<div class="ess-nohits">
|
|
287
|
+
<div class="ess-spelling">
|
|
288
|
+
<span class="ess-qmark">?</span>
|
|
289
|
+
|
|
290
|
+
<h3 class="ess-helpheadline">Sökfrågan kan vara felstavad</h3>
|
|
291
|
+
|
|
292
|
+
<p>Menade du <strong><strong>barnomsorg</strong></strong> eller <strong><strong>backgrounds</strong></strong>?</p>
|
|
293
|
+
|
|
294
|
+
<ul>
|
|
295
|
+
|
|
296
|
+
<li>
|
|
297
|
+
Jag vill söka efter
|
|
298
|
+
<a href="http://foo.appliance.siteseeker.se/search/bar/?q=barnomsorg&t=simple&ls=2&d=0&d1=01&d2=1&d3=1970&d4=26&d5=11&d6=2013&s=0&so=1&h=0&hn=20&hd=1&i=sv&sc=click&p=&b=1&c=0&t=s&l=0&ll=-2&f=0&ff=0&oenc=UTF-8&ua=f111c901872e2ee9f3fea7552d9dde4d"><strong>barnomsorg</strong></a>.
|
|
299
|
+
</li>
|
|
300
|
+
|
|
301
|
+
<li>
|
|
302
|
+
Jag vill söka efter
|
|
303
|
+
<a href="http://foo.appliance.siteseeker.se/search/bar/?q=backgrounds&t=simple&ls=2&d=0&d1=01&d2=1&d3=1970&d4=26&d5=11&d6=2013&s=0&so=1&h=0&hn=20&hd=1&i=sv&sc=click&p=&b=1&c=0&t=s&l=0&ll=-2&f=0&ff=0&oenc=UTF-8&ua=f111c901872e2ee9f3fea7552d9dde4d"><strong>backgrounds</strong></a>.
|
|
304
|
+
</li>
|
|
305
|
+
|
|
306
|
+
</ul>
|
|
307
|
+
|
|
308
|
+
</div>
|
|
309
|
+
</div>
|
|
310
|
+
|
|
311
|
+
|
|
312
|
+
|
|
313
|
+
|
|
314
|
+
|
|
315
|
+
<dl class="ess-hits ess-hit">
|
|
316
|
+
|
|
317
|
+
|
|
318
|
+
|
|
319
|
+
|
|
320
|
+
|
|
321
|
+
</dl>
|
|
322
|
+
|
|
323
|
+
|
|
324
|
+
|
|
325
|
+
|
|
326
|
+
|
|
327
|
+
|
|
328
|
+
|
|
329
|
+
|
|
330
|
+
<div id="essi-footer-logo">
|
|
331
|
+
<br >
|
|
332
|
+
<a href="http://www.siteseeker.se/">
|
|
333
|
+
<img src="http://foo.appliance.siteseeker.se/images/modular/poweredbysiteseeker.gif" alt="Sökning levererad av Euroling SiteSeeker"
|
|
334
|
+
style="border: none;" ></a>
|
|
335
|
+
</div>
|
|
336
|
+
|
|
337
|
+
|
|
338
|
+
<script type="text/javascript" src="http://foo.appliance.siteseeker.se/click/bar/?ua=f111c901872e2ee9f3fea7552d9dde4d&pageloading=1"></script>
|
|
339
|
+
|
|
340
|
+
</td>
|
|
341
|
+
<td valign="top">
|
|
342
|
+
|
|
343
|
+
</td>
|
|
344
|
+
</tr>
|
|
345
|
+
</table>
|
|
346
|
+
</body>
|
|
347
|
+
</html>
|
|
@@ -0,0 +1,69 @@
|
|
|
1
|
+
# -*- coding: utf-8 -*-
|
|
2
|
+
require 'spec_helper'
|
|
3
|
+
|
|
4
|
+
describe SiteseekerNormalizer do
|
|
5
|
+
before(:each) do
|
|
6
|
+
raw_results = open("spec/fixtures/barn.html").read
|
|
7
|
+
@results = SiteseekerNormalizer::Parse.new(raw_results, encoding: "UTF-8")
|
|
8
|
+
end
|
|
9
|
+
|
|
10
|
+
it "should have a number of hits" do
|
|
11
|
+
@results.total.should be_a Fixnum
|
|
12
|
+
end
|
|
13
|
+
|
|
14
|
+
it "should have results" do
|
|
15
|
+
@results.entries.count.should > 0
|
|
16
|
+
end
|
|
17
|
+
|
|
18
|
+
describe "result entry" do
|
|
19
|
+
it "should have an order number" do
|
|
20
|
+
@results.entries.first.number.should eq 1
|
|
21
|
+
end
|
|
22
|
+
|
|
23
|
+
it "should have a title" do
|
|
24
|
+
@results.entries.first.title.should be_a String
|
|
25
|
+
end
|
|
26
|
+
|
|
27
|
+
it "should have an extract" do
|
|
28
|
+
@results.entries.first.summary.should be_a String
|
|
29
|
+
end
|
|
30
|
+
|
|
31
|
+
it "should have a breadcrumb" do
|
|
32
|
+
@results.entries.first.breadcrumbs.should be_an Array
|
|
33
|
+
end
|
|
34
|
+
|
|
35
|
+
it "should have a category" do
|
|
36
|
+
@results.entries.first.category.should be_a String
|
|
37
|
+
end
|
|
38
|
+
|
|
39
|
+
it "should have a date string" do
|
|
40
|
+
@results.entries.first.date.should be_a String
|
|
41
|
+
end
|
|
42
|
+
end
|
|
43
|
+
|
|
44
|
+
it "should have sorting" do
|
|
45
|
+
@results.sorting.should be_an Array
|
|
46
|
+
end
|
|
47
|
+
|
|
48
|
+
it "should have a first sorting entry with text" do
|
|
49
|
+
@results.sorting.first.text.should be_a String
|
|
50
|
+
end
|
|
51
|
+
|
|
52
|
+
it "should have a second sorting entry with an url" do
|
|
53
|
+
@results.sorting[1].query.should be_a String
|
|
54
|
+
end
|
|
55
|
+
|
|
56
|
+
it "should have a query string for getting more results" do
|
|
57
|
+
@results.more_query.should be_a String
|
|
58
|
+
end
|
|
59
|
+
|
|
60
|
+
it "should have a categories" do
|
|
61
|
+
@results.category_groups.should be_an Array
|
|
62
|
+
end
|
|
63
|
+
|
|
64
|
+
it "should show a spelling suggestions" do
|
|
65
|
+
raw_results = open("spec/fixtures/barnomsrg.html").read
|
|
66
|
+
results = SiteseekerNormalizer::Parse.new(raw_results, encoding: "UTF-8")
|
|
67
|
+
results.suggestions.count.should > 0
|
|
68
|
+
end
|
|
69
|
+
end
|
data/spec/spec_helper.rb
ADDED
metadata
ADDED
|
@@ -0,0 +1,122 @@
|
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
|
2
|
+
name: siteseeker_normalizer
|
|
3
|
+
version: !ruby/object:Gem::Version
|
|
4
|
+
version: 0.0.6
|
|
5
|
+
platform: ruby
|
|
6
|
+
authors:
|
|
7
|
+
- martent
|
|
8
|
+
autorequire:
|
|
9
|
+
bindir: bin
|
|
10
|
+
cert_chain: []
|
|
11
|
+
date: 2013-11-26 00:00:00.000000000 Z
|
|
12
|
+
dependencies:
|
|
13
|
+
- !ruby/object:Gem::Dependency
|
|
14
|
+
name: nokogiri
|
|
15
|
+
requirement: !ruby/object:Gem::Requirement
|
|
16
|
+
requirements:
|
|
17
|
+
- - ~>
|
|
18
|
+
- !ruby/object:Gem::Version
|
|
19
|
+
version: '1.5'
|
|
20
|
+
type: :runtime
|
|
21
|
+
prerelease: false
|
|
22
|
+
version_requirements: !ruby/object:Gem::Requirement
|
|
23
|
+
requirements:
|
|
24
|
+
- - ~>
|
|
25
|
+
- !ruby/object:Gem::Version
|
|
26
|
+
version: '1.5'
|
|
27
|
+
- !ruby/object:Gem::Dependency
|
|
28
|
+
name: bundler
|
|
29
|
+
requirement: !ruby/object:Gem::Requirement
|
|
30
|
+
requirements:
|
|
31
|
+
- - ~>
|
|
32
|
+
- !ruby/object:Gem::Version
|
|
33
|
+
version: '1.3'
|
|
34
|
+
type: :development
|
|
35
|
+
prerelease: false
|
|
36
|
+
version_requirements: !ruby/object:Gem::Requirement
|
|
37
|
+
requirements:
|
|
38
|
+
- - ~>
|
|
39
|
+
- !ruby/object:Gem::Version
|
|
40
|
+
version: '1.3'
|
|
41
|
+
- !ruby/object:Gem::Dependency
|
|
42
|
+
name: rake
|
|
43
|
+
requirement: !ruby/object:Gem::Requirement
|
|
44
|
+
requirements:
|
|
45
|
+
- - '>='
|
|
46
|
+
- !ruby/object:Gem::Version
|
|
47
|
+
version: '0'
|
|
48
|
+
type: :development
|
|
49
|
+
prerelease: false
|
|
50
|
+
version_requirements: !ruby/object:Gem::Requirement
|
|
51
|
+
requirements:
|
|
52
|
+
- - '>='
|
|
53
|
+
- !ruby/object:Gem::Version
|
|
54
|
+
version: '0'
|
|
55
|
+
- !ruby/object:Gem::Dependency
|
|
56
|
+
name: rspec
|
|
57
|
+
requirement: !ruby/object:Gem::Requirement
|
|
58
|
+
requirements:
|
|
59
|
+
- - '>='
|
|
60
|
+
- !ruby/object:Gem::Version
|
|
61
|
+
version: '0'
|
|
62
|
+
type: :development
|
|
63
|
+
prerelease: false
|
|
64
|
+
version_requirements: !ruby/object:Gem::Requirement
|
|
65
|
+
requirements:
|
|
66
|
+
- - '>='
|
|
67
|
+
- !ruby/object:Gem::Version
|
|
68
|
+
version: '0'
|
|
69
|
+
description: Siteseeker integration library
|
|
70
|
+
email:
|
|
71
|
+
- marten@thavenius.se
|
|
72
|
+
executables: []
|
|
73
|
+
extensions: []
|
|
74
|
+
extra_rdoc_files: []
|
|
75
|
+
files:
|
|
76
|
+
- .gitignore
|
|
77
|
+
- .rspec
|
|
78
|
+
- COPYING
|
|
79
|
+
- Gemfile
|
|
80
|
+
- README.md
|
|
81
|
+
- Rakefile
|
|
82
|
+
- lib/siteseeker_normalizer.rb
|
|
83
|
+
- lib/siteseeker_normalizer/client.rb
|
|
84
|
+
- lib/siteseeker_normalizer/parse.rb
|
|
85
|
+
- lib/siteseeker_normalizer/parse/category.rb
|
|
86
|
+
- lib/siteseeker_normalizer/parse/entry.rb
|
|
87
|
+
- lib/siteseeker_normalizer/version.rb
|
|
88
|
+
- siteseeker_normalizer.gemspec
|
|
89
|
+
- spec/fixtures/barn.html
|
|
90
|
+
- spec/fixtures/barnomsrg.html
|
|
91
|
+
- spec/siteseeeker_normalizer_spec.rb
|
|
92
|
+
- spec/spec_helper.rb
|
|
93
|
+
homepage: https://github.com/malmostad/siteseeker_normalizer
|
|
94
|
+
licenses:
|
|
95
|
+
- AGPL v3
|
|
96
|
+
metadata: {}
|
|
97
|
+
post_install_message:
|
|
98
|
+
rdoc_options: []
|
|
99
|
+
require_paths:
|
|
100
|
+
- lib
|
|
101
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
|
102
|
+
requirements:
|
|
103
|
+
- - '>='
|
|
104
|
+
- !ruby/object:Gem::Version
|
|
105
|
+
version: '0'
|
|
106
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
|
107
|
+
requirements:
|
|
108
|
+
- - '>='
|
|
109
|
+
- !ruby/object:Gem::Version
|
|
110
|
+
version: '0'
|
|
111
|
+
requirements: []
|
|
112
|
+
rubyforge_project:
|
|
113
|
+
rubygems_version: 2.1.11
|
|
114
|
+
signing_key:
|
|
115
|
+
specification_version: 4
|
|
116
|
+
summary: A Ruby Gem for making requests and parsing the response from Siteseeker to
|
|
117
|
+
a structured object.
|
|
118
|
+
test_files:
|
|
119
|
+
- spec/fixtures/barn.html
|
|
120
|
+
- spec/fixtures/barnomsrg.html
|
|
121
|
+
- spec/siteseeeker_normalizer_spec.rb
|
|
122
|
+
- spec/spec_helper.rb
|