rroonga 1.2.5-x86-mingw32 → 1.2.7-x86-mingw32

data/vendor/local/share/man/man1/groonga.1

@@ -1,4 +1,4 @@
-.TH "GROONGA" "1" "August 05, 2011" "1.2.4" "groonga"
+.TH "GROONGA" "1" "August 28, 2011" "1.2.5" "groonga"
 .SH NAME
 groonga \- groonga documentation
 .
@@ -257,7 +257,16 @@ Install:
 % brew install groonga
 .ft P
 .fi
-.SS Windows
+.SS Windows (Installer)
+.IP Note
+.
+We only distribute the package for x64.
+.RE
+.sp
+You just download installer (.exe file) from
+\fI\%packages.groonga.org/windows/\fP
+and execute it.
+.SS Windows (zip)
 .IP Note
 .
 We only distribute the package for x64.
@@ -265,7 +274,7 @@ We only distribute the package for x64.
 .sp
 You just download zip\-file from
 \fI\%packages.groonga.org/windows/\fP
-and extract it. We doesn\(aqt distribute installer for now. Sorry.
+and extract it.
 .SS Others
 .sp
 If you will use indexes of tokenizing of each morpheme for
@@ -381,7 +390,7 @@ groonga実行ファイルやgroongaサーバを介して様々なコマンドを
 .sp
 書式2において、空白や、記号「"\(aq()」のうちいずれかを含む値を指定したい場合は、シングルクォート(\(aq)かダブルクォート(")で値を囲みます。
 .sp
-詳しくは、 \fB/execfile\fP のコマンドの項を参考にしてください。
+詳しくは、 \fB/executables/groonga\fP のコマンドの項を参考にしてください。
 .SS 主なコマンド
 .INDENT 0.0
 .INDENT 3.5
@@ -1828,6 +1837,826 @@ Commentsテーブルのlast_modifiedカラムは、Time型です。この型で
 .fi
 .sp
 2010/03/17 14:00:00以前に書かれたコメント5件がヒットしました。また、ドリルダウンの結果も返ってきており、「hsiomaneki」が2件、「グニャラくん」が3件ヒットしていることがわかります。
+.SH SUGGEST
+.IP Note
+.
+The suggest feature specification isn\(aqt stable. The
+specification may be changed.
+.RE
+.sp
+Groonga has the suggest feature. This section describes how to use it and how it works.
+.SS Introduction
+.sp
+The suggest feature in groonga provides the following features:
+.INDENT 0.0
+.IP \(bu 2
+.
+Completion
+.IP \(bu 2
+.
+Correction
+.IP \(bu 2
+.
+Suggestion
+.UNINDENT
+.SS Completion
+.sp
+Completion helps user input. If user inputs a partial word,
+groonga can return complete words from registered
+words.
+.sp
+For example, there are registered words:
+.INDENT 0.0
+.IP \(bu 2
+.
+"groonga"
+.IP \(bu 2
+.
+"complete"
+.IP \(bu 2
+.
+"correction"
+.IP \(bu 2
+.
+"suggest"
+.UNINDENT
+.sp
+An user inputs "co" and groonga returns "complete" and
+"correction" because they starts with "co".
+.sp
+An user inputs "sug" and groonga returns "suggest" because
+"suggest" starts with "sug".
+.sp
+An user inputs "ab" and groonga returns nothing because no
+word starts with "ab".
+.SS Correction
+.sp
+Correction also helps user input. If user inputs a wrong
+word, groonga can return correct words from registered
+correction pairs.
+.sp
+For example, there are registered correction pairs:
+.TS
+center;
+|l|l|.
+_
+T{
+wrong word
+T}	T{
+correct word
+T}
+_
+T{
+grroonga
+gronga
+gronnga
+T}	T{
+groonga
+groonga
+groonga
+T}
+_
+.TE
+.sp
+An user inputs "gronga" and groonga returns "groonga" because
+"gronga" is in wrong word and corresponding correct word is
+"groonga".
+.sp
+An user inputs "roonga" and groonga returns nothing because
+"roonga" isn\(aqt in wrong word.
+.SS Suggestion
+.sp
+Suggestion helps that user filters many found documents. If
+user inputs a query, groonga can return new queries that has
+more additional keywords from registered related query
+pairs.
+.sp
+For example, there are registered related query pairs:
+.TS
+center;
+|l|l|.
+_
+T{
+keyword
+T}	T{
+related query
+T}
+_
+T{
+groonga
+T}	T{
+groonga search engine
+T}
+_
+T{
+search
+T}	T{
+Google search
+T}
+_
+T{
+speed
+T}	T{
+groonga speed
+T}
+_
+.TE
+.sp
+An user inputs "groonga" and groonga returns "groonga search
+engine" because "groonga" is in keyword column and related
+query column is "groonga search engine".
+.sp
+An user inputs "MySQL" and groonga returns nothing because
+"MySQL" isn\(aqt in keyword column values.
+.SS Learning
+.sp
+The suggest feature requires registered data before using
+the feature. Those data can be registered from user inputs.
+Gronnga\-suggest\-httpd and groonga\-suggest\-learner commands
+are provided for the propose.
+.SS Tutorial
+.sp
+TODO...
+.SS Completion
+.sp
+This section describes about the following completion
+features:
+.INDENT 0.0
+.IP \(bu 2
+.
+How it works
+.IP \(bu 2
+.
+How to use
+.IP \(bu 2
+.
+How to learn
+.UNINDENT
+.SS How it works
+.sp
+The completion feature uses three searches to compute
+completed words:
+.INDENT 0.0
+.INDENT 3.5
+.INDENT 0.0
+.IP 1. 3
+.
+Prefix RK search against registered words.
+.IP 2. 3
+.
+Cooccurrence search against learned data.
+.IP 3. 3
+.
+Prefix search against registered words. (optional)
+.UNINDENT
+.UNINDENT
+.UNINDENT
+.SS Prefix RK search
+.sp
+RK means Romaji and Katakana. Prefix RK search can find
+registered words that start with user\(aqs input by romaji,
+katakana or hiragana. It\(aqs useful for searching in Japanese.
+.sp
+For example, there is a registered word "日本". And "ニホン"
+(it must be katakana) is registered as its reading. An user
+can find "日本" by "ni", "二" or "に".
+.SS Cooccurrence search
+.sp
+Cooccurrence search can find registered words from user\(aqs
+partial input. It uses user input sequences that will be
+learned from query logs, access logs and so on.
+.sp
+For example, there is the following user input sequence:
+.TS
+center;
+|l|l|.
+_
+T{
+input
+T}	T{
+submit
+T}
+_
+T{
+s
+T}	T{
+no
+T}
+_
+T{
+se
+T}	T{
+no
+T}
+_
+T{
+sea
+T}	T{
+no
+T}
+_
+T{
+sear
+T}	T{
+no
+T}
+_
+T{
+searc
+T}	T{
+no
+T}
+_
+T{
+search
+T}	T{
+yes
+T}
+_
+T{
+e
+T}	T{
+no
+T}
+_
+T{
+en
+T}	T{
+no
+T}
+_
+T{
+eng
+T}	T{
+no
+T}
+_
+T{
+engi
+T}	T{
+no
+T}
+_
+T{
+engin
+T}	T{
+no
+T}
+_
+T{
+engine
+T}	T{
+no
+T}
+_
+T{
+enginen
+T}	T{
+no (typo!)
+T}
+_
+T{
+engine
+T}	T{
+yes
+T}
+_
+.TE
+.sp
+Groonga creates the following completion pairs:
+.TS
+center;
+|l|l|.
+_
+T{
+input
+T}	T{
+completed word
+T}
+_
+T{
+s
+T}	T{
+search
+T}
+_
+T{
+se
+T}	T{
+search
+T}
+_
+T{
+sea
+T}	T{
+search
+T}
+_
+T{
+sear
+T}	T{
+search
+T}
+_
+T{
+searc
+T}	T{
+search
+T}
+_
+T{
+e
+T}	T{
+engine
+T}
+_
+T{
+en
+T}	T{
+engine
+T}
+_
+T{
+eng
+T}	T{
+engine
+T}
+_
+T{
+engi
+T}	T{
+engine
+T}
+_
+T{
+engin
+T}	T{
+engine
+T}
+_
+T{
+engine
+T}	T{
+engine
+T}
+_
+T{
+enginen
+T}	T{
+engine
+T}
+_
+.TE
+.sp
+All user not\-submitted inputs (e.g. "s", "se" and so on)
+before each an user submission maps to the submitted input
+(e.g. "search").
+.sp
+To be precise, this description isn\(aqt correct because it
+omits about time stamp. Groonga doesn\(aqt case about "all user
+not\-submitted inputs before each an user
+submission". Groonga just case about "all user not\-submitted
+inputs within a minute from an user submission before each
+an user submission". Groonga doesn\(aqt treat user inputs
+before a minute ago.
+.sp
+If an user inputs "sea" and cooccurrence search returns
+"search" because "sea" is in input column and corresponding
+completed word column value is "search".
+.SS Prefix search
+.sp
+Prefix search can find registered word that start with
+user\(aqs input. This search doesn\(aqt care about romaji,
+katakana and hiragana not like Prefix RK search.
+.sp
+This search isn\(aqt always ran. It\(aqs just ran when it\(aqs
+requested explicitly or both prefix RK search and
+cooccurrence search return nothing.
+.sp
+For example, there is a registered word "search". An user
+can find "search" by "s", "se", "sea", "sear", "searc" and
+"search".
+.SS How to use
+.sp
+Groonga provides \fB/commands/suggest\fP command to use
+completion. \fI\-\-type complete\fP option requests completion.
+.sp
+For example, here is an command to get completion results by
+"en":
+.sp
+Execution example:
+.sp
+.nf
+.ft C
+> suggest \-\-table item_query \-\-column kana \-\-types complete \-\-threshold 1 \-\-query en
+[[0,1313131879.64588,0.000913584],{"complete":[[1],[["_key","ShortText"],["_score","Int32"]],["engine",1]]}]
+.ft P
+.fi
+.SS How it learns
+.sp
+Cooccurrence search uses learned data. They are based on
+query logs, access logs and so on. To create learned data,
+groonga needs user input sequence with time stamp and user
+submit input with time stamp.
+.sp
+For example, an user wants to search by "engine". The user
+inputs the query with the following sequence:
+.INDENT 0.0
+.INDENT 3.5
+.INDENT 0.0
+.IP 1. 3
+.
+2011\-08\-10T13:33:23+09:00: e
+.IP 2. 3
+.
+2011\-08\-10T13:33:23+09:00: en
+.IP 3. 3
+.
+2011\-08\-10T13:33:24+09:00: eng
+.IP 4. 3
+.
+2011\-08\-10T13:33:24+09:00: engi
+.IP 5. 3
+.
+2011\-08\-10T13:33:24+09:00: engin
+.IP 6. 3
+.
+2011\-08\-10T13:33:25+09:00: engine (submit!)
+.UNINDENT
+.UNINDENT
+.UNINDENT
+.sp
+Groonga can be learned from the input sequence by the
+following command:
+.sp
+.nf
+.ft C
+load \-\-table event_query \-\-each \(aqsuggest_preparer(_id, type, item, sequence, time, pair_query)\(aq
+[
+{"sequence": "1", "time": 1312950803.86057, "item": "e"},
+{"sequence": "1", "time": 1312950803.96857, "item": "en"},
+{"sequence": "1", "time": 1312950804.26057, "item": "eng"},
+{"sequence": "1", "time": 1312950804.56057, "item": "engi"},
+{"sequence": "1", "time": 1312950804.76057, "item": "engin"},
+{"sequence": "1", "time": 1312950805.86057, "item": "engine", "type": "submit"}
+]
+.ft P
+.fi
+.SS Correction
+.sp
+This section describes about the following correction
+features:
+.INDENT 0.0
+.IP \(bu 2
+.
+How it works
+.IP \(bu 2
+.
+How to use
+.IP \(bu 2
+.
+How to learn
+.UNINDENT
+.SS How it works
+.sp
+The correction feature uses three searches to compute corrected
+words:
+.INDENT 0.0
+.INDENT 3.5
+.INDENT 0.0
+.IP 1. 3
+.
+Cooccurrence search against learned data.
+.IP 2. 3
+.
+Similar search against registered words. (optional)
+.UNINDENT
+.UNINDENT
+.UNINDENT
+.SS Cooccurrence search
+.sp
+Cooccurrence search can find registered words from user\(aqs
+wrong input. It uses user submit sequences that will be
+learned from query logs, access logs and so on.
+.sp
+For example, there are the following user submissions:
+.TS
+center;
+|l|l|.
+_
+T{
+query
+T}	T{
+time
+T}
+_
+T{
+serach (typo!)
+T}	T{
+2011\-08\-10T22:20:50+09:00
+T}
+_
+T{
+search (fixed!)
+T}	T{
+2011\-08\-10T22:20:52+09:00
+T}
+_
+.TE
+.sp
+Groonga creates the following correction pair from the above
+submissions:
+.TS
+center;
+|l|l|.
+_
+T{
+input
+T}	T{
+corrected word
+T}
+_
+T{
+serach
+T}	T{
+search
+T}
+_
+.TE
+.sp
+Groonga treats continuous submissions within a minute as
+input correction by user. Not submitted user input sequence
+between two submissions isn\(aqt used as learned data for
+correction.
+.sp
+If an user inputs "serach" and cooccurrence search returns
+"search" because "serach" is in input column and
+corresponding corrected word column value is "search".
+.SS Similar search
+.sp
+Similar search can find registered words that has one or
+more the same tokens as user input. TokenBigram tokenizer is
+used for tokenization because suggest dataset schema
+created by \fB/executables/groonga\-suggest\-create\-dataset\fP
+uses TokenBigram tokenizer as the default tokenizer.
+.sp
+For example, there is a registered query "search engine". An
+user can find "search engine" by "web search service",
+"sound engine" and so on. Because "search engine" and "web
+search engine" have the same token "search" and "search
+engine" and "sound engine" have the same token "engine".
+.sp
+"search engine" is tokenized to "search" and "engine"
+tokens. (Groonga\(aqs TokenBigram tokenizer doesn\(aqt tokenize
+two characters for continuous alphabets and continuous
+digits for reducing search
+noise. TokenBigramSplitSymbolAlphaDigit tokenizer should be
+used to ensure tokenizing to two characters.) "web search
+service" is tokenized to "web", "search" and
+"service". "sound engine" is tokenized to "sound" and
+"engine".
+.SS How to use
+.sp
+Groonga provides \fB/commands/suggest\fP command to use
+correction. \fI\-\-type correct\fP option requests corrections.
+.sp
+For example, here is an command to get correction results by
+"saerch":
+.sp
+Execution example:
+.sp
+.nf
+.ft C
+> suggest \-\-table item_query \-\-column kana \-\-types correction \-\-threshold 1 \-\-query saerch
+[[0,1313131812.39207,0.000783031],{"correct":[[1],[["_key","ShortText"],["_score","Int32"]],["search",1]]}]
+.ft P
+.fi
+.SS How it learns
+.sp
+Cooccurrence search uses learned data. They are based on
+query logs, access logs and so on. To create learned data,
+groonga needs user submit inputs with time stamp.
+.sp
+For example, an user wants to search by "search" but the
+user has typo "saerch" before inputs the correct query. The
+user inputs the query with the following sequence:
+.INDENT 0.0
+.INDENT 3.5
+.INDENT 0.0
+.IP 1. 3
+.
+2011\-08\-10T13:33:23+09:00: s
+.IP 2. 3
+.
+2011\-08\-10T13:33:23+09:00: sa
+.IP 3. 3
+.
+2011\-08\-10T13:33:24+09:00: sae
+.IP 4. 3
+.
+2011\-08\-10T13:33:24+09:00: saer
+.IP 5. 3
+.
+2011\-08\-10T13:33:24+09:00: saerc
+.IP 6. 3
+.
+2011\-08\-10T13:33:25+09:00: saerch (submit!)
+.IP 7. 3
+.
+2011\-08\-10T13:33:29+09:00: serch (correcting...)
+.IP 8. 3
+.
+2011\-08\-10T13:33:30+09:00: search (submit!)
+.UNINDENT
+.UNINDENT
+.UNINDENT
+.sp
+Groonga can be learned from the input sequence by the
+following command:
+.sp
+.nf
+.ft C
+load \-\-table event_query \-\-each \(aqsuggest_preparer(_id, type, item, sequence, time, pair_query)\(aq
+[
+{"sequence": "1", "time": 1312950803.86057, "item": "s"},
+{"sequence": "1", "time": 1312950803.96857, "item": "sa"},
+{"sequence": "1", "time": 1312950804.26057, "item": "sae"},
+{"sequence": "1", "time": 1312950804.56057, "item": "saer"},
+{"sequence": "1", "time": 1312950804.76057, "item": "saerc"},
+{"sequence": "1", "time": 1312950805.76057, "item": "saerch", "type": "submit"},
+{"sequence": "1", "time": 1312950809.76057, "item": "serch"},
+{"sequence": "1", "time": 1312950810.86057, "item": "search", "type": "submit"}
+]
+.ft P
+.fi
+.SS Suggestion
+.sp
+This section describes about the following completion
+features:
+.INDENT 0.0
+.IP \(bu 2
+.
+How it works
+.IP \(bu 2
+.
+How to use
+.IP \(bu 2
+.
+How to learn
+.UNINDENT
+.SS How it works
+.sp
+The suggestion feature uses a search to compute suggested
+words:
+.INDENT 0.0
+.INDENT 3.5
+.INDENT 0.0
+.IP 1. 3
+.
+Cooccurrence search against learned data.
+.UNINDENT
+.UNINDENT
+.UNINDENT
+.SS Cooccurrence search
+.sp
+Cooccurrence search can find related words from user\(aqs
+input. It uses user submissions that will be learned
+from query logs, access logs and so on.
+.sp
+For example, there are the following user submissions:
+.TS
+center;
+|l|.
+_
+T{
+query
+T}
+_
+T{
+search engine
+T}
+_
+T{
+web search realtime
+T}
+_
+.TE
+.sp
+Groonga creates the following suggestion pairs:
+.TS
+center;
+|l|l|.
+_
+T{
+input
+T}	T{
+suggested words
+T}
+_
+T{
+search
+T}	T{
+search engine
+T}
+_
+T{
+engine
+T}	T{
+search engine
+T}
+_
+T{
+web
+T}	T{
+web search realtime
+T}
+_
+T{
+search
+T}	T{
+web search realtime
+T}
+_
+T{
+realtime
+T}	T{
+web search realtime
+T}
+_
+.TE
+.sp
+Those pairs are created by the following steps:
+.INDENT 0.0
+.INDENT 3.5
+.INDENT 0.0
+.IP 1. 3
+.
+Tokenizes user input query by TokenDelimit tokenizer
+that uses a space as token delimiter. (e.g. "search
+engine" is tokenized to two tokens "search" and
+"engine".)
+.IP 2. 3
+.
+Creates a pair that is consists of a token and original
+query for each token.
+.UNINDENT
+.UNINDENT
+.UNINDENT
+.sp
+If an user inputs "search" and cooccurrence search returns
+"search engine" and "web search realtime" because "search" is
+in two input columns and corresponding suggested word
+columns have "search engine" and "web search realtime".
+.SS How to use
+.sp
+Groonga provides \fB/commands/suggest\fP command to use
+suggestion. \fI\-\-type suggest\fP option requests suggestion
+.sp
+For example, here is an command to get suggestion results by
+"search":
+.sp
+Execution example:
+.sp
+.nf
+.ft C
+> suggest \-\-table item_query \-\-column kana \-\-types suggest \-\-threshold 1 \-\-query search
+[[0,1313131808.92868,0.000607197],{"suggest":[[2],[["_key","ShortText"],["_score","Int32"]],["search engine",1],["web search realtime",1]]}]
+.ft P
+.fi
+.SS How it learns
+.sp
+Cooccurrence search uses learned data. They are based on
+query logs, access logs and so on. To create learned data,
+groonga needs user input sequence with time stamp and user
+submit input with time stamp.
+.sp
+For example, an user wants to search by "engine". The user
+inputs the query with the following sequence:
+.INDENT 0.0
+.INDENT 3.5
+.INDENT 0.0
+.IP 1. 3
+.
+2011\-08\-10T13:33:23+09:00: search engine (submit)
+.IP 2. 3
+.
+2011\-08\-10T13:33:28+09:00: web search realtime (submit)
+.UNINDENT
+.UNINDENT
+.UNINDENT
+.sp
+Groonga can be learned from the submissions by the
+following command:
+.sp
+.nf
+.ft C
+load \-\-table event_query \-\-each \(aqsuggest_preparer(_id, type, item, sequence, time, pair_query)\(aq
+[
+{"sequence": "1", "time": 1312950803.86057, "item": "search engine", "type": "submit"},
+{"sequence": "1", "time": 1312950808.86057, "item": "web search realtime", "type": "submit"}
+]
+.ft P
+.fi
 .SH リファレンスマニュアル
 .SS 実行ファイル
 .sp
@@ -2070,7 +2899,7 @@ out_gqtp コマンドファイル 出力ファイル名
 .UNINDENT
 .SS コマンドファイル
 .sp
-コマンドファイルは、groonga組み込みコマンドを1行に1つずつ記述したテキストファイルです。拡張子に制限はありません。groonga組み込みコマンドに関しては \fBcommands\fP を参照してください。
+コマンドファイルは、groonga組み込みコマンドを1行に1つずつ記述したテキストファイルです。拡張子に制限はありません。groonga組み込みコマンドに関しては \fB/commands\fP を参照してください。
 .SS サンプル
 .sp
 スクリプトファイルのサンプルです。:
@@ -2591,10 +3420,41 @@ GETメソッドのみが使用可能です。
 .sp
 output_typeにはjson, tsv, xmlが指定可能です。
 .sp
-command_versionはコマンドの仕様の互換性を指定します。詳細は \fBcommand_version\fP を参照してください。
+command_versionはコマンドの仕様の互換性を指定します。詳細は \fB/command_version\fP を参照してください。
 .SS 返値
 .sp
 output_typeの指定に従って、コマンドの実行結果を出力します。
+.SS groonga\-suggest\-create\-dataset
+.SS NAME
+.sp
+groonga\-suggest\-create\-dataset \- Defines schema for a suggestion dataset
+.SS SYNOPSTIS
+.sp
+.nf
+.ft C
+groonga\-suggest\-create\-dataset [options] DATABASE DATASET
+.ft P
+.fi
+.SS DESCTIPION
+.sp
+groonga\-suggest\-create\-dataset creates a dataset for \fB/suggest\fP. A database has many datasets. This command just defines schea for a suggestion dataset.
+.SS OPTIONS
+.sp
+None.
+.SS EXIT STATUS
+.sp
+TODO
+.SS FILES
+.sp
+TODO
+.SS EXAMPLE
+.sp
+TODO
+.SS SEE ALSO
+.sp
+\fB/suggest\fP
+.. : doc:\fIgroonga\-suggest\-httpd\fP
+.. : doc:\fIgroonga\-suggest\-learner\fP
 .SS コマンド
 .sp
 まず、すべてのコマンドに関係するコマンドバージョンについて説明します。その後、組み込みのコマンドを順に説明します。
@@ -4484,100 +5344,366 @@ status
 .ft P
 .fi
 .SS suggest
+.IP Note
+.
+The suggest feature specification isn\(aqt stable. The
+specification may be changed.
+.RE
+.SS NAME
 .sp
-** 未だ仕様は確定しておりません。利用のためには特定のデータを持つテーブルが必要となります。 **
-.SS 名前
-.sp
-suggest \- 指定したクエリに対する補完・修正・提案を行う。
-.SS 書式
+suggest \- returns completion, correction and/or suggestion for a query.
+.SS SYNOPSIS
 .sp
 .nf
 .ft C
-suggest types table column query
+suggest types table column query [sortby [output_columns [offset [limit [frequency_threshold [conditional_probability_threshold [prefix_search]]]]]]]
 .ft P
 .fi
-.SS 説明
+.SS DESCRIPTION
 .sp
-suggestコマンドは、指定したクエリに対して補完・修正・提案を行います。
+The suggest command returns completion, correction and/or
+suggestion for a specified query.
 .sp
-補完とは、クエリの未入力・未変換部分を補うことです。修正とは、クエリ内のミスタイプや誤変換を修正することです。提案とは、さらなる絞込みを行うための候補となるクエリを提案することです。
-.SS 引数
-.sp
-\fBtypes\fP
+See \fB/suggest/introduction\fP about completion,
+correction and suggestion.
+.SS OPTIONS
 .INDENT 0.0
-.INDENT 3.5
+.TP
+.B \fBtypes\fP
 .sp
-補完・修正・提案のいずれを行うかを、パイプ(\(aq|\(aq)で組み合わせたシンボル名で指定します(例: correct|suggest)
+It specifies what types are returned by the suggest
+command.
 .sp
-\fBcomplete\fP
+Here are available types:
+.INDENT 7.0
+.INDENT 3.5
 .INDENT 0.0
+.TP
+.B \fBcomplete\fP
+.sp
+The suggest command does completion.
+.TP
+.B \fBcorrect\fP
+.sp
+The suggest command does correction.
+.TP
+.B \fBsuggest\fP
+.sp
+The suggest command does suggestion.
+.UNINDENT
+.UNINDENT
+.UNINDENT
+.sp
+You can specify one or more types separated by \fB|\fP.
+Here are examples:
+.INDENT 7.0
 .INDENT 3.5
 .sp
-補完を行います。
+It returns correction:
+.sp
+.nf
+.ft C
+correct
+.ft P
+.fi
+.sp
+It returns correction and suggestion:
+.sp
+.nf
+.ft C
+correct|suggest
+.ft P
+.fi
+.sp
+It returns complete, correction and suggestion:
+.sp
+.nf
+.ft C
+complete|correct|suggest
+.ft P
+.fi
 .UNINDENT
 .UNINDENT
+.TP
+.B \fBtable\fP
+.sp
+It specifies table name that has \fBitem_${DATA_SET_NAME}\(ga format.
+For example, \(ga\(gaitem_query\fP is a table name if you created
+dataset by the following command:
+.sp
+.nf
+.ft C
+groonga\-suggest\-create\-dataset /tmp/db\-path query
+.ft P
+.fi
+.TP
+.B \fBcolumn\fP
 .sp
-\fBcorrect\fP
+It specifies a column name that has furigana in
+Katakana in \fBtable\fP table.
+.TP
+.B \fBquery\fP
+.sp
+It specifies query for completion, correction and/or
+suggestion.
+.TP
+.B \fBsortby\fP
+.sp
+It specifies sort key.
+.INDENT 7.0
+.TP
+.B Default:
+.
+\fB\-_score\fP
+.UNINDENT
+.TP
+.B \fBoutput_columns\fP
+.sp
+It specifies output columns.
+.INDENT 7.0
+.TP
+.B Default:
+.
+\fB_key,_score\fP
+.UNINDENT
+.TP
+.B \fBoffset\fP
+.sp
+It specifies returned records offset.
+.INDENT 7.0
+.TP
+.B Default:
+.
+\fB0\fP
+.UNINDENT
+.TP
+.B \fBlimit\fP
+.sp
+It specifies number of returned records.
+.INDENT 7.0
+.TP
+.B Default:
+.
+\fB10\fP
+.UNINDENT
+.TP
+.B \fBfrequency_threshold\fP
+.sp
+It specifies threshold for item frequency. Returned records must have
+\fB_score\fP that is greater than or equal to \fBfrequency_threshold\fP.
+.INDENT 7.0
+.TP
+.B Default:
+.
+\fB100\fP
+.UNINDENT
+.UNINDENT
+.sp
+\fBconditional_probability_threshold\fP
 .INDENT 0.0
 .INDENT 3.5
 .sp
-修正を行います。
+It specifies threshold for conditional
+probability. Conditional probability is used for learned
+data. It is probability of query submission when \fBquery\fP
+is occurred. Returned records must have conditional
+probability that is greater than or equal to
+\fBconditional_probability_threshold\fP.
+.INDENT 0.0
+.TP
+.B Default:
+.
+\fB0.2\fP
+.UNINDENT
 .UNINDENT
 .UNINDENT
-.sp
-\fBsuggest\fP
 .INDENT 0.0
+.TP
+.B \fBprefix_search\fP
+.sp
+It specifies whether optional prefix search is used or not
+in completion.
+.sp
+Here are available values:
+.INDENT 7.0
 .INDENT 3.5
+.INDENT 0.0
+.TP
+.B \fByes\fP
 .sp
-提案を行います。
+Prefix search is always used.
+.TP
+.B \fBno\fP
+.sp
+Prefix search is never used.
+.TP
+.B \fBauto\fP
+.sp
+Prefix search is used only when other search can\(aqt
+find any records.
+.UNINDENT
 .UNINDENT
 .UNINDENT
+.INDENT 7.0
+.TP
+.B Default:
+.
+\fBauto\fP
 .UNINDENT
 .UNINDENT
+.SS RETURN VALUE
+.SS JSON format
 .sp
-\fBtable\fP
+Here is a returned JSON format:
+.sp
+.nf
+.ft C
+{"type1": [["candidate1", score of candidate1],
+           ["candidate2", score of candidate2],
+           ...],
+ "type2": [["candidate1", score of candidate1],
+           ["candidate2", score of candidate2],
+           ...],
+ ...}
+.ft P
+.fi
+.sp
+\fBtype\fP
 .INDENT 0.0
 .INDENT 3.5
 .sp
-基礎となるデータが格納されているテーブル名を指定します。
+A type specified by \fBtypes\fP.
 .UNINDENT
 .UNINDENT
 .sp
-\fBcolumn\fP
+\fBcandidate\fP
 .INDENT 0.0
 .INDENT 3.5
 .sp
-基礎となるデータが格納されているテーブルにおいて、「ヨミガナ」が格納されているカラム名を指定します。
+A candidate for completion, correction or suggestion.
 .UNINDENT
 .UNINDENT
 .sp
-\fBquery\fP
+\fBscore of candidate\fP
 .INDENT 0.0
 .INDENT 3.5
 .sp
-補完・修正・提案の対象となるクエリを指定します。
+A score of corresponding \fBcandidate\fP. It means that
+higher score candidate is more likely candidate for
+completion, correction or suggestion. Returned candidates
+are sorted by \fBscore of candidate\fP descending by
+default.
 .UNINDENT
 .UNINDENT
-.SS 返値
-.SS json形式
+.SS EXAMPLE
+.sp
+Here are learned data for completion.
+.sp
+Execution example:
 .sp
 .nf
 .ft C
-{"タイプ1":[["候補1", 候補1のスコア], ["候補2", 候補2のスコア], ...], ...}
-
-\(ga\(gaタイプ\(ga\(ga
-
-  types引数に指定した値。
-
-\(ga\(ga候補\(ga\(ga
-
-  補完・修正・提案の候補。
-
-\(ga\(ga候補のスコア\(ga\(ga
-
-  補完・修正・提案として優れたものほど高くなる数値。結果は、スコア順にソートされている。
+> load \-\-table event_query \-\-each \(aqsuggest_preparer(_id, type, item, sequence, time, pair_query)\(aq
+> [
+> {"sequence": "1", "time": 1312950803.86057, "item": "e"},
+> {"sequence": "1", "time": 1312950803.96857, "item": "en"},
+> {"sequence": "1", "time": 1312950804.26057, "item": "eng"},
+> {"sequence": "1", "time": 1312950804.56057, "item": "engi"},
+> {"sequence": "1", "time": 1312950804.76057, "item": "engin"},
+> {"sequence": "1", "time": 1312950805.86057, "item": "engine", "type": "submit"}
+> ]
+[[0,1313562872.82845,1.577141877],6]
+.ft P
+.fi
+.sp
+Here are learned data for correction.
+.sp
+Execution example:
+.sp
+.nf
+.ft C
+> load \-\-table event_query \-\-each \(aqsuggest_preparer(_id, type, item, sequence, time, pair_query)\(aq
+> [
+> {"sequence": "2", "time": 1312950803.86057, "item": "s"},
+> {"sequence": "2", "time": 1312950803.96857, "item": "sa"},
+> {"sequence": "2", "time": 1312950804.26057, "item": "sae"},
+> {"sequence": "2", "time": 1312950804.56057, "item": "saer"},
+> {"sequence": "2", "time": 1312950804.76057, "item": "saerc"},
+> {"sequence": "2", "time": 1312950805.76057, "item": "saerch", "type": "submit"},
+> {"sequence": "2", "time": 1312950809.76057, "item": "serch"},
+> {"sequence": "2", "time": 1312950810.86057, "item": "search", "type": "submit"}
+> ]
+[[0,1313562874.6075,2.004191545],8]
+.ft P
+.fi
+.sp
+Here are learned data for suggestion.
+.sp
+Execution example:
+.sp
+.nf
+.ft C
+> load \-\-table event_query \-\-each \(aqsuggest_preparer(_id, type, item, sequence, time, pair_query)\(aq
+> [
+> {"sequence": "3", "time": 1312950803.86057, "item": "search engine", "type": "submit"},
+> {"sequence": "3", "time": 1312950808.86057, "item": "web search realtime", "type": "submit"}
+> ]
+[[0,1313562876.81305,0.801821271],2]
 .ft P
 .fi
+.sp
+Here is a completion example.
+.sp
+Execution example:
+.sp
+.nf
+.ft C
+> suggest \-\-table item_query \-\-column kana \-\-types complete \-\-threshold 1 \-\-query en
+[[0,1313562877.81745,0.001827712],{"complete":[[1],[["_key","ShortText"],["_score","Int32"]],["engine",1]]}]
+.ft P
+.fi
+.sp
+Here is a correction example.
+.sp
+Execution example:
+.sp
+.nf
+.ft C
+> suggest \-\-table item_query \-\-column kana \-\-types correct \-\-threshold 1 \-\-query saerch
+[[0,1313562878.02203,0.000442789],{"correct":[[1],[["_key","ShortText"],["_score","Int32"]],["search",1]]}]
+.ft P
+.fi
+.sp
+Here is a suggestion example.
+.sp
+Execution example:
+.sp
+.nf
+.ft C
+> suggest \-\-table item_query \-\-column kana \-\-types suggest \-\-threshold 1 \-\-query search
+[[0,1313562878.22422,0.000812428],{"suggest":[[2],[["_key","ShortText"],["_score","Int32"]],["search engine",1],["web search realtime",1]]}]
+.ft P
+.fi
+.sp
+Here is a mixed example.
+.sp
+Execution example:
+.sp
+.nf
+.ft C
+> suggest \-\-table item_query \-\-column kana \-\-types complete|correct|suggest \-\-threshold 1 \-\-query search
+[[0,1313562878.42973,0.001223259],{"complete":[[2],[["_key","ShortText"],["_score","Int32"]],["search",2],["search engine",2]],"correct":[[1],[["_key","ShortText"],["_score","Int32"]],["search",2]],"suggest":[[2],[["_key","ShortText"],["_score","Int32"]],["search engine",1],["web search realtime",1]]}]
+.ft P
+.fi
+.SS SEE ALSO
+.INDENT 0.0
+.IP \(bu 2
+.
+\fB/suggest\fP
+.IP \(bu 2
+.
+\fB/executables/groonga\-suggest\-create\-dataset\fP
+.UNINDENT
 .SS table_create
 .SS 名前
 .sp