Skip to content
This repository has been archived by the owner on Feb 25, 2023. It is now read-only.

Commit

Permalink
Customizable sentence parsing (#1217)
Browse files Browse the repository at this point in the history 
* 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
  • Loading branch information
@toasted-nutbread
toasted-nutbread authored Jan 10, 2021
1 parent 083da93 commit f6a38f4
Show file tree
Hide file tree
Showing 12 changed files with 638 additions and 62 deletions.
75 changes: 75 additions & 0 deletions ext/bg/css/settings2.css
View file
Original file line number Diff line number Diff line change
Expand Up @@ -1856,6 +1856,81 @@ input.translation-text-replacement-test-output {
display: none;
}

.sentence-termination-character-list-table {
width: 100%;
table-layout: fixed;
border-spacing: 0.25em;
margin-top: 0.5em;
min-width: 400px;
}
.sentence-termination-character-list-table thead td {
white-space: nowrap;
font-size: var(--font-size-small);
line-height: 1;
text-align: left;
vertical-align: bottom;
font-weight: normal;
}
.sentence-termination-character-list-table td {
vertical-align: middle;
padding: 0;
}
.sentence-termination-character-list-table td:nth-child(1) {
width: 2em;
}
.sentence-termination-character-list-table td:nth-child(2) {
width: 4em;
}
.sentence-termination-character-list-table td:nth-child(3) {
width: 25%;
}
.sentence-termination-character-list-table td:nth-child(4) {
width: 18.5%;
}
.sentence-termination-character-list-table td:nth-child(5) {
width: 18.5%;
}
.sentence-termination-character-list-table td:nth-child(6) {
width: 40%;
}
.sentence-termination-character-list-table td:nth-child(7) {
width: 3em;
}
select.sentence-termination-character-type,
input.sentence-termination-character-input1,
input.sentence-termination-character-input2 {
width: 100%;
}
.sentence-termination-character-input2-alt {
text-align: center;
}
.sentence-termination-character-entry:not([data-type=quote]) .sentence-termination-character-input2 {
display: none;
}
.sentence-termination-character-entry[data-type=quote] .sentence-termination-character-input2-alt {
display: none;
}
.sentence-termination-character-include-list {
display: flex;
flex-flow: row nowrap;
}
.sentence-termination-character-include {
display: flex;
flex-flow: row nowrap;
white-space: nowrap;
align-items: center;
}
.sentence-termination-character-include>:first-child {
margin-right: 0.375em;
}
.sentence-termination-character-include+.sentence-termination-character-include {
margin-left: 1em;
}
#sentence-termination-character-list-empty {
flex: 0 1 auto;
margin-top: 0.5em;
}


/* Generic layouts */
.margin-above {
Expand Down
61 changes: 60 additions & 1 deletion ext/bg/data/options-schema.json
View file
Original file line number Diff line number Diff line change
Expand Up @@ -846,13 +846,72 @@
"sentenceParsing": {
"type": "object",
"required": [
"scanExtent"
"scanExtent",
"enableTerminationCharacters",
"terminationCharacters"
],
"properties": {
"scanExtent": {
"type": "integer",
"minimum": 0,
"default": 200
},
"enableTerminationCharacters": {
"type": "boolean",
"default": true
},
"terminationCharacters": {
"type": "array",
"items": {
"type": "object",
"required": [
"enabled",
"character1",
"character2",
"includeCharacterAtStart",
"includeCharacterAtEnd"
],
"properties": {
"enabled": {
"type": "boolean",
"default": true
},
"character1": {
"type": "string",
"default": "\"",
"minLength": 1,
"maxLength": 1
},
"character2": {
"type": ["string", "null"],
"default": "\"",
"minLength": 1,
"maxLength": 1
},
"includeCharacterAtStart": {
"type": "boolean",
"default": false
},
"includeCharacterAtEnd": {
"type": "boolean",
"default": false
}
}
},
"default": [
{"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}
]
}
}
}
Expand Down
19 changes: 18 additions & 1 deletion ext/bg/js/options.js
View file
Original file line number Diff line number Diff line change
Expand Up @@ -684,13 +684,30 @@ class OptionsUtil {
// Version 8 changes:
// Added translation.textReplacements.
// Moved anki.sentenceExt to sentenceParsing.scanExtent.
// Added sentenceParsing.enableTerminationCharacters.
// Added sentenceParsing.terminationCharacters.
for (const profile of options.profiles) {
profile.options.translation.textReplacements = {
searchOriginal: true,
groups: []
};
profile.options.sentenceParsing = {
scanExtent: profile.options.anki.sentenceExt
scanExtent: profile.options.anki.sentenceExt,
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}
]
};
delete profile.options.anki.sentenceExt;
}
Expand Down
Loading

0 comments on commit f6a38f4

Please sign in to comment.