text-icu- Bindings to the ICU library

Safe HaskellNone




Commonly used functions for Unicode, implemented as bindings to the International Components for Unicode (ICU) libraries.

This module contains only the most commonly used types and functions. Other modules in this package expose richer interfaces.


Data representation

The Haskell Text type is implemented as an array in the Haskell heap. This means that its location is not pinned; it may be copied during a garbage collection pass. ICU, on the other hand, works with strings that are allocated in the normal system heap and have a fixed address.

To accommodate this need, these bindings use the functions from Foreign to copy data between the Haskell heap and the system heap. The copied strings are still managed automatically, but the need to duplicate data does add some performance and memory overhead.


data LocaleName Source

The name of a locale.



The root locale. For a description of resource bundles and the root resource, see http://userguide.icu-project.org/locale/resources.

Locale String

A specific locale.


The program's current locale.

Boundary analysis

Text boundary analysis is the process of locating linguistic boundaries while formatting and handling text. Examples of this process include:

  • Locating appropriate points to word-wrap text to fit within specific margins while displaying or printing.
  • Counting characters, words, sentences, or paragraphs.
  • Making a list of the unique words in a document.
  • Figuring out if a given range of text contains only whole words.
  • Capitalizing the first letter of each word.
  • Locating a particular unit of the text (For example, finding the third word in the document).

The Breaker type was designed to support these kinds of tasks.

For the impure boundary analysis API (which is richer, but less easy to use than the pure API), see the Data.Text.ICU.Break module. The impure API supports some uses that may be less efficient via the pure API, including:

  • Locating the beginning of a word that the user has selected.
  • Determining how far to move the text cursor when the user hits an arrow key (Some characters require more than one position in the text store and some characters in the text store do not display at all).

data Breaker a Source

A boundary analyser.

data Break a Source

A break in a string.


Eq a => Eq (Break a) 
Show a => Show (Break a) 

brkPrefix :: Break a -> TextSource

Prefix of the current break.

brkBreak :: Break a -> TextSource

Text of the current break.

brkSuffix :: Break a -> TextSource

Suffix of the current break.

brkStatus :: Break a -> aSource

Status of the current break (only meaningful if Line or Word).

data Line Source

Line break status.



A soft line break is a position at which a line break is acceptable, but not required.



data Word Source

Word break status.



A "word" that does not fit into another category. Includes spaces and most punctuation.


A word that appears to be a number.


A word containing letters, excluding hiragana, katakana or ideographic characters.


A word containing kana characters.


A word containing ideographic characters.


breakCharacter :: LocaleName -> Breaker ()Source

Break a string on character boundaries.

Character boundary analysis identifies the boundaries of Extended Grapheme Clusters, which are groupings of codepoints that should be treated as character-like units for many text operations. Please see Unicode Standard Annex #29, Unicode Text Segmentation, http://www.unicode.org/reports/tr29/ for additional information on grapheme clusters and guidelines on their use.

breakLine :: LocaleName -> Breaker LineSource

Break a string on line boundaries.

Line boundary analysis determines where a text string can be broken when line wrapping. The mechanism correctly handles punctuation and hyphenated words.

breakSentence :: LocaleName -> Breaker ()Source

Break a string on sentence boundaries.

Sentence boundary analysis allows selection with correct interpretation of periods within numbers and abbreviations, and trailing punctuation marks such as quotation marks and parentheses.

breakWord :: LocaleName -> Breaker WordSource

Break a string on word boundaries.

Word boundary analysis is used by search and replace functions, as well as within text editing applications that allow the user to select words with a double click. Word selection provides correct interpretation of punctuation marks within and following words. Characters that are not part of a word, such as symbols or punctuation marks, have word breaks on both sides.

breaks :: Breaker a -> Text -> [Break a]Source

Return a list of all breaks in a string, from left to right.

breaksRight :: Breaker a -> Text -> [Break a]Source

Return a list of all breaks in a string, from right to left.

Case mapping



:: Bool

Whether to include or exclude mappings for dotted and dotless I and i that are marked with I in CaseFolding.txt.

-> Text 
-> Text 

Case-fold the characters in a string.

Case folding is locale independent and not context sensitive, but there is an option for treating the letter I specially for Turkic languages. The result may be longer or shorter than the original.

toLower :: LocaleName -> Text -> TextSource

Lowercase the characters in a string.

Casing is locale dependent and context sensitive. The result may be longer or shorter than the original.

toUpper :: LocaleName -> Text -> TextSource

Uppercase the characters in a string.

Casing is locale dependent and context sensitive. The result may be longer or shorter than the original.


data CharIterator Source

A type that supports efficient iteration over Unicode characters.

As an example of where this may be useful, a function using this type may be able to iterate over a UTF-8 ByteString directly, rather than first copying and converting it to an intermediate form. This type also allows e.g. comparison between Text and ByteString, with minimal overhead.

fromString :: String -> CharIteratorSource

Construct a CharIterator from a Unicode string.

fromText :: Text -> CharIteratorSource

Construct a CharIterator from a Unicode string.

fromUtf8 :: ByteString -> CharIteratorSource

Construct a CharIterator from a Unicode string encoded as a UTF-8 ByteString. The validity of the encoded string is *not* checked.


data NormalizationMode Source

Normalization modes.



No decomposition/composition.


Canonical decomposition.


Compatibility decomposition.


Canonical decomposition followed by canonical composition.


Compatibility decomposition followed by canonical composition.


"Fast C or D" form.

normalize :: NormalizationMode -> Text -> TextSource

Normalize a string according the specified normalization mode.

quickCheck :: NormalizationMode -> Text -> Maybe BoolSource

Perform an efficient check on a string, to quickly determine if the string is in a particular normalization form.

A Nothing result indicates that a definite answer could not be determined quickly, and a more thorough check is required, e.g. with isNormalized. The user may have to convert the string to its normalized form and compare the results.

A result of Just True or Just False indicates that the string definitely is, or is not, in the given normalization form.

isNormalized :: NormalizationMode -> Text -> BoolSource

Indicate whether a string is in a given normalization form.

Unlike quickCheck, this function returns a definitive result. For NFD, NFKD, and FCD normalization forms, both functions work in exactly the same ways. For NFC and NFKC forms, where quickCheck may return Nothing, this function will perform further tests to arrive at a definitive result.

String comparison

Normalization-sensitive string comparison

data CompareOption Source

Options to compare.



The caller knows that both strings fulfill the FCD conditions. If not set, compare will quickCheck for FCD and normalize if necessary.


Compare strings case-insensitively using case folding, instead of case-sensitively. If set, then the following case folding options are used.


When case folding, exclude the special I character. For use with Turkic (Turkish/Azerbaijani) text data.

compare :: [CompareOption] -> Text -> Text -> OrderingSource

Compare two strings for canonical equivalence. Further options include case-insensitive comparison and code point order (as opposed to code unit order).

Canonical equivalence between two strings is defined as their normalized forms (NFD or NFC) being identical. This function compares strings incrementally instead of normalizing (and optionally case-folding) both strings entirely, improving performance significantly.

Bulk normalization is only necessary if the strings do not fulfill the FCD conditions. Only in this case, and only if the strings are relatively long, is memory allocated temporarily. For FCD strings and short non-FCD strings there is no memory allocation.

Locale-sensitive string collation

For the impure collation API (which is richer, but less easy to use than the pure API), see the Data.Text.ICU.Collate module.

data Collator Source

String collator type. Collators are considered equal if they will sort strings identically.

collator :: LocaleName -> CollatorSource

Create an immutable Collator for comparing strings.

If Root is passed as the locale, UCA collation rules will be used.

collatorWith :: LocaleName -> [Attribute] -> CollatorSource

Create an immutable Collator with the given Attributes.

collate :: Collator -> Text -> Text -> OrderingSource

Compare two strings.

collateIter :: Collator -> CharIterator -> CharIterator -> OrderingSource

Compare two CharIterators.

If either iterator was constructed from a ByteString, it does not need to be copied or converted beforehand, so this function can be quite cheap.

sortKey :: Collator -> Text -> ByteStringSource

Create a key for sorting the Text using the given Collator. The result of comparing two ByteStrings that have been transformed with sortKey will be the same as the result of collate on the two untransformed Texts.

uca :: CollatorSource

A Collator that uses the Unicode Collation Algorithm (UCA).

Regular expressions

data MatchOption Source

Options for controlling matching behaviour.



Enable case insensitive matching.


Allow comments and white space within patterns.


If set, '.' matches line terminators. Otherwise '.' matching stops at line end.


If set, treat the entire pattern as a literal string. Metacharacters or escape sequences in the input sequence will be given no special meaning.

The option CaseInsensitive retains its meanings on matching when used in conjunction with this option. Other options become superfluous.


Control behaviour of '$' and '^'. If set, recognize line terminators within string, Otherwise, match only at start and end of input string.


Haskell-only line endings. When this mode is enabled, only '\n' is recognized as a line ending in the behavior of '.', '^', and '$'.


Unicode word boundaries. If set, '\\b' uses the Unicode TR 29 definition of word boundaries.

Warning: Unicode word boundaries are quite different from traditional regular expression word boundaries. See http://unicode.org/reports/tr29/#Word_Boundaries.


Throw an error on unrecognized backslash escapes. If set, fail with an error on patterns that contain backslash-escaped ASCII letters without a known special meaning. If this flag is not set, these escaped letters represent themselves.

WorkLimit Int

Set a processing limit for match operations.

Some patterns, when matching certain strings, can run in exponential time. For practical purposes, the match operation may appear to be in an infinite loop. When a limit is set a match operation will fail with an error if the limit is exceeded.

The units of the limit are steps of the match engine. Correspondence with actual processor time will depend on the speed of the processor and the details of the specific pattern, but will typically be on the order of milliseconds.

By default, the matching time is not limited.

StackLimit Int

Set the amount of heap storage avaliable for use by the match backtracking stack.

ICU uses a backtracking regular expression engine, with the backtrack stack maintained on the heap. This function sets the limit to the amount of memory that can be used for this purpose. A backtracking stack overflow will result in an error from the match operation that caused it.

A limit is desirable because a malicious or poorly designed pattern can use excessive memory, potentially crashing the process. A limit is enabled by default.

data ParseError Source

Detailed information about parsing errors. Used by ICU parsing engines that parse long rules, patterns, or programs, where the text being parsed is long enough that more information than an ICUError is needed to localize the error.

data Match Source

A match for a regular expression.


data Regex Source

A compiled regular expression.

Regex values are usually constructed using the regex or regex' functions. This type is also an instance of IsString, so if you have the OverloadedStrings language extension enabled, you can construct a Regex by simply writing the pattern in quotes (though this does not allow you to specify any Options).

class Regular r Source

A typeclass for functions common to both Match and Regex types.


regex :: [MatchOption] -> Text -> RegexSource

Compile a regular expression with the given options. This function throws a ParseError if the pattern is invalid, so it is best for use when the pattern is statically known.

regex' :: [MatchOption] -> Text -> Either ParseError RegexSource

Compile a regular expression with the given options. This is safest to use when the pattern is constructed at run time.


pattern :: Regular r => r -> TextSource

Return the source form of the pattern used to construct this regular expression or match.


find :: Regex -> Text -> Maybe MatchSource

Find the first match for the regular expression in the given text.

findAll :: Regex -> Text -> [Match]Source

Lazily find all matches for the regular expression in the given text.

Match groups

Capturing groups are numbered starting from zero. Group zero is always the entire matching text. Groups greater than zero contain the text matching each capturing group in a regular expression.

groupCount :: Regular r => r -> IntSource

Return the number of capturing groups in this regular expression or match's pattern.

unfold :: (Int -> Match -> Maybe Text) -> Match -> [Text]Source

A combinator for returning a list of all capturing groups on a Match.

span :: Match -> TextSource

Return the span of text between the end of the previous match and the beginning of the current match.

group :: Int -> Match -> Maybe TextSource

Return the nth capturing group in a match, or Nothing if n is out of bounds.

prefix :: Int -> Match -> Maybe TextSource

Return the prefix of the nth capturing group in a match (the text from the start of the string to the start of the match), or Nothing if n is out of bounds.

suffix :: Int -> Match -> Maybe TextSource

Return the suffix of the nth capturing group in a match (the text from the end of the match to the end of the string), or Nothing if n is out of bounds.