rroonga 5.0.1 → 5.0.2
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/README.md +1 -1
- data/Rakefile +0 -6
- data/example/bookmark.rb +1 -6
- data/example/index-html.rb +0 -1
- data/ext/groonga/extconf.rb +1 -6
- data/ext/groonga/rb-grn-logger.c +159 -3
- data/ext/groonga/rb-grn-plugin.c +17 -1
- data/ext/groonga/rb-grn-query-logger.c +129 -3
- data/ext/groonga/rb-grn.h +1 -1
- data/lib/groonga/query-logger.rb +48 -3
- data/lib/groonga/schema.rb +16 -3
- data/rroonga-build.rb +3 -3
- data/test/test-logger.rb +97 -2
- data/test/test-plugin.rb +5 -1
- data/test/test-query-logger.rb +149 -0
- data/test/test-schema.rb +19 -0
- metadata +4 -4
- data/doc/text/install.textile +0 -145
- data/doc/text/tutorial.textile +0 -510
data/doc/text/install.textile
DELETED
@@ -1,145 +0,0 @@
|
|
1
|
-
h1. Install
|
2
|
-
|
3
|
-
This document describes how to install Rroonga.
|
4
|
-
|
5
|
-
You can install Rroonga by RubyGems. It is the standard way for Ruby
|
6
|
-
libraries.
|
7
|
-
|
8
|
-
Rroonga is depends on Groonga. So you need to install both Groonga and
|
9
|
-
Rroonga. You can't install Groonga by RubyGems because it isn't Ruby
|
10
|
-
library. But don't worry. Rroonga provides the following options for
|
11
|
-
easy to install:
|
12
|
-
|
13
|
-
* Rroonga downloads, builds and installs Groonga automatically. You
|
14
|
-
don't need to do them explicitly.
|
15
|
-
* Rroonga uses Groonga installed by your packaging system.
|
16
|
-
|
17
|
-
The following sections describe the above in detail.
|
18
|
-
|
19
|
-
h2. Install with auto Groonga build
|
20
|
-
|
21
|
-
Rroonga searches Groonga on install. If Rroonga can't find
|
22
|
-
Groonga, Rroonga downloads, builds and installs Groonga
|
23
|
-
automatically.
|
24
|
-
|
25
|
-
Type the following command to install Rroonga and Groonga. You don't
|
26
|
-
need to install Groonga explicitly:
|
27
|
-
|
28
|
-
<pre>
|
29
|
-
!!!command_line
|
30
|
-
% gem install rroonga
|
31
|
-
</pre>
|
32
|
-
|
33
|
-
h2. Install with Groonga package
|
34
|
-
|
35
|
-
You can use Groonga package on you packaging system instead of
|
36
|
-
building Groonga by yourself. There are the following advantages for
|
37
|
-
this option:
|
38
|
-
|
39
|
-
* It reduces installation time.
|
40
|
-
* It doesn't fail on building Groonga.
|
41
|
-
|
42
|
-
h3. Windows
|
43
|
-
|
44
|
-
Rroonga gem for Windows includes both pre-compiled Rroonga and Groonga
|
45
|
-
in the gem. So what you need to do is you just install rroonga gem.
|
46
|
-
|
47
|
-
Type the following command on Ruby console:
|
48
|
-
|
49
|
-
<pre>
|
50
|
-
!!!command_line
|
51
|
-
> gem install rroonga
|
52
|
-
</pre>
|
53
|
-
|
54
|
-
This document assumes that you're using "RubyInstaller for
|
55
|
-
Windows":http://rubyinstaller.org/ .
|
56
|
-
|
57
|
-
h3. OS X
|
58
|
-
|
59
|
-
There are Groonga packages for OS X environment.
|
60
|
-
|
61
|
-
h4. MacPorts
|
62
|
-
|
63
|
-
If you're using "MacPorts":http://www.macports.org/ , type the
|
64
|
-
following commands on your terminal:
|
65
|
-
|
66
|
-
<pre>
|
67
|
-
!!!command_line
|
68
|
-
% sudo port install groonga
|
69
|
-
% sudo gem install rroonga
|
70
|
-
</pre>
|
71
|
-
|
72
|
-
h4. Homebrew
|
73
|
-
|
74
|
-
If you're using "Homebrew":http://brew.sh/ , type the
|
75
|
-
following commands on your terminal:
|
76
|
-
|
77
|
-
<pre>
|
78
|
-
!!!command_line
|
79
|
-
% brew install groonga
|
80
|
-
% gem install rroonga
|
81
|
-
</pre>
|
82
|
-
|
83
|
-
h3. Debian GNU/Linux
|
84
|
-
|
85
|
-
You can install the Groonga package by apt. See "Groonga
|
86
|
-
documentation":http://groonga.org/docs/install/debian.html how to set
|
87
|
-
apt-line up.
|
88
|
-
|
89
|
-
Type the following commands on your terminal after you finish to set
|
90
|
-
apt-line up.
|
91
|
-
|
92
|
-
<pre>
|
93
|
-
!!!command_line
|
94
|
-
% sudo apt-get install -y libgroonga-dev
|
95
|
-
% sudo gem install rroonga
|
96
|
-
</pre>
|
97
|
-
|
98
|
-
h3. Ubuntu
|
99
|
-
|
100
|
-
You can install the Groonga package by apt. See "Groonga
|
101
|
-
documentation":http://groonga.org/docs/install/ubuntu.html how to set
|
102
|
-
apt-line up.
|
103
|
-
|
104
|
-
Type the following commands on your terminal after you finish to set
|
105
|
-
apt-line up.
|
106
|
-
|
107
|
-
<pre>
|
108
|
-
!!!command_line
|
109
|
-
% sudo apt-get install -y libgroonga-dev
|
110
|
-
% sudo gem install rroonga
|
111
|
-
</pre>
|
112
|
-
|
113
|
-
h3. CentOS
|
114
|
-
|
115
|
-
You can install the Groonga package by yum. See "Groonga
|
116
|
-
documentation":http://groonga.org/docs/install/centos.html how to set
|
117
|
-
yum repository up.
|
118
|
-
|
119
|
-
But you need to install Ruby 1.9.3 or later by yourself. Both CentOS 5
|
120
|
-
and 6 ship Ruby 1.8. Rroonga doesn't support Ruby 1.8.
|
121
|
-
|
122
|
-
Type the following commands on your terminal after you finish to set
|
123
|
-
yum repository up and installing Ruby 1.9.3 or later.
|
124
|
-
|
125
|
-
<pre>
|
126
|
-
!!!command_line
|
127
|
-
% sudo yum install groonga-devel -y
|
128
|
-
% gem install rroonga
|
129
|
-
</pre>
|
130
|
-
|
131
|
-
h3. Fedora
|
132
|
-
|
133
|
-
You can install the Groonga package by yum. The Groonga package is
|
134
|
-
included in the official Fedora repository.
|
135
|
-
|
136
|
-
<pre>
|
137
|
-
!!!command_line
|
138
|
-
% sudo yum install groonga-devel -y
|
139
|
-
% sudo gem install rroonga
|
140
|
-
</pre>
|
141
|
-
|
142
|
-
h2. Links
|
143
|
-
|
144
|
-
* "2. Install - Groonga documentation":http://groonga.org/docs/install.html
|
145
|
-
|
data/doc/text/tutorial.textile
DELETED
@@ -1,510 +0,0 @@
|
|
1
|
-
h1. Tutorial
|
2
|
-
|
3
|
-
This page introduce how to use Rroonga via a simple application making.
|
4
|
-
|
5
|
-
h2. Install
|
6
|
-
|
7
|
-
You can install Rroonga in your compter with RubyGems.
|
8
|
-
|
9
|
-
<pre>
|
10
|
-
!!!command_line
|
11
|
-
% sudo gem install rroonga
|
12
|
-
</pre>
|
13
|
-
|
14
|
-
h2. Create Database
|
15
|
-
|
16
|
-
Let's create database for simple bookmark application.
|
17
|
-
Please execute irb with loading Rroonga with this command:
|
18
|
-
|
19
|
-
<pre>
|
20
|
-
!!!irb
|
21
|
-
% irb --simple-prompt -r groonga
|
22
|
-
>>
|
23
|
-
</pre>
|
24
|
-
|
25
|
-
Now you use UTF-8 as the encoding of database.
|
26
|
-
|
27
|
-
<pre>
|
28
|
-
!!!irb
|
29
|
-
>> Groonga::Context.default_options = {:encoding => :utf8}
|
30
|
-
=> {:encoding=>:utf8}
|
31
|
-
</pre>
|
32
|
-
|
33
|
-
Then, try to create database in a file.
|
34
|
-
|
35
|
-
<pre>
|
36
|
-
!!!irb
|
37
|
-
>> Groonga::Database.create(:path => "/tmp/bookmark.db")
|
38
|
-
=> #<Groonga::Database ...>
|
39
|
-
</pre>
|
40
|
-
|
41
|
-
From now, the created database is used implicitly.
|
42
|
-
You don't have to be aware of it after you created a database first.
|
43
|
-
|
44
|
-
h2. Define table
|
45
|
-
|
46
|
-
Groonga supports 4 types of tables.
|
47
|
-
|
48
|
-
- Groonga::Hash :=
|
49
|
-
Hash table. It manages records via each primary key. It supports
|
50
|
-
very quickly exact match search.
|
51
|
-
=:
|
52
|
-
|
53
|
-
- Groonga::PatriciaTrie :=
|
54
|
-
Patricia Trie. It supports some search such as predictive search and
|
55
|
-
common prefix search, but it provides a little slowly exact match search
|
56
|
-
than Groonga::Hash. It provides cursor to take records in ascending
|
57
|
-
or descending order.
|
58
|
-
=:
|
59
|
-
|
60
|
-
- Groonga::DoubleArrayTrie :=
|
61
|
-
Double Array Trie. It requires large spaces rather than other
|
62
|
-
tables, but it can update key without ID change. It provides exract
|
63
|
-
match search, predictive search and common prefix search and cursor
|
64
|
-
like Groonga::PatriciaTrie.
|
65
|
-
=:
|
66
|
-
|
67
|
-
- Groonga::Array :=
|
68
|
-
Array. It doesn't have primary keys. It manages records by ID.
|
69
|
-
=:
|
70
|
-
|
71
|
-
Now, you use Groonga::Hash and create the table named @Items@. The type
|
72
|
-
of its primary key is String.
|
73
|
-
|
74
|
-
<pre>
|
75
|
-
!!!irb
|
76
|
-
>> Groonga::Schema.create_table("Items", :type => :hash)
|
77
|
-
=> [...]
|
78
|
-
</pre>
|
79
|
-
|
80
|
-
You have @Items@ table by this code.
|
81
|
-
You can refer the defined table with Groonga.[] like below:
|
82
|
-
|
83
|
-
<pre>
|
84
|
-
!!!irb
|
85
|
-
>> items = Groonga["Items"]
|
86
|
-
=> #<Groonga::Hash ...>
|
87
|
-
</pre>
|
88
|
-
|
89
|
-
You can treat it like Hash.
|
90
|
-
For example, let's type @items.size@ to get the number of records in
|
91
|
-
the table.
|
92
|
-
|
93
|
-
<pre>
|
94
|
-
!!!irb
|
95
|
-
>> items.size
|
96
|
-
=> 0
|
97
|
-
</pre>
|
98
|
-
|
99
|
-
h2. Add records
|
100
|
-
|
101
|
-
Let's add records to @Items@ table.
|
102
|
-
|
103
|
-
<pre>
|
104
|
-
!!!irb
|
105
|
-
>> items.add("http://en.wikipedia.org/wiki/Ruby")
|
106
|
-
=> #<Groonga::Record ...>
|
107
|
-
>> items.add("http://www.ruby-lang.org/")
|
108
|
-
=> #<Groonga::Record ...>
|
109
|
-
</pre>
|
110
|
-
|
111
|
-
Please check the number of records. It increases from 0 to 2.
|
112
|
-
|
113
|
-
<pre>
|
114
|
-
!!!irb
|
115
|
-
>> items.size
|
116
|
-
=> 2
|
117
|
-
</pre>
|
118
|
-
|
119
|
-
If you can get record by primary key, type like below:
|
120
|
-
|
121
|
-
<pre>
|
122
|
-
!!!irb
|
123
|
-
>> items["http://en.wikipedia.org/wiki/Ruby"]
|
124
|
-
=> #<Groonga::Record ...>
|
125
|
-
</pre>
|
126
|
-
|
127
|
-
h2. Full text search
|
128
|
-
|
129
|
-
Let's add item's title to full text search.
|
130
|
-
|
131
|
-
first, you add the @Text@ type column "@title@" to @Items@ table.
|
132
|
-
|
133
|
-
<pre>
|
134
|
-
!!!irb
|
135
|
-
>> Groonga::Schema.change_table("Items") do |table|
|
136
|
-
?> table.text("title")
|
137
|
-
>> end
|
138
|
-
=> [...]
|
139
|
-
</pre>
|
140
|
-
|
141
|
-
Defined columns is named as @#{TABLE_NAME}.#{COLUMN_NAME}@.
|
142
|
-
You can refer them with {Groonga.[]} as same as tables.
|
143
|
-
|
144
|
-
<pre>
|
145
|
-
!!!irb
|
146
|
-
>> title_column = Groonga["Items.title"]
|
147
|
-
=> #<Groonga::VariableSizeColumn ...>
|
148
|
-
</pre>
|
149
|
-
|
150
|
-
|
151
|
-
Secondly, let's add the table containing terms from splited from texts.
|
152
|
-
Then you define the @Terms@ for it.
|
153
|
-
|
154
|
-
<pre>
|
155
|
-
!!!irb
|
156
|
-
>> Groonga::Schema.create_table("Terms",
|
157
|
-
?> :type => :patricia_trie,
|
158
|
-
?> :normalizer => :NormalizerAuto,
|
159
|
-
?> :default_tokenizer => "TokenBigram")
|
160
|
-
</pre>
|
161
|
-
|
162
|
-
You specify @:default_tokenzier => "TokenBigram"@ for "Tokenizer" in
|
163
|
-
the above code.
|
164
|
-
"Tokenizer" is the object to split terms from texts. The default value
|
165
|
-
for it is none.
|
166
|
-
Full text search requires a tokenizer, so you specify "Bigram", a type
|
167
|
-
of N-gram.
|
168
|
-
Full text search with N-gram uses splited N characters and their
|
169
|
-
position in texts. "N" in N-gram specifies the number of each terms.
|
170
|
-
Groonga supports Unigram (N=1), Bigram (N=2) and Trigram (N=3).
|
171
|
-
|
172
|
-
You also specify @:normalizer => :NormalizerAuto@ to search texts with
|
173
|
-
ignoring the case.
|
174
|
-
|
175
|
-
Now, you ready table for terms, so you define the index of
|
176
|
-
@Items.tiltle@ column.
|
177
|
-
|
178
|
-
<pre>
|
179
|
-
!!!irb
|
180
|
-
>> Groonga::Schema.change_table("Terms") do |table|
|
181
|
-
?> table.index("Items.title")
|
182
|
-
>> end
|
183
|
-
=> [...]
|
184
|
-
</pre>
|
185
|
-
|
186
|
-
You may feel a few unreasonable code. The index of @Items@ table's
|
187
|
-
column is defined as the column in @Terms@.
|
188
|
-
|
189
|
-
When a record is added to @Items@, groonga adds records associated
|
190
|
-
each terms in it to @Terms@ automatically.
|
191
|
-
|
192
|
-
|
193
|
-
@Terms@ is a few particular table, but you can add some columns to term
|
194
|
-
table such as @Terms@ and manage many attributes of each terms. It is
|
195
|
-
very useful to process particular search.
|
196
|
-
|
197
|
-
Now, you finished table definition.
|
198
|
-
Let's put some values to @title@ of each record you added before.
|
199
|
-
|
200
|
-
<pre>
|
201
|
-
!!!irb
|
202
|
-
>> items["http://en.wikipedia.org/wiki/Ruby"].title = "Ruby"
|
203
|
-
=> "Ruby"
|
204
|
-
>> items["http://www.ruby-lang.org/"].title = "Ruby Programming Language"
|
205
|
-
"Ruby Programming Language"
|
206
|
-
</pre>
|
207
|
-
|
208
|
-
Now, you can do full text search like above:
|
209
|
-
|
210
|
-
<pre>
|
211
|
-
!!!irb
|
212
|
-
>> ruby_items = items.select {|record| record.title =~ "Ruby"}
|
213
|
-
=> #<Groonga::Hash ..., normalizer: (nil)>
|
214
|
-
</pre>
|
215
|
-
|
216
|
-
Groonga returns the search result as Groonga::Hash.
|
217
|
-
Keys in this hash table is records of hitted @Items@.
|
218
|
-
|
219
|
-
<pre>
|
220
|
-
!!!irb
|
221
|
-
>> ruby_items.collect {|record| record.key.key}
|
222
|
-
=> ["http://en.wikipedia.org/wiki/Ruby", "http://www.ruby-lang.org/"]
|
223
|
-
</pre>
|
224
|
-
|
225
|
-
In above example, you get records in @Items@ with @record.key@, and
|
226
|
-
keys of them with @record.key.key@.
|
227
|
-
|
228
|
-
You can access a refered key in records briefly with @record["_key"]@.
|
229
|
-
|
230
|
-
<pre>
|
231
|
-
!!!irb
|
232
|
-
>> ruby_items.collect {|record| record["_key"]}
|
233
|
-
=> ["http://en.wikipedia.org/wiki/Ruby", "http://www.ruby-lang.org/"]
|
234
|
-
</pre>
|
235
|
-
|
236
|
-
h2. Improve the simple bookmark application
|
237
|
-
|
238
|
-
Let's try to improve this simple application a little. You can create
|
239
|
-
bookmark application for multi users and they can comment to each
|
240
|
-
bookmarks.
|
241
|
-
|
242
|
-
First, you add tables for users and for comments like below:
|
243
|
-
|
244
|
-
!http://qwik.jp/senna/senna2.files/rect4605.png!
|
245
|
-
|
246
|
-
Let's add the table for users, @Users@.
|
247
|
-
|
248
|
-
<pre>
|
249
|
-
!!!irb
|
250
|
-
>> Groonga::Schema.create_table("Users", :type => :hash) do |table|
|
251
|
-
?> table.text("name")
|
252
|
-
>> end
|
253
|
-
=> [...]
|
254
|
-
</pre>
|
255
|
-
|
256
|
-
|
257
|
-
Next, let's add the table for comments as @Comments@.
|
258
|
-
|
259
|
-
<pre>
|
260
|
-
!!!irb
|
261
|
-
>> Groonga::Schema.create_table("Comments") do |table|
|
262
|
-
?> table.reference("item")
|
263
|
-
>> table.reference("author", "Users")
|
264
|
-
>> table.text("content")
|
265
|
-
>> table.time("issued")
|
266
|
-
>> end
|
267
|
-
=> [...]
|
268
|
-
</pre>
|
269
|
-
|
270
|
-
Then you define the index of @content@ column in @Comments@ for full
|
271
|
-
text search.
|
272
|
-
|
273
|
-
<pre>
|
274
|
-
!!!irb
|
275
|
-
>> Groonga::Schema.change_table("Terms") do |table|
|
276
|
-
?> table.index("Comments.content")
|
277
|
-
>> end
|
278
|
-
=> [...]
|
279
|
-
</pre>
|
280
|
-
|
281
|
-
You finish table definition by above code.
|
282
|
-
|
283
|
-
Secondly, you add some users to @Users@.
|
284
|
-
|
285
|
-
<pre>
|
286
|
-
!!!irb
|
287
|
-
>> users = Groonga["Users"]
|
288
|
-
=> #<Groonga::Hash ...>
|
289
|
-
>> users.add("alice", :name => "Alice")
|
290
|
-
=> #<Groonga::Record ...>
|
291
|
-
>> users.add("bob", :name => "Bob")
|
292
|
-
=> #<Groonga::Record ...>
|
293
|
-
</pre>
|
294
|
-
|
295
|
-
Now, let's write the process to bookmark by a user.
|
296
|
-
You assume that the user, @moritan@, bookmark a page including
|
297
|
-
infomation related Ruby.
|
298
|
-
|
299
|
-
First, you check if the page has been added @Items@ already.
|
300
|
-
|
301
|
-
<pre>
|
302
|
-
!!!irb
|
303
|
-
>> items.has_key?("http://www.ruby-doc.org/")
|
304
|
-
=> false
|
305
|
-
</pre>
|
306
|
-
|
307
|
-
The page hasn't been added, so you add it to @Items@.
|
308
|
-
|
309
|
-
<pre>
|
310
|
-
!!!irb
|
311
|
-
>> items.add("http://www.ruby-doc.org/",
|
312
|
-
?> :title => "Ruby-Doc.org: Documenting the Ruby Language")
|
313
|
-
=> #<Groonga::Record ...>
|
314
|
-
</pre>
|
315
|
-
|
316
|
-
Next, you add the record to @Comments@. This record contains this page
|
317
|
-
as its @item@ column.
|
318
|
-
|
319
|
-
<pre>
|
320
|
-
!!!irb
|
321
|
-
>> require "time"
|
322
|
-
=> true
|
323
|
-
>> comments = Groonga["Comments"]
|
324
|
-
=> #<Groonga::Array ...>
|
325
|
-
>> comments.add(:item => "http://www.ruby-doc.org/",
|
326
|
-
?> :author => "alice",
|
327
|
-
?> :content => "Ruby documents",
|
328
|
-
?> :issued => Time.parse("2010-11-20T18:01:22+09:00"))
|
329
|
-
=> #<Groonga::Record ...>
|
330
|
-
</pre>
|
331
|
-
|
332
|
-
h2. Define methods for this process
|
333
|
-
|
334
|
-
For usefull, you define methods for above processes.
|
335
|
-
|
336
|
-
<pre>
|
337
|
-
!!!irb
|
338
|
-
>> @items = items
|
339
|
-
=> #<Groonga::Hash ...>
|
340
|
-
>> @comments = comments
|
341
|
-
=> #<Groonga::Array ...>
|
342
|
-
>> def add_bookmark(url, title, author, content, issued)
|
343
|
-
>> item = @items[url] || @items.add(url, :title => title)
|
344
|
-
>> @comments.add(:item => item,
|
345
|
-
?> :author => author,
|
346
|
-
?> :content => content,
|
347
|
-
?> :issued => issued)
|
348
|
-
>> end
|
349
|
-
=> nil
|
350
|
-
</pre>
|
351
|
-
|
352
|
-
You assign @items@ and @comments@ to each instance variable, so you can
|
353
|
-
use them in @add_bookmark@ method.
|
354
|
-
|
355
|
-
@add_bookmark@ executes processes like below:
|
356
|
-
|
357
|
-
* Check if the record associated the page exists in @Items@ table.
|
358
|
-
* If not, add the record to it.
|
359
|
-
* Add the record to @Comments@ table.
|
360
|
-
|
361
|
-
With this method, lets bookmark some pages.
|
362
|
-
|
363
|
-
<pre>
|
364
|
-
!!!irb
|
365
|
-
>> add_bookmark("https://rubygems.org/",
|
366
|
-
?> "RubyGems.org | your community gem host", "alice", "Ruby gems",
|
367
|
-
?> Time.parse("2010-10-07T14:18:28+09:00"))
|
368
|
-
=> #<Groonga::Record ...>
|
369
|
-
>> add_bookmark("http://ranguba.org/",
|
370
|
-
?> "Fulltext search by Ruby with groonga - Ranguba", "bob",
|
371
|
-
?> "Ruby groonga fulltextsearch",
|
372
|
-
?> Time.parse("2010-11-11T12:39:59+09:00"))
|
373
|
-
=> #<Groonga::Record ...>
|
374
|
-
>> add_bookmark("http://www.ruby-doc.org/",
|
375
|
-
?> "ruby-doc", "bob", "ruby documents",
|
376
|
-
?> Time.parse("2010-07-28T20:46:23+09:00"))
|
377
|
-
=> #<Groonga::Record ...>
|
378
|
-
</pre>
|
379
|
-
|
380
|
-
h2. Full text search part 2
|
381
|
-
|
382
|
-
Let's do full text search for added records.
|
383
|
-
|
384
|
-
<pre>
|
385
|
-
!!!irb
|
386
|
-
>> records = comments.select do |record|
|
387
|
-
?> record["content"] =~ "Ruby"
|
388
|
-
>> end
|
389
|
-
=> #<Groonga::Hash ...>
|
390
|
-
>> records.each do |record|
|
391
|
-
?> comment = record
|
392
|
-
>> p [comment.id,
|
393
|
-
?> comment.issued,
|
394
|
-
?> comment.item.title,
|
395
|
-
?> comment.author.name,
|
396
|
-
?> comment.content]
|
397
|
-
>> end
|
398
|
-
[1, 2010-11-20 18:01:22 +0900, "Ruby-Doc.org: Documenting the Ruby Language", "Alice", "Ruby documents"]
|
399
|
-
[2, 2010-10-07 14:18:28 +0900, "RubyGems.org | your community gem host", "Alice", "Ruby gems"]
|
400
|
-
[3, 2010-11-11 12:39:59 +0900, "Fulltext search by Ruby with groonga - Ranguba", "Bob", "Ruby groonga fulltextsearch"]
|
401
|
-
[4, 2010-07-28 20:46:23 +0900, "Ruby-Doc.org: Documenting the Ruby Language", "Bob", "ruby documents"]
|
402
|
-
</pre>
|
403
|
-
|
404
|
-
You can access the columns with the same name method as each them.
|
405
|
-
These methods suport to access the complex data type.
|
406
|
-
(In usually RDB, you should namage JOIN tables, @Items@, @Comments@,
|
407
|
-
@Users@.)
|
408
|
-
|
409
|
-
The search is finished when the first sentence in this codes. The
|
410
|
-
results of this search is the object as records set.
|
411
|
-
|
412
|
-
<pre>
|
413
|
-
!!!irb
|
414
|
-
>> records
|
415
|
-
#<Groonga::Hash ..., size: <4>>
|
416
|
-
</pre>
|
417
|
-
|
418
|
-
You can arrange this records set before output.
|
419
|
-
For example, sort these records in the descending order by date.
|
420
|
-
|
421
|
-
<pre>
|
422
|
-
!!!irb
|
423
|
-
>> records.sort([{:key => "issued", :order => "descending"}]).each do |record|
|
424
|
-
?> comment = record
|
425
|
-
>> p [comment.id,
|
426
|
-
?> comment.issued,
|
427
|
-
?> comment.item.title,
|
428
|
-
?> comment.author.name,
|
429
|
-
?> comment.content]
|
430
|
-
>> end
|
431
|
-
[1, 2010-11-20 18:01:22 +0900, "Ruby-Doc.org: Documenting the Ruby Language", "Alice", "Ruby documents"]
|
432
|
-
[2, 2010-11-11 12:39:59 +0900, "Fulltext search by Ruby with groonga - Ranguba", "Bob", "Ruby groonga fulltextsearch"]
|
433
|
-
[3, 2010-10-07 14:18:28 +0900, "RubyGems.org | your community gem host", "Alice", "Ruby gems"]
|
434
|
-
[4, 2010-07-28 20:46:23 +0900, "Ruby-Doc.org: Documenting the Ruby Language", "Bob", "ruby documents"]
|
435
|
-
=> [...]
|
436
|
-
</pre>
|
437
|
-
|
438
|
-
Let's group the result by each item for easy view.
|
439
|
-
|
440
|
-
<pre>
|
441
|
-
!!!irb
|
442
|
-
>> records.group("item").each do |record|
|
443
|
-
?> item = record.key
|
444
|
-
>> p [record.n_sub_records,
|
445
|
-
?> item.key,
|
446
|
-
?> item.title]
|
447
|
-
>> end
|
448
|
-
[2, "http://www.ruby-doc.org/", "Ruby-Doc.org: Documenting the Ruby Language"]
|
449
|
-
[1, "https://rubygems.org/", "RubyGems.org | your community gem host"]
|
450
|
-
[1, "http://ranguba.org/", "Fulltext search by Ruby with groonga - Ranguba"]
|
451
|
-
=> nil
|
452
|
-
</pre>
|
453
|
-
|
454
|
-
@n_sub_records@ is the number of records in each group.
|
455
|
-
It is similar value as count() function of a query including "GROUP
|
456
|
-
BY" in SQL.
|
457
|
-
|
458
|
-
h2. more complex search
|
459
|
-
|
460
|
-
Now, you challenge the more useful search.
|
461
|
-
|
462
|
-
You should calcurate goodness of fit of search explicitly.
|
463
|
-
|
464
|
-
You can use @Items.title@ and @Comments.content@ as search targets now.
|
465
|
-
@Items.title@ is the a few reliable information taken from each
|
466
|
-
original pages. On the other hands, @Comments.content@ is the less
|
467
|
-
reliable information because this depends on users of bookmark
|
468
|
-
application.
|
469
|
-
|
470
|
-
Then, you search records with this policy:
|
471
|
-
|
472
|
-
* Search item matched @Items.title@ or @Comments.content@.
|
473
|
-
* Add 10 times heavier weight to socres of each record matched
|
474
|
-
@Items.title@ than ones of @Comments.comment@.
|
475
|
-
* If multi @comment@ of one item are matched keyword, specify the sum
|
476
|
-
of scores of each @coments@ as score of the item.
|
477
|
-
|
478
|
-
On this policy, you try to type below:
|
479
|
-
|
480
|
-
<pre>
|
481
|
-
!!!irb
|
482
|
-
>> ruby_comments = @comments.select {|record| record.content =~ "Ruby"}
|
483
|
-
=> #<Groonga::Hash ..., size: <4>
|
484
|
-
>> ruby_items = @items.select do |record|
|
485
|
-
?> target = record.match_target do |match_record|
|
486
|
-
?> match_record.title * 10
|
487
|
-
>> end
|
488
|
-
>> target =~ "Ruby"
|
489
|
-
>> end
|
490
|
-
#<Groonga::Hash ..., size: <4>>
|
491
|
-
</pre>
|
492
|
-
|
493
|
-
You group the results of _ruby_comments_ in each item and union
|
494
|
-
_ruby_items_ .
|
495
|
-
|
496
|
-
<pre>
|
497
|
-
!!!irb
|
498
|
-
>> ruby_items = ruby_comments.group("item").union!(ruby_items)
|
499
|
-
#<Groonga::Hash ..., size: <5>>
|
500
|
-
>> ruby_items.sort([{:key => "_score", :order => "descending"}]).each do |record|
|
501
|
-
>> p [record.score, record.title]
|
502
|
-
>> end
|
503
|
-
[22, "Ruby-Doc.org: Documenting the Ruby Language"]
|
504
|
-
[11, "Fulltext search by Ruby with groonga - Ranguba"]
|
505
|
-
[10, "Ruby Programming Language"]
|
506
|
-
[10, "Ruby"]
|
507
|
-
[1, "RubyGems.org | your community gem host"]
|
508
|
-
</pre>
|
509
|
-
|
510
|
-
Then, you get the result.
|