From f6a38f40dc52c4517e41ddb381278ecf5efba056 Mon Sep 17 00:00:00 2001 From: toasted-nutbread Date: Sun, 10 Jan 2021 14:43:06 -0500 Subject: Customizable sentence parsing (#1217) * Add new sentenceParsing options * Update TextScanner.setOptions * Assign terminator/quote maps * Pass sentence parsing info to extractSentence * Simplify setting * Add setting for enableTerminationCharacters * Create new settings for sentence termination characters --- test/test-options-util.js | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) (limited to 'test/test-options-util.js') diff --git a/test/test-options-util.js b/test/test-options-util.js index 8b7254c6..5ce9313e 100644 --- a/test/test-options-util.js +++ b/test/test-options-util.js @@ -420,7 +420,22 @@ function createProfileOptionsUpdatedTestData1() { fieldTemplates: null }, sentenceParsing: { - scanExtent: 200 + scanExtent: 200, + enableTerminationCharacters: true, + terminationCharacters: [ + {enabled: true, character1: '「', character2: '」', includeCharacterAtStart: false, includeCharacterAtEnd: false}, + {enabled: true, character1: '『', character2: '』', includeCharacterAtStart: false, includeCharacterAtEnd: false}, + {enabled: true, character1: '"', character2: '"', includeCharacterAtStart: false, includeCharacterAtEnd: false}, + {enabled: true, character1: '\'', character2: '\'', includeCharacterAtStart: false, includeCharacterAtEnd: false}, + {enabled: true, character1: '.', character2: null, includeCharacterAtStart: false, includeCharacterAtEnd: true}, + {enabled: true, character1: '!', character2: null, includeCharacterAtStart: false, includeCharacterAtEnd: true}, + {enabled: true, character1: '?', character2: null, includeCharacterAtStart: false, includeCharacterAtEnd: true}, + {enabled: true, character1: '.', character2: null, includeCharacterAtStart: false, includeCharacterAtEnd: true}, + {enabled: true, character1: '。', character2: null, includeCharacterAtStart: false, includeCharacterAtEnd: true}, + {enabled: true, character1: '!', character2: null, includeCharacterAtStart: false, includeCharacterAtEnd: true}, + {enabled: true, character1: '?', character2: null, includeCharacterAtStart: false, includeCharacterAtEnd: true}, + {enabled: true, character1: '…', character2: null, includeCharacterAtStart: false, includeCharacterAtEnd: true} + ] } }; } -- cgit v1.2.3