HAYASHI Kentaro
null+****@clear*****
Fri Apr 26 12:28:48 JST 2013
HAYASHI Kentaro 2013-04-26 12:28:48 +0900 (Fri, 26 Apr 2013) New Revision: 59f4008af8688b753b2a2d1e325a1de6ed1306e5 https://github.com/groonga/groonga.org/commit/59f4008af8688b753b2a2d1e325a1de6ed1306e5 Message: blog en: add 3.0.3 release entry Added files: en/_posts/2013-04-29-release.textile Added: en/_posts/2013-04-29-release.textile (+155 -0) 100644 =================================================================== --- /dev/null +++ en/_posts/2013-04-29-release.textile 2013-04-26 12:28:48 +0900 (9d9d8b1) @@ -0,0 +1,155 @@ +--- +layout: post.en +title: Groonga 3.0.3 has been released +description: Groonga 3.0.3 has been released! +published: false +--- + + +h2. Groonga 3.0.3 has been released + +"Groonga 3.0.3":/docs/news.html#release-3-0-3 has been released! + +How to install: "Install":/docs/install.html + +There are two topics for this release. + +* Supported to learn with the value of customized weight +* Supported to build customized groonga without changing spec file +* Notice release process about Fedora 18 +* Supported Ubuntu 13.04 Raring Ringtail + +h3. Supported to learn with the value of customized weight + +In this release, learning with the value of customized weight has been supported. + +groonga supports three "suggest":http://groonga.org/docs/suggest/introduction.html related functionality such as completion, correction and suggestion. + +There is a need to learn properly for satisfying your demands. + +If you want to change the value of weight for each season, you need to re-learn some times for same data. + +This way has a performance penalty for big data. + +So, we has supported to customize the value of weight, there is no need to learn multiple times for same data. + +There is only one thing you need to do is customising configuration table before loading target data. + +Here is the concrete example such as suggesting in season vegetable. + +First, make database for example. + +<pre> +% mkdir -p testdb +% groonga -n testdb/db shutdown +</pre> + +Second, make a table for learning by @groonga-suggest-create-dataset@ command. + +<pre> +% groonga-suggest-create-dataset testdb/db query +</pre> + +At last, prepare dataset for learning. Note that changing the value of weight configuration before loading "cabbage" data. + +<pre> +load --table event_query --each 'suggest_preparer(_id, type, item, sequence, time, pair_query)' +[ +{"sequence": "1", "time": 1312950808.86057, "item": "tomato in season vegetables", "type": "submit"} +{"sequence": "1", "time": 1312950808.86057, "item": "sweet potato in season vegetables", "type": "submit"} +{"sequence": "1", "time": 1312950808.86057, "item": "chinese cabbage in season vegetables", "type": "submit"} +] + +load --table configuration +[ +{"_key": "query", "weight": 10} +] + +load --table event_query --each 'suggest_preparer(_id, type, item, sequence, time, pair_query)' +[ +{"sequence": "1", "time": 1312950803.86057, "item": "cabbage in season vegetables", "type": "submit"}, +] +</pre> + +So, let's try to use suggestion functionality. Here is the query to get suggestion result in season vegetables. + +<pre> +suggest --table item_query --column kana --types suggest --frequency_threshold 1 --query season +</pre> + +As a result, there are four results and you know that cabbage is in the best season by suggestion score. + +<pre> +[ + [0,1366613297.59312,0.000118494033813477], + { + "suggest":[[4], + [ + ["_key","ShortText"],["_score","Int32"]], + ["cabbage in season vegetables",10], + ["tomato in season vegetables",1], + ["chinese cabbage in season vegetables",1], + ["sweet potato in season vegetables",1] + ] + } +] +</pre> + +Previously, you need to do learning many times by followings way to achieve same effects. + +<pre> +load --table event_query --each 'suggest_preparer(_id, type, item, sequence, time, pair_query)' +[ +{"sequence": "1", "time": 1312950803.86057, "item": "cabbage in season vegetables", "type": "submit"}, +{"sequence": "1", "time": 1312950803.86057, "item": "cabbage in season vegetables", "type": "submit"}, +{"sequence": "1", "time": 1312950803.86057, "item": "cabbage in season vegetables", "type": "submit"}, +{"sequence": "1", "time": 1312950803.86057, "item": "cabbage in season vegetables", "type": "submit"}, +... +</pre> + +h3. Supported to build customized groonga without changing spec file + +This release began to add a parameter for building customized version of groonga. + +For example, if you want to customize stack size which groonga uses, you need to change the spec file about @--with-stack-size=@ option. + +<pre> +%configure \ + --disable-static \ + --with-package-platform=redhat \ + --with-stack-size=4096 \ + --with-zlib \ +%if %{use_lzo} + --with-lzo \ +%endif + --with-munin-plugins +</pre> + +In this release, you can specify additional parameter as a rpmbuild command's one. There is no need to modify spec file directly. + +<pre> +% rpmbuild --rebuild --define 'additional_configure_options --with-stach-size=4096' groonga-3.0.3-0.el6.src.rpm +</pre> + +h3. Notice release process about Fedora 18 + +Since groonga 3.0.2 release, we had stopped to provide groonga packages by self hosted yum repository. +Instead, we only provides groonga packages via Fedora official yum repository. + +There is no need for Fedora users to regist groonga self hosted repository now. + +We cooperate with Mr. Daiki Ueno who maintaines groonga packages so far in Fedora official repository. + +But there is a bit time lag about two or three weeks until you can install latest release. + +h3. Supported Ubuntu 13.04 Raring Ringtail + +This release began to support Ubuntu 13.04 Raring Ringtail which is released at Apr 25, 2013. + +See the "install documentation":http://groonga.org/docs/install/ubuntu.html#raring-ringrail about details. + +h3. Conclusion + +See "Release 3.0.3 2013/04/29":/docs/news.html#release-3-0-3 about detailed changes since 3.0.2. + +Let's search by groonga! -------------- next part -------------- HTML����������������������������...Download