ICU Analysis Pluginedit

The ICU analysis plugin allows for unicode normalization, collation and folding. The plugin is called elasticsearch-analysis-icu.

The plugin includes the following analysis components:

ICU Normalizationedit

Normalizes characters as explained here. It registers itself by default under icu_normalizer or icuNormalizer using the default settings. Allows for the name parameter to be provided which can include the following values: nfc, nfkc, and nfkc_cf. Here is a sample settings:

{
    "index" : {
        "analysis" : {
            "analyzer" : {
                "normalization" : {
                    "tokenizer" : "keyword",
                    "filter" : ["icu_normalizer"]
                }
            }
        }
    }
}

ICU Foldingedit

Folding of unicode characters based on UTR#30. It registers itself under icu_folding and icuFolding names. The filter also does lowercasing, which means the lowercase filter can normally be left out. Sample setting:

{
    "index" : {
        "analysis" : {
            "analyzer" : {
                "folding" : {
                    "tokenizer" : "keyword",
                    "filter" : ["icu_folding"]
                }
            }
        }
    }
}

Filteringedit

The folding can be filtered by a set of unicode characters with the parameter unicodeSetFilter. This is useful for a non-internationalized search engine where retaining a set of national characters which are primary letters in a specific language is wanted. See syntax for the UnicodeSet here.

The Following example exempts Swedish characters from the folding. Note that the filtered characters are NOT lowercased which is why we add that filter below.

{
    "index" : {
        "analysis" : {
            "analyzer" : {
                "folding" : {
                    "tokenizer" : "standard",
                    "filter" : ["my_icu_folding", "lowercase"]
                }
            }
            "filter" : {
                "my_icu_folding" : {
                    "type" : "icu_folding"
                    "unicodeSetFilter" : "[^åäöÅÄÖ]"
                }
            }
        }
    }
}

ICU Collationedit

Uses collation token filter. Allows to either specify the rules for collation (defined here) using the rules parameter (can point to a location or expressed in the settings, location can be relative to config location), or using the language parameter (further specialized by country and variant). By default registers under icu_collation or icuCollation and uses the default locale.

Here is a sample settings:

{
    "index" : {
        "analysis" : {
            "analyzer" : {
                "collation" : {
                    "tokenizer" : "keyword",
                    "filter" : ["icu_collation"]
                }
            }
        }
    }
}

And here is a sample of custom collation:

{
    "index" : {
        "analysis" : {
            "analyzer" : {
                "collation" : {
                    "tokenizer" : "keyword",
                    "filter" : ["myCollator"]
                }
            },
            "filter" : {
                "myCollator" : {
                    "type" : "icu_collation",
                    "language" : "en"
                }
            }
        }
    }
}

Optionsedit

strength

The strength property determines the minimum level of difference considered significant during comparison.

The default strength for the Collator is tertiary, unless specified otherwise by the locale used to create the Collator. Possible values: primary, secondary, tertiary, quaternary or identical.

See ICU Collation documentation for a more detailed explanation for the specific values.

decomposition

Possible values: no or canonical. Defaults to no. Setting this decomposition property with canonical allows the Collator to handle un-normalized text properly, producing the same results as if the text were normalized. If no is set, it is the user’s responsibility to insure that all text is already in the appropriate form before a comparison or before getting a CollationKey. Adjusting decomposition mode allows the user to select between faster and more complete collation behavior. Since a great many of the world’s languages do not require text normalization, most locales set no as the default decomposition mode.

Expert options:edit

alternate

Possible values: shifted or non-ignorable. Sets the alternate handling for strength quaternary to be either shifted or non-ignorable. What boils down to ignoring punctuation and whitespace.

caseLevel

Possible values: true or false. Default is false. Whether case level sorting is required. When strength is set to primary this will ignore accent differences.

caseFirst

Possible values: lower or upper. Useful to control which case is sorted first when case is not ignored for strength tertiary.

numeric

Possible values: true or false. Whether digits are sorted according to numeric representation. For example the value egg-9 is sorted before the value egg-21. Defaults to false.

variableTop

Single character or contraction. Controls what is variable for alternate.

hiraganaQuaternaryMode

Possible values: true or false. Defaults to false. Distinguishing between Katakana and Hiragana characters in quaternary strength .

ICU Tokenizeredit

Breaks text into words according to UAX #29: Unicode Text Segmentation http://www.unicode.org/reports/tr29/.

{
    "index" : {
        "analysis" : {
            "analyzer" : {
                "collation" : {
                    "tokenizer" : "icu_tokenizer",
                }
            }
        }
    }
}