[apple/icu.git] / icuSources / test / testdata / break_rules / readme.txt

file: testdata/break_rules/readme.txt
Copyright (C) 2016 and later: Unicode, Inc. and others.
License & terms of use: http://www.unicode.org/copyright.html#License

Copyright (c) 2015-2016, International Business Machines Corporation and others. All Rights Reserved.

This directory contains the break iterator reference rule files used by intltest rbbi/RBBIMonkeyTest/testMonkey.
The rules in this directory track the boundary rules from Unicode UAX 14 and 29. They are interpretted
to provide an expected set of boundary positions to compare with the results from ICU break iteration.

Each set of reference break rules lives in a separate file.
The list of rule files to run by default is hardcoded into the test code, in rbbimonkeytest.cpp.

Each test file includes
  - The type of ICU break interator to create (word, line, sentence, etc.)
  - The locale to use
  - Character Class definitions
  - Rule definitions

To Do
  - Syntax for tailoring.


Character Class Definition: 
    name = set_regular_expression;

Rule Definition:
    rule_regular_expression;

name:
    [A-Za-z_][A-Za-z0-9_]*

set_regular_expression:
    The intersection of an ICU regular expression [set] expression and a UnicodeSet pattern.
    (They are mostly the same)
    May include previously defined set names, which are logically expanded in-place.

rule_regular_expresson:
    An ICU Regular Expression.
    May include set names, which are logically expanded in-place.
    May include a '÷', which defines a boundary position.

Application of the rules:
    Matching begins at the start of text, or after a previously identified boundary.
    The pseudo-code below finds the next boundary.

    while position < end of text
        for each rule
            if the text at position matches this rule
                if the rule has a '÷'
                    Boundary is found.
                    return the position of the '÷' within the match.
                else
                    position = last character of the rule match.
                    break from the rule loop, continue the outer loop.

    This differs from the Unicode UAX algorithm in that each position in the text is
    not tested separately. Instead, when a rule match is found, rule application restarts with the last
    character of the preceding rule match. ICU's break rules also operate this way.

    Expressing rules this way simplifies UAX rules that have leading or trailing context; it
    is no longer necessary to write expressions that match the context starting from
    any position within it.

    This rule form differs from ICU rules in that the rules are applied sequentially, as they
    are with the Unicode UAX rules. With the main ICU break rules, all are applied in parallel.

Word Dictionaries
    The monkey test does not test dictionary based breaking. The set named 'dicitionary' is special,
    as it is in the main ICU rules. For the monkey test, no characters from the dictionary set are
    included in the randomly-generated test data.
Commit	Line	Data
2ca993e8	1	file: testdata/break_rules/readme.txt
f3c0d7a5 A	2	Copyright (C) 2016 and later: Unicode, Inc. and others.
	3	License & terms of use: http://www.unicode.org/copyright.html#License
	4
2ca993e8 A	5	Copyright (c) 2015-2016, International Business Machines Corporation and others. All Rights Reserved.
	6
	7	This directory contains the break iterator reference rule files used by intltest rbbi/RBBIMonkeyTest/testMonkey.
	8	The rules in this directory track the boundary rules from Unicode UAX 14 and 29. They are interpretted
	9	to provide an expected set of boundary positions to compare with the results from ICU break iteration.
	10
	11	Each set of reference break rules lives in a separate file.
	12	The list of rule files to run by default is hardcoded into the test code, in rbbimonkeytest.cpp.
	13
	14	Each test file includes
	15	- The type of ICU break interator to create (word, line, sentence, etc.)
	16	- The locale to use
	17	- Character Class definitions
	18	- Rule definitions
	19
	20	To Do
	21	- Syntax for tailoring.
	22
	23
	24	Character Class Definition:
	25	name = set_regular_expression;
	26
	27	Rule Definition:
	28	rule_regular_expression;
	29
	30	name:
	31	[A-Za-z_][A-Za-z0-9_]*
	32
	33	set_regular_expression:
	34	The intersection of an ICU regular expression [set] expression and a UnicodeSet pattern.
	35	(They are mostly the same)
	36	May include previously defined set names, which are logically expanded in-place.
	37
	38	rule_regular_expresson:
	39	An ICU Regular Expression.
	40	May include set names, which are logically expanded in-place.
	41	May include a '÷', which defines a boundary position.
	42
	43	Application of the rules:
	44	Matching begins at the start of text, or after a previously identified boundary.
	45	The pseudo-code below finds the next boundary.
	46
	47	while position < end of text
	48	for each rule
	49	if the text at position matches this rule
	50	if the rule has a '÷'
	51	Boundary is found.
	52	return the position of the '÷' within the match.
	53	else
	54	position = last character of the rule match.
	55	break from the rule loop, continue the outer loop.
	56
	57	This differs from the Unicode UAX algorithm in that each position in the text is
	58	not tested separately. Instead, when a rule match is found, rule application restarts with the last
	59	character of the preceding rule match. ICU's break rules also operate this way.
	60
	61	Expressing rules this way simplifies UAX rules that have leading or trailing context; it
	62	is no longer necessary to write expressions that match the context starting from
	63	any position within it.
	64
	65	This rule form differs from ICU rules in that the rules are applied sequentially, as they
	66	are with the Unicode UAX rules. With the main ICU break rules, all are applied in parallel.
	67
	68	Word Dictionaries
69	The monkey test does not test dictionary based breaking. The set named 'dicitionary' is special,
70	as it is in the main ICU rules. For the monkey test, no characters from the dictionary set are
71	included in the randomly-generated test data.
72