Caution
This feature is experimental. API will be changed.
highlight_full tags target text. It can use to highlight the search keyword. It can specify use/not use HTML escape, the normalizer name and change the tag for each keyword.
highlight_full has required parameter and optional parameter:
highlight_full(column, normalizer_name, use_html_escape,
keyword1, open_tag1, close_tag1,
...
[keywordN, open_tagN, close_tagN])
Here are a schema definition and sample data to show usage.
Execution example:
table_create Entries TABLE_NO_KEY
# [[0,1407692435.35498,0.0364797115325928],true]
column_create Entries body COLUMN_SCALAR ShortText
# [[0,1407692435.39156,0.0256640911102295],true]
table_create Terms TABLE_PAT_KEY ShortText --default_tokenizer TokenBigram --normalizer NormalizerAuto
# [[0,1407692435.423,0.0274741649627686],true]
column_create Terms document_index COLUMN_INDEX|WITH_POSITION Entries body
# [[0,1407692435.45051,0.0534985065460205],true]
load --table Entries
[
{"body": "Mroonga is a ＭｙＳＱＬ storage engine based on Groonga. <b>Rroonga</b> is a Ruby binding of Groonga."}
]
# [[0,1407692435.50406,0.378907442092896],1]
highlight_full can be used in only --output_columns in select.
highlight_full requires Groonga 4.0.5 or later.
highlight_full requires コマンドバージョン 2 or later.
The following example uses HTML escape and normalzier is NormalizeAuto. It specifies the tags <span class="keyword1"> and </span> of the keyword groonga, and the tags <span class="keyword2"> and </span> of the keyword mysql.
Execution example:
select Entries --output_columns 'highlight_full(body, "NormalizerAuto", true, "Groonga", "<span class=\\"keyword1\\">", "</span>", "mysql", "<span class=\\"keyword2\\">", "</span>")' --command_version 2
# [
# [
# 0,
# 1407695996.52987,
# 0.00151872634887695
# ],
# [
# [
# [
# 1
# ],
# [
# [
# "highlight_full",
# "null"
# ]
# ],
# [
# "Mroonga is a <span class=\"keyword2\">ＭｙＳＱＬ</span> storage engine based on <span class=\"keyword1\">Groonga</span>. <b>Rroonga</b> is a Ruby binding of <span class=\"keyword1\">Groonga</span>."
# ]
# ]
# ]
# ]
The text are scanned by the keywords for tagging after they are normalized by NormalizerAuto normalizer.
--query "groonga mysql" matches to the first record's body. highight_full surrounds the keywords groonga contained in the text with <span class="keyword1"> and </span>, and the keywords mysql contained in the text with with <span class="keyword2"> and </span>.
Special characters such as < and > are escapsed as < and >.
You can specify string literal instead of column.
Execution example:
select Entries --output_columns 'highlight_full("Groonga is very fast fulltext search engine.", "NormalizerAuto", true, "Groonga", "<span class=\\"keyword1\\">", "</span>", "mysql", "<span class=\\"keyword2\\">", "</span>")' --command_version 2 --match_columns body --query "groonga"
# [
# [
# 0,
# 1407696157.1849,
# 0.00164437294006348
# ],
# [
# [
# [
# 1
# ],
# [
# [
# "highlight_full",
# "null"
# ]
# ],
# [
# "<span class=\"keyword1\">Groonga</span> is very fast fulltext search engine."
# ]
# ]
# ]
# ]
There are three required parameters, column, normalizer_name and use_html_escape. There are three or over optional parameters, keywordN, open_tagN and end_tagN.
Specifies a column of the table.
Specifies a normalizer name.
Specifies use or not use HTML escape. If it is true , use HTML escape. If it is false , not use HTML escape.
Specifies a keyword for tagging. You can specify multiple keywords for each three arguments.
Specifies a open tag. You can specify multiple open tags for each three arguments.
Specifies a close tag. You can specify multiple close tags for each three arguments.
highlight_full returns a tagged string or null. If highlight_full can't find any keywords, it returns null.