chaffjs 0.13.0 → 0.15.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/dist/address-chain.d.ts +4 -1
- package/dist/address-chain.d.ts.map +1 -1
- package/dist/address-chain.js +5 -2
- package/dist/address-chain.js.map +1 -1
- package/dist/cli-text.d.ts +4 -1
- package/dist/cli-text.d.ts.map +1 -1
- package/dist/cli-text.js +14 -2
- package/dist/cli-text.js.map +1 -1
- package/dist/cli.d.ts.map +1 -1
- package/dist/cli.js +21 -21
- package/dist/cli.js.map +1 -1
- package/dist/commands/eval.d.ts +4 -0
- package/dist/commands/eval.d.ts.map +1 -1
- package/dist/commands/eval.js +20 -6
- package/dist/commands/eval.js.map +1 -1
- package/dist/commands/feedback.d.ts.map +1 -1
- package/dist/commands/feedback.js +2 -1
- package/dist/commands/feedback.js.map +1 -1
- package/dist/commands/test.d.ts.map +1 -1
- package/dist/commands/test.js +3 -2
- package/dist/commands/test.js.map +1 -1
- package/dist/commands/tree.d.ts.map +1 -1
- package/dist/commands/tree.js +2 -1
- package/dist/commands/tree.js.map +1 -1
- package/dist/detectors/acronym-compound.d.ts +8 -0
- package/dist/detectors/acronym-compound.d.ts.map +1 -0
- package/dist/detectors/acronym-compound.js +8 -0
- package/dist/detectors/acronym-compound.js.map +1 -0
- package/dist/detectors/acronym-context.d.ts +6 -0
- package/dist/detectors/acronym-context.d.ts.map +1 -1
- package/dist/detectors/acronym-context.js +24 -1
- package/dist/detectors/acronym-context.js.map +1 -1
- package/dist/detectors/acronym-expansion.d.ts +10 -0
- package/dist/detectors/acronym-expansion.d.ts.map +1 -1
- package/dist/detectors/acronym-expansion.js +63 -5
- package/dist/detectors/acronym-expansion.js.map +1 -1
- package/dist/detectors/char-shape.d.ts.map +1 -1
- package/dist/detectors/char-shape.js +38 -16
- package/dist/detectors/char-shape.js.map +1 -1
- package/dist/detectors/citation-key.d.ts +13 -0
- package/dist/detectors/citation-key.d.ts.map +1 -0
- package/dist/detectors/citation-key.js +9 -0
- package/dist/detectors/citation-key.js.map +1 -0
- package/dist/detectors/cited-name.d.ts +11 -0
- package/dist/detectors/cited-name.d.ts.map +1 -0
- package/dist/detectors/cited-name.js +15 -0
- package/dist/detectors/cited-name.js.map +1 -0
- package/dist/detectors/concrete-evidence.d.ts +7 -0
- package/dist/detectors/concrete-evidence.d.ts.map +1 -0
- package/dist/detectors/concrete-evidence.js +21 -0
- package/dist/detectors/concrete-evidence.js.map +1 -0
- package/dist/detectors/conjugated-form.d.ts +3 -0
- package/dist/detectors/conjugated-form.d.ts.map +1 -0
- package/dist/detectors/conjugated-form.js +29 -0
- package/dist/detectors/conjugated-form.js.map +1 -0
- package/dist/detectors/counter-edge.d.ts +8 -0
- package/dist/detectors/counter-edge.d.ts.map +1 -0
- package/dist/detectors/counter-edge.js +27 -0
- package/dist/detectors/counter-edge.js.map +1 -0
- package/dist/detectors/dotted-name.d.ts +13 -0
- package/dist/detectors/dotted-name.d.ts.map +1 -0
- package/dist/detectors/dotted-name.js +10 -0
- package/dist/detectors/dotted-name.js.map +1 -0
- package/dist/detectors/doubled-word.d.ts +6 -1
- package/dist/detectors/doubled-word.d.ts.map +1 -1
- package/dist/detectors/doubled-word.js +22 -2
- package/dist/detectors/doubled-word.js.map +1 -1
- package/dist/detectors/en-shape.d.ts +1 -1
- package/dist/detectors/en-shape.d.ts.map +1 -1
- package/dist/detectors/en-shape.js +11 -22
- package/dist/detectors/en-shape.js.map +1 -1
- package/dist/detectors/enumerated-runs.d.ts +22 -0
- package/dist/detectors/enumerated-runs.d.ts.map +1 -0
- package/dist/detectors/enumerated-runs.js +53 -0
- package/dist/detectors/enumerated-runs.js.map +1 -0
- package/dist/detectors/gram-place.d.ts +0 -4
- package/dist/detectors/gram-place.d.ts.map +1 -1
- package/dist/detectors/gram-place.js +6 -1
- package/dist/detectors/gram-place.js.map +1 -1
- package/dist/detectors/gram-predicate.d.ts +7 -0
- package/dist/detectors/gram-predicate.d.ts.map +1 -0
- package/dist/detectors/gram-predicate.js +44 -0
- package/dist/detectors/gram-predicate.js.map +1 -0
- package/dist/detectors/heading-case.d.ts +21 -0
- package/dist/detectors/heading-case.d.ts.map +1 -0
- package/dist/detectors/heading-case.js +38 -0
- package/dist/detectors/heading-case.js.map +1 -0
- package/dist/detectors/heading-echo.d.ts.map +1 -1
- package/dist/detectors/heading-echo.js +4 -1
- package/dist/detectors/heading-echo.js.map +1 -1
- package/dist/detectors/lexicon.d.ts.map +1 -1
- package/dist/detectors/lexicon.js +7 -4
- package/dist/detectors/lexicon.js.map +1 -1
- package/dist/detectors/middle-dot.d.ts +6 -0
- package/dist/detectors/middle-dot.d.ts.map +1 -0
- package/dist/detectors/middle-dot.js +9 -0
- package/dist/detectors/middle-dot.js.map +1 -0
- package/dist/detectors/name-numeral.d.ts +16 -0
- package/dist/detectors/name-numeral.d.ts.map +1 -0
- package/dist/detectors/name-numeral.js +30 -0
- package/dist/detectors/name-numeral.js.map +1 -0
- package/dist/detectors/name-part.d.ts +8 -0
- package/dist/detectors/name-part.d.ts.map +1 -0
- package/dist/detectors/name-part.js +20 -0
- package/dist/detectors/name-part.js.map +1 -0
- package/dist/detectors/name-run.d.ts +8 -0
- package/dist/detectors/name-run.d.ts.map +1 -0
- package/dist/detectors/name-run.js +13 -0
- package/dist/detectors/name-run.js.map +1 -0
- package/dist/detectors/orthography.d.ts.map +1 -1
- package/dist/detectors/orthography.js +18 -6
- package/dist/detectors/orthography.js.map +1 -1
- package/dist/detectors/placeholder.d.ts +14 -0
- package/dist/detectors/placeholder.d.ts.map +1 -0
- package/dist/detectors/placeholder.js +24 -0
- package/dist/detectors/placeholder.js.map +1 -0
- package/dist/detectors/polite-word.d.ts +4 -0
- package/dist/detectors/polite-word.d.ts.map +1 -0
- package/dist/detectors/polite-word.js +10 -0
- package/dist/detectors/polite-word.js.map +1 -0
- package/dist/detectors/quoted-variant.d.ts +7 -0
- package/dist/detectors/quoted-variant.d.ts.map +1 -0
- package/dist/detectors/quoted-variant.js +18 -0
- package/dist/detectors/quoted-variant.js.map +1 -0
- package/dist/detectors/register.d.ts +7 -5
- package/dist/detectors/register.d.ts.map +1 -1
- package/dist/detectors/register.js +6 -4
- package/dist/detectors/register.js.map +1 -1
- package/dist/detectors/roman-numeral.d.ts +6 -0
- package/dist/detectors/roman-numeral.d.ts.map +1 -0
- package/dist/detectors/roman-numeral.js +6 -0
- package/dist/detectors/roman-numeral.js.map +1 -0
- package/dist/detectors/sentence-ending.d.ts.map +1 -1
- package/dist/detectors/sentence-ending.js +14 -4
- package/dist/detectors/sentence-ending.js.map +1 -1
- package/dist/detectors/series-label.d.ts +13 -0
- package/dist/detectors/series-label.d.ts.map +1 -0
- package/dist/detectors/series-label.js +26 -0
- package/dist/detectors/series-label.js.map +1 -0
- package/dist/detectors/signals.d.ts.map +1 -1
- package/dist/detectors/signals.js +37 -25
- package/dist/detectors/signals.js.map +1 -1
- package/dist/detectors/structure-tree.js +1 -1
- package/dist/detectors/structure-tree.js.map +1 -1
- package/dist/detectors/structure.d.ts.map +1 -1
- package/dist/detectors/structure.js +2 -1
- package/dist/detectors/structure.js.map +1 -1
- package/dist/detectors/superlative-name.d.ts +9 -0
- package/dist/detectors/superlative-name.d.ts.map +1 -0
- package/dist/detectors/superlative-name.js +24 -0
- package/dist/detectors/superlative-name.js.map +1 -0
- package/dist/detectors/token-shape.d.ts.map +1 -1
- package/dist/detectors/token-shape.js +10 -2
- package/dist/detectors/token-shape.js.map +1 -1
- package/dist/document.d.ts.map +1 -1
- package/dist/document.js +10 -4
- package/dist/document.js.map +1 -1
- package/dist/genre-check.d.ts +34 -0
- package/dist/genre-check.d.ts.map +1 -0
- package/dist/genre-check.js +21 -0
- package/dist/genre-check.js.map +1 -0
- package/dist/genre.d.ts +8 -1
- package/dist/genre.d.ts.map +1 -1
- package/dist/genre.js +29 -5
- package/dist/genre.js.map +1 -1
- package/dist/init.d.ts +0 -1
- package/dist/init.d.ts.map +1 -1
- package/dist/init.js +0 -12
- package/dist/init.js.map +1 -1
- package/dist/joined-view.d.ts +25 -0
- package/dist/joined-view.d.ts.map +1 -0
- package/dist/joined-view.js +76 -0
- package/dist/joined-view.js.map +1 -0
- package/dist/line-continues.d.ts +6 -0
- package/dist/line-continues.d.ts.map +1 -0
- package/dist/line-continues.js +33 -0
- package/dist/line-continues.js.map +1 -0
- package/dist/measure.d.ts +1 -0
- package/dist/measure.d.ts.map +1 -1
- package/dist/measure.js +5 -1
- package/dist/measure.js.map +1 -1
- package/dist/number-name.d.ts +7 -3
- package/dist/number-name.d.ts.map +1 -1
- package/dist/number-name.js +51 -15
- package/dist/number-name.js.map +1 -1
- package/dist/orthography.d.ts +4 -0
- package/dist/orthography.d.ts.map +1 -1
- package/dist/orthography.js +11 -9
- package/dist/orthography.js.map +1 -1
- package/dist/plain-source.d.ts +6 -0
- package/dist/plain-source.d.ts.map +1 -0
- package/dist/plain-source.js +6 -0
- package/dist/plain-source.js.map +1 -0
- package/dist/plugin.d.ts +18 -1
- package/dist/plugin.d.ts.map +1 -1
- package/dist/preamble-paragraphs.d.ts +7 -0
- package/dist/preamble-paragraphs.d.ts.map +1 -0
- package/dist/preamble-paragraphs.js +7 -0
- package/dist/preamble-paragraphs.js.map +1 -0
- package/dist/quoted-span.d.ts +10 -0
- package/dist/quoted-span.d.ts.map +1 -0
- package/dist/quoted-span.js +20 -0
- package/dist/quoted-span.js.map +1 -0
- package/dist/render/eval.d.ts +2 -1
- package/dist/render/eval.d.ts.map +1 -1
- package/dist/render/eval.js +84 -43
- package/dist/render/eval.js.map +1 -1
- package/dist/resolve-genre.d.ts +11 -0
- package/dist/resolve-genre.d.ts.map +1 -0
- package/dist/resolve-genre.js +18 -0
- package/dist/resolve-genre.js.map +1 -0
- package/dist/soft-break.d.ts +19 -0
- package/dist/soft-break.d.ts.map +1 -0
- package/dist/soft-break.js +78 -0
- package/dist/soft-break.js.map +1 -0
- package/dist/structure/build.d.ts.map +1 -1
- package/dist/structure/build.js +23 -10
- package/dist/structure/build.js.map +1 -1
- package/dist/structure/date-order.d.ts +3 -0
- package/dist/structure/date-order.d.ts.map +1 -1
- package/dist/structure/date-order.js +9 -1
- package/dist/structure/date-order.js.map +1 -1
- package/dist/structure/heading-leaves.d.ts +7 -0
- package/dist/structure/heading-leaves.d.ts.map +1 -0
- package/dist/structure/heading-leaves.js +6 -0
- package/dist/structure/heading-leaves.js.map +1 -0
- package/dist/structure/issues.d.ts +1 -1
- package/dist/structure/issues.d.ts.map +1 -1
- package/dist/structure/issues.js +14 -3
- package/dist/structure/issues.js.map +1 -1
- package/dist/structure/listed-tags.d.ts +3 -0
- package/dist/structure/listed-tags.d.ts.map +1 -0
- package/dist/structure/listed-tags.js +6 -0
- package/dist/structure/listed-tags.js.map +1 -0
- package/dist/structure/number-in-sentence.d.ts +15 -0
- package/dist/structure/number-in-sentence.d.ts.map +1 -0
- package/dist/structure/number-in-sentence.js +13 -0
- package/dist/structure/number-in-sentence.js.map +1 -0
- package/dist/structure/unit-word.d.ts +8 -0
- package/dist/structure/unit-word.d.ts.map +1 -0
- package/dist/structure/unit-word.js +10 -0
- package/dist/structure/unit-word.js.map +1 -0
- package/dist/structure/wrapped-tail.d.ts +18 -0
- package/dist/structure/wrapped-tail.d.ts.map +1 -0
- package/dist/structure/wrapped-tail.js +11 -0
- package/dist/structure/wrapped-tail.js.map +1 -0
- package/package.json +3 -3
- package/rules/dangling-reference.yaml +2 -2
- package/rules/date-order.yaml +2 -2
- package/rules/duplicate-definition.yaml +2 -2
- package/rules/latin-spacing.yaml +4 -2
- package/rules/max-kanji-continuous.yaml +1 -1
- package/rules/no-doubled-joshi.yaml +2 -2
- package/rules/no-nakaguro-parallel.yaml +2 -2
- package/rules/title-case-consistency.yaml +2 -2
- package/rules/undefined-acronym.yaml +15 -1
- package/rules/unqualified-superlative.yaml +1 -1
- package/src/address-chain.ts +5 -2
- package/src/cli-text.ts +19 -3
- package/src/cli.ts +22 -19
- package/src/commands/eval.ts +26 -5
- package/src/commands/feedback.ts +2 -1
- package/src/commands/test.ts +3 -2
- package/src/commands/tree.ts +2 -1
- package/src/detectors/acronym-compound.ts +9 -0
- package/src/detectors/acronym-context.ts +32 -1
- package/src/detectors/acronym-expansion.ts +72 -5
- package/src/detectors/char-shape.ts +44 -17
- package/src/detectors/citation-key.ts +13 -0
- package/src/detectors/cited-name.ts +23 -0
- package/src/detectors/concrete-evidence.ts +31 -0
- package/src/detectors/conjugated-form.ts +40 -0
- package/src/detectors/counter-edge.ts +29 -0
- package/src/detectors/dotted-name.ts +14 -0
- package/src/detectors/doubled-word.ts +25 -2
- package/src/detectors/en-shape.ts +17 -30
- package/src/detectors/enumerated-runs.ts +68 -0
- package/src/detectors/gram-place.ts +8 -1
- package/src/detectors/gram-predicate.ts +55 -0
- package/src/detectors/heading-case.ts +41 -0
- package/src/detectors/heading-echo.ts +5 -1
- package/src/detectors/lexicon.ts +8 -5
- package/src/detectors/middle-dot.ts +10 -0
- package/src/detectors/name-numeral.ts +37 -0
- package/src/detectors/name-part.ts +24 -0
- package/src/detectors/name-run.ts +16 -0
- package/src/detectors/orthography.ts +20 -9
- package/src/detectors/placeholder.ts +38 -0
- package/src/detectors/polite-word.ts +14 -0
- package/src/detectors/quoted-variant.ts +20 -0
- package/src/detectors/register.ts +11 -7
- package/src/detectors/roman-numeral.ts +5 -0
- package/src/detectors/sentence-ending.ts +23 -6
- package/src/detectors/series-label.ts +33 -0
- package/src/detectors/signals.ts +44 -31
- package/src/detectors/structure-tree.ts +2 -2
- package/src/detectors/structure.ts +2 -1
- package/src/detectors/superlative-name.ts +30 -0
- package/src/detectors/token-shape.ts +12 -3
- package/src/document.ts +12 -5
- package/src/genre-check.ts +46 -0
- package/src/genre.ts +34 -5
- package/src/init.ts +0 -13
- package/src/joined-view.ts +96 -0
- package/src/line-continues.ts +43 -0
- package/src/measure.ts +9 -1
- package/src/number-name.ts +65 -15
- package/src/orthography.ts +15 -9
- package/src/plain-source.ts +5 -0
- package/src/plugin.ts +23 -2
- package/src/preamble-paragraphs.ts +9 -0
- package/src/quoted-span.ts +25 -0
- package/src/render/eval.ts +101 -38
- package/src/resolve-genre.ts +20 -0
- package/src/soft-break.ts +94 -0
- package/src/structure/build.ts +29 -17
- package/src/structure/date-order.ts +16 -1
- package/src/structure/heading-leaves.ts +7 -0
- package/src/structure/issues.ts +18 -3
- package/src/structure/listed-tags.ts +7 -0
- package/src/structure/number-in-sentence.ts +20 -0
- package/src/structure/unit-word.ts +11 -0
- package/src/structure/wrapped-tail.ts +22 -0
|
@@ -0,0 +1,38 @@
|
|
|
1
|
+
const WORD = /[A-Za-z][A-Za-z'-]*/gu;
|
|
2
|
+
/** 小さい語は Title Case でも小文字のままなので、大文字化の判定から外す。 */
|
|
3
|
+
const MINOR = new Set(["a", "an", "the", "and", "or", "but", "of", "in", "on", "at", "to", "for", "with", "as", "by", "from", "is"]);
|
|
4
|
+
const TITLE_DEPTH = 1;
|
|
5
|
+
/** 見出しが Title Case か。判定できなければ undefined。 */
|
|
6
|
+
export const isTitleCase = (heading) => {
|
|
7
|
+
const words = [...heading.matchAll(WORD)].map((match) => match[0]).filter((word) => !MINOR.has(word.toLowerCase()));
|
|
8
|
+
// 1 語の見出しは、どちらの流儀でも先頭が大文字になる。判定できない。
|
|
9
|
+
if (words.length < 2)
|
|
10
|
+
return undefined;
|
|
11
|
+
const capitalized = words.filter((word) => word[0] === word[0]?.toUpperCase()).length;
|
|
12
|
+
return capitalized === words.length;
|
|
13
|
+
};
|
|
14
|
+
/**
|
|
15
|
+
* 文書の題名。最初の見出しが深さ 1 で、深さ 1 の見出しがほかに無いときの、その見出し。
|
|
16
|
+
* 深さ 1 が複数あると、最初のものが題名なのか章なのかを構造からは決められないので、題名とは呼ばない。
|
|
17
|
+
* 見出しより前の導入部(深さ 0)は見出しではないので数えない。
|
|
18
|
+
*/
|
|
19
|
+
export const pageTitleOf = (sections) => {
|
|
20
|
+
const headed = sections.filter((section) => section.depth > 0);
|
|
21
|
+
const first = headed[0];
|
|
22
|
+
if (first?.depth !== TITLE_DEPTH)
|
|
23
|
+
return undefined;
|
|
24
|
+
return headed.filter((section) => section.depth === TITLE_DEPTH).length === 1 ? first : undefined;
|
|
25
|
+
};
|
|
26
|
+
/**
|
|
27
|
+
* 節の見出しのうち少数派の流儀(true なら Title Case)。少数派が無ければ undefined。
|
|
28
|
+
* 題名は数に入れない。題名は名前なので、節と流儀を分けるのはよくある揃えかた。
|
|
29
|
+
* 同数のときだけ題名の流儀で決める。題名と同じ流儀のほうが、書き手の選んだ流儀である見込みが高い。
|
|
30
|
+
*/
|
|
31
|
+
export const minorityCase = (sections, pageTitleCase) => {
|
|
32
|
+
if (sections.titleCase === sections.sentenceCase)
|
|
33
|
+
return sections.titleCase === 0 || pageTitleCase === undefined ? undefined : !pageTitleCase;
|
|
34
|
+
const minorityIsTitle = sections.titleCase < sections.sentenceCase;
|
|
35
|
+
const few = minorityIsTitle ? sections.titleCase : sections.sentenceCase;
|
|
36
|
+
return few === 0 ? undefined : minorityIsTitle;
|
|
37
|
+
};
|
|
38
|
+
//# sourceMappingURL=heading-case.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"heading-case.js","sourceRoot":"","sources":["../../src/detectors/heading-case.ts"],"names":[],"mappings":"AAAA,MAAM,IAAI,GAAG,uBAAuB,CAAC;AAErC,gDAAgD;AAChD,MAAM,KAAK,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,CAAC,CAAC,CAAC;AAErI,MAAM,WAAW,GAAG,CAAC,CAAC;AAEtB,4CAA4C;AAC5C,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,OAAe,EAAuB,EAAE;IAClE,MAAM,KAAK,GAAG,CAAC,GAAG,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;IACpH,qCAAqC;IACrC,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,SAAS,CAAC;IACvC,MAAM,WAAW,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC,EAAE,WAAW,EAAE,CAAC,CAAC,MAAM,CAAC;IACtF,OAAO,WAAW,KAAK,KAAK,CAAC,MAAM,CAAC;AACtC,CAAC,CAAC;AAEF;;;;GAIG;AACH,MAAM,CAAC,MAAM,WAAW,GAAG,CAAuC,QAAsB,EAAiB,EAAE;IACzG,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;IAC/D,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;IACxB,IAAI,KAAK,EAAE,KAAK,KAAK,WAAW;QAAE,OAAO,SAAS,CAAC;IACnD,OAAO,MAAM,CAAC,MAAM,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,KAAK,KAAK,WAAW,CAAC,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAAC;AACpG,CAAC,CAAC;AAIF;;;;GAIG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,QAAoB,EAAE,aAAkC,EAAuB,EAAE;IAC5G,IAAI,QAAQ,CAAC,SAAS,KAAK,QAAQ,CAAC,YAAY;QAAE,OAAO,QAAQ,CAAC,SAAS,KAAK,CAAC,IAAI,aAAa,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,aAAa,CAAC;IAC9I,MAAM,eAAe,GAAG,QAAQ,CAAC,SAAS,GAAG,QAAQ,CAAC,YAAY,CAAC;IACnE,MAAM,GAAG,GAAG,eAAe,CAAC,CAAC,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC,QAAQ,CAAC,YAAY,CAAC;IACzE,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,eAAe,CAAC;AACjD,CAAC,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"heading-echo.d.ts","sourceRoot":"","sources":["../../src/detectors/heading-echo.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"heading-echo.d.ts","sourceRoot":"","sources":["../../src/detectors/heading-echo.ts"],"names":[],"mappings":"AAGA,OAAO,KAAK,EAAE,QAAQ,EAA+C,MAAM,cAAc,CAAC;AA6D1F,eAAO,MAAM,WAAW,EAAE,QAezB,CAAC"}
|
|
@@ -1,5 +1,6 @@
|
|
|
1
1
|
import { lengthOf } from "../measure.js";
|
|
2
2
|
import { handsOver } from "./lead-in.js";
|
|
3
|
+
import { withoutQuotedVariants } from "./quoted-variant.js";
|
|
3
4
|
/** 見出しが短すぎると、偶然の一致で 100% になる。これ未満の見出しは見ない。 */
|
|
4
5
|
const MIN_GRAMS = 4;
|
|
5
6
|
/**
|
|
@@ -50,12 +51,14 @@ const leadsIn = (doc, section, phrases) => {
|
|
|
50
51
|
const first = section.firstSentence;
|
|
51
52
|
return first !== undefined && handsOver(first.text, doc.source.slice(first.span.end, section.span.end), phrases);
|
|
52
53
|
};
|
|
54
|
+
/** 見出しとの重なりを測る文。用語集や表記の手引きは見出しの語の別の書き方を引用する(Not “datacentre”)ので、それは数えない。 */
|
|
55
|
+
const echoedText = (section) => withoutQuotedVariants(section.firstSentence?.text ?? "", section.heading);
|
|
53
56
|
export const headingEcho = (doc, options) => {
|
|
54
57
|
const leadIns = (doc.lexicons["lead-in"] ?? []).map((entry) => entry.pattern);
|
|
55
58
|
return doc.sections
|
|
56
59
|
.filter((section) => section.heading.length > 0 && section.firstSentence !== undefined && addsLittle(section, doc.lengthUnit))
|
|
57
60
|
.filter((section) => !leadsIn(doc, section, leadIns))
|
|
58
|
-
.map((section) => ({ section, overlap: Math.round(containment(trigrams(section.heading), trigrams(section
|
|
61
|
+
.map((section) => ({ section, overlap: Math.round(containment(trigrams(section.heading), trigrams(echoedText(section))) * 100) }))
|
|
59
62
|
.filter(({ overlap }) => overlap >= options.limit)
|
|
60
63
|
.map(({ section, overlap }) => ({
|
|
61
64
|
rule: "heading-echo",
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"heading-echo.js","sourceRoot":"","sources":["../../src/detectors/heading-echo.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,eAAe,CAAC;AACzC,OAAO,EAAE,SAAS,EAAE,MAAM,cAAc,CAAC;
|
|
1
|
+
{"version":3,"file":"heading-echo.js","sourceRoot":"","sources":["../../src/detectors/heading-echo.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,eAAe,CAAC;AACzC,OAAO,EAAE,SAAS,EAAE,MAAM,cAAc,CAAC;AACzC,OAAO,EAAE,qBAAqB,EAAE,MAAM,qBAAqB,CAAC;AAG5D,8CAA8C;AAC9C,MAAM,SAAS,GAAG,CAAC,CAAC;AAEpB;;;GAGG;AACH,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAe,EAAE;IAC7C,2CAA2C;IAC3C,sDAAsD;IACtD,MAAM,KAAK,GAAG,CAAC,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC;IAC3D,OAAO,IAAI,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;AAC/H,CAAC,CAAC;AAEF;;;;;;;GAOG;AACH,MAAM,WAAW,GAAG,CAAC,OAAoB,EAAE,QAAqB,EAAU,EAAE;IAC1E,IAAI,OAAO,CAAC,IAAI,GAAG,SAAS,IAAI,QAAQ,CAAC,IAAI,KAAK,CAAC;QAAE,OAAO,CAAC,CAAC;IAC9D,OAAO,CAAC,GAAG,OAAO,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC;AACjF,CAAC,CAAC;AAEF;;;;;;GAMG;AACH,MAAM,YAAY,GAAG,EAAE,IAAI,EAAE,CAAC,EAAE,IAAI,EAAE,EAAE,EAAE,CAAC;AAE3C,MAAM,YAAY,GAAG,CAAC,OAAe,EAAE,IAAgB,EAAU,EAAE,CACjE,IAAI,KAAK,MAAM;IACb,CAAC,CAAC,OAAO;SACJ,IAAI,EAAE;SACN,KAAK,CAAC,MAAM,CAAC;SACb,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,MAAM;IAC7C,CAAC,CAAC,OAAO,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC;AAE1C,MAAM,UAAU,GAAG,CAAC,OAAgB,EAAE,IAAgB,EAAW,EAAE;IACjE,MAAM,KAAK,GAAG,OAAO,CAAC,aAAa,CAAC;IACpC,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC;IACtC,OAAO,QAAQ,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,YAAY,CAAC,OAAO,CAAC,OAAO,EAAE,IAAI,CAAC,IAAI,YAAY,CAAC,IAAI,CAAC,CAAC;AAC3F,CAAC,CAAC;AAEF,yCAAyC;AACzC,MAAM,OAAO,GAAG,CAAC,GAAkB,EAAE,OAAgB,EAAE,OAA0B,EAAW,EAAE;IAC5F,MAAM,KAAK,GAAG,OAAO,CAAC,aAAa,CAAC;IACpC,OAAO,KAAK,KAAK,SAAS,IAAI,SAAS,CAAC,KAAK,CAAC,IAAI,EAAE,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,EAAE,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,OAAO,CAAC,CAAC;AACnH,CAAC,CAAC;AAEF,4EAA4E;AAC5E,MAAM,UAAU,GAAG,CAAC,OAAgB,EAAU,EAAE,CAAC,qBAAqB,CAAC,OAAO,CAAC,aAAa,EAAE,IAAI,IAAI,EAAE,EAAE,OAAO,CAAC,OAAO,CAAC,CAAC;AAE3H,MAAM,CAAC,MAAM,WAAW,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAC/D,MAAM,OAAO,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IAC9E,OAAO,GAAG,CAAC,QAAQ;SAChB,MAAM,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,IAAI,OAAO,CAAC,aAAa,KAAK,SAAS,IAAI,UAAU,CAAC,OAAO,EAAE,GAAG,CAAC,UAAU,CAAC,CAAC;SAC7H,MAAM,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,GAAG,EAAE,OAAO,EAAE,OAAO,CAAC,CAAC;SACpD,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,EAAE,OAAO,EAAE,OAAO,EAAE,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,QAAQ,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,EAAE,CAAC,CAAC;SACjI,MAAM,CAAC,CAAC,EAAE,OAAO,EAAE,EAAE,EAAE,CAAC,OAAO,IAAI,OAAO,CAAC,KAAK,CAAC;SACjD,GAAG,CAAC,CAAC,EAAE,OAAO,EAAE,OAAO,EAAE,EAAE,EAAE,CAAC,CAAC;QAC9B,IAAI,EAAE,cAAc;QACpB,QAAQ,EAAE,SAAS;QACnB,IAAI,EAAE,CAAC;QACP,MAAM,EAAE,CAAC;QACT,KAAK,EAAE,OAAO,CAAC,aAAa,EAAE,IAAI,CAAC,IAAI,EAAE,IAAI,EAAE;QAC/C,MAAM,EAAE,EAAE,OAAO,EAAE,OAAO,CAAC,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,aAAa,EAAE,IAAI,CAAC,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC,KAAK,EAAE;KAC5I,CAAC,CAAC,CAAC;AACR,CAAC,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"lexicon.d.ts","sourceRoot":"","sources":["../../src/detectors/lexicon.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,QAAQ,EAA2D,MAAM,cAAc,CAAC;
|
|
1
|
+
{"version":3,"file":"lexicon.d.ts","sourceRoot":"","sources":["../../src/detectors/lexicon.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,QAAQ,EAA2D,MAAM,cAAc,CAAC;AAqCtG,eAAO,MAAM,cAAc,UAAmC,CAAC;AAC/D,eAAO,MAAM,cAAc,UAAwC,CAAC;AA8BpE;;;;GAIG;AACH,eAAO,MAAM,sBAAsB,EAAE,QAWpC,CAAC;AAQF;;;GAGG;AACH,eAAO,MAAM,mBAAmB,EAAE,QA0BjC,CAAC;AAEF;;;GAGG;AACH,eAAO,MAAM,MAAM,EAAE,QAyBpB,CAAC;AAiBF,eAAO,MAAM,cAAc,EAAE,QAwB5B,CAAC"}
|
|
@@ -2,6 +2,7 @@ import { joinWords } from "./word-list.js";
|
|
|
2
2
|
import { wordsOf } from "./structure.js";
|
|
3
3
|
import { entryIn, entryOpens, entryRanges } from "./lexicon-match.js";
|
|
4
4
|
import { scoped, scopeMarkersOf } from "./superlative-scope.js";
|
|
5
|
+
import { namesQuantity } from "./superlative-name.js";
|
|
5
6
|
const PER = 1000;
|
|
6
7
|
/** 密度を見る rule は短い文書を測らない。単位は言語で違うので床も分ける。 */
|
|
7
8
|
const FLOOR = { word: 200, char: 500 };
|
|
@@ -33,13 +34,15 @@ const DIGIT = /\d/u;
|
|
|
33
34
|
const qualifiersOf = (doc) => ({
|
|
34
35
|
comparison: doc.lexicons["comparison-marker"] ?? [],
|
|
35
36
|
scope: scopeMarkersOf(doc.lexicons["superlative-scope"] ?? []),
|
|
37
|
+
quantityNouns: doc.lexicons["quantity-noun"] ?? [],
|
|
36
38
|
});
|
|
37
|
-
/**
|
|
38
|
-
const
|
|
39
|
+
/** どの出現も範囲を持つか量の名前のときだけ。1 つでもそうでない出現があれば、その文には限定の無い最上級がある。 */
|
|
40
|
+
const everyQualified = (sentence, entry, qualifiers) => {
|
|
41
|
+
const tokens = sentence.tokens ?? [];
|
|
39
42
|
const ranges = entryRanges(sentence, entry);
|
|
40
|
-
return ranges.length > 0 && ranges.every((range) => scoped(
|
|
43
|
+
return ranges.length > 0 && ranges.every((range) => scoped(tokens, range, qualifiers.scope) || namesQuantity(tokens, range, qualifiers.quantityNouns));
|
|
41
44
|
};
|
|
42
|
-
const qualified = (sentence, entry, qualifiers) => DIGIT.test(sentence.text) || qualifiers.comparison.some((marker) => entryIn(sentence, marker)) ||
|
|
45
|
+
const qualified = (sentence, entry, qualifiers) => DIGIT.test(sentence.text) || qualifiers.comparison.some((marker) => entryIn(sentence, marker)) || everyQualified(sentence, entry, qualifiers);
|
|
43
46
|
const bareHits = (doc, lexicon) => {
|
|
44
47
|
const qualifiers = qualifiersOf(doc);
|
|
45
48
|
return doc.sentences.flatMap((sentence) => lexicon.filter((entry) => entryIn(sentence, entry) && !qualified(sentence, entry, qualifiers)).map((entry) => ({ sentence, matched: entry.pattern })));
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"lexicon.js","sourceRoot":"","sources":["../../src/detectors/lexicon.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,SAAS,EAAE,MAAM,gBAAgB,CAAC;AAC3C,OAAO,EAAE,OAAO,EAAE,MAAM,gBAAgB,CAAC;AAEzC,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,WAAW,EAAE,MAAM,oBAAoB,CAAC;AACtE,OAAO,EAAE,MAAM,EAAE,cAAc,EAAqB,MAAM,wBAAwB,CAAC;
|
|
1
|
+
{"version":3,"file":"lexicon.js","sourceRoot":"","sources":["../../src/detectors/lexicon.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,SAAS,EAAE,MAAM,gBAAgB,CAAC;AAC3C,OAAO,EAAE,OAAO,EAAE,MAAM,gBAAgB,CAAC;AAEzC,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,WAAW,EAAE,MAAM,oBAAoB,CAAC;AACtE,OAAO,EAAE,MAAM,EAAE,cAAc,EAAqB,MAAM,wBAAwB,CAAC;AACnF,OAAO,EAAE,aAAa,EAAE,MAAM,uBAAuB,CAAC;AAEtD,MAAM,GAAG,GAAG,IAAI,CAAC;AAEjB,6CAA6C;AAC7C,MAAM,KAAK,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,GAAG,EAAE,CAAC;AAIvC,MAAM,OAAO,GAAG,CAAC,GAAkB,EAAE,OAAgB,EAAS,EAAE,CAC9D,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,OAAO,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,OAAO,EAAE,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,CAAC;AAElJ;;;GAGG;AACH,MAAM,WAAW,GACf,CAAC,IAAY,EAAY,EAAE,CAC3B,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAC1B,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,EAAE,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC;IACjD,MAAM,MAAM,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC;IAC5B,MAAM,IAAI,GAAG,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC,GAAG,GAAG,CAAC,CAAC;IACzE,MAAM,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;IACtB,IAAI,MAAM,GAAG,KAAK,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,KAAK,KAAK,SAAS,IAAI,IAAI,IAAI,OAAO,CAAC,KAAK;QAAE,OAAO,EAAE,CAAC;IAC9F,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;QACxB,IAAI;QACJ,QAAQ,EAAE,SAAS;QACnB,IAAI,EAAE,CAAC;QACP,MAAM,EAAE,CAAC;QACT,KAAK,EAAE,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;QAC/B,MAAM,EAAE,EAAE,OAAO,EAAE,GAAG,CAAC,OAAO,EAAE,KAAK,EAAE,IAAI,CAAC,MAAM,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE;KAC3H,CAAC,CAAC,CAAC;AACN,CAAC,CAAC;AAEJ,MAAM,CAAC,MAAM,cAAc,GAAG,WAAW,CAAC,mBAAmB,CAAC,CAAC;AAC/D,MAAM,CAAC,MAAM,cAAc,GAAG,WAAW,CAAC,wBAAwB,CAAC,CAAC;AAEpE,qFAAqF;AACrF,MAAM,KAAK,GAAG,KAAK,CAAC;AAIpB,MAAM,YAAY,GAAG,CAAC,GAAkB,EAAc,EAAE,CAAC,CAAC;IACxD,UAAU,EAAE,GAAG,CAAC,QAAQ,CAAC,mBAAmB,CAAC,IAAI,EAAE;IACnD,KAAK,EAAE,cAAc,CAAC,GAAG,CAAC,QAAQ,CAAC,mBAAmB,CAAC,IAAI,EAAE,CAAC;IAC9D,aAAa,EAAE,GAAG,CAAC,QAAQ,CAAC,eAAe,CAAC,IAAI,EAAE;CACnD,CAAC,CAAC;AAEH,8DAA8D;AAC9D,MAAM,cAAc,GAAG,CAAC,QAAkB,EAAE,KAAmB,EAAE,UAAsB,EAAW,EAAE;IAClG,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,IAAI,EAAE,CAAC;IACrC,MAAM,MAAM,GAAG,WAAW,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC;IAC5C,OAAO,MAAM,CAAC,MAAM,GAAG,CAAC,IAAI,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,MAAM,EAAE,KAAK,EAAE,UAAU,CAAC,KAAK,CAAC,IAAI,aAAa,CAAC,MAAM,EAAE,KAAK,EAAE,UAAU,CAAC,aAAa,CAAC,CAAC,CAAC;AACzJ,CAAC,CAAC;AAEF,MAAM,SAAS,GAAG,CAAC,QAAkB,EAAE,KAAmB,EAAE,UAAsB,EAAW,EAAE,CAC7F,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,UAAU,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,OAAO,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC,IAAI,cAAc,CAAC,QAAQ,EAAE,KAAK,EAAE,UAAU,CAAC,CAAC;AAEhJ,MAAM,QAAQ,GAAG,CAAC,GAAkB,EAAE,OAAgB,EAAS,EAAE;IAC/D,MAAM,UAAU,GAAG,YAAY,CAAC,GAAG,CAAC,CAAC;IACrC,OAAO,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CACxC,OAAO,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,OAAO,CAAC,QAAQ,EAAE,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,QAAQ,EAAE,KAAK,EAAE,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,OAAO,EAAE,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC,CACtJ,CAAC;AACJ,CAAC,CAAC;AAEF;;;;GAIG;AACH,MAAM,CAAC,MAAM,sBAAsB,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAC1E,MAAM,IAAI,GAAG,QAAQ,CAAC,GAAG,EAAE,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC;IAClD,IAAI,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK;QAAE,OAAO,EAAE,CAAC;IAC3C,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;QACxB,IAAI,EAAE,yBAAyB;QAC/B,QAAQ,EAAE,SAAS;QACnB,IAAI,EAAE,CAAC;QACP,MAAM,EAAE,CAAC;QACT,KAAK,EAAE,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;QAC/B,MAAM,EAAE,EAAE,OAAO,EAAE,GAAG,CAAC,OAAO,EAAE,KAAK,EAAE,IAAI,CAAC,MAAM,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE;KAC5G,CAAC,CAAC,CAAC;AACN,CAAC,CAAC;AAEF,MAAM,QAAQ,GAAG,CAAC,GAAkB,EAAE,OAAgB,EAA0B,EAAE,CAChF,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,SAAS,EAAE,EAAE;IAC/B,MAAM,IAAI,GAAG,SAAS,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC;IACpC,OAAO,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,UAAU,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC;AACpG,CAAC,CAAC,CAAC;AAEL;;;GAGG;AACH,MAAM,CAAC,MAAM,mBAAmB,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IACvE,MAAM,OAAO,GAAG,QAAQ,CAAC,GAAG,EAAE,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC;IACrD,MAAM,IAAI,GAAG,OAAO,CAAC,MAAM,CACzB,CAAC,GAAG,EAAE,MAAM,EAAE,KAAK,EAAE,EAAE;QACrB,IAAI,MAAM,KAAK,SAAS;YAAE,OAAO,EAAE,IAAI,EAAE,CAAC,GAAG,GAAG,CAAC,IAAI,EAAE,GAAG,CAAC,OAAO,CAAC,EAAE,OAAO,EAAE,EAAE,EAAE,CAAC;QACnF,OAAO,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,OAAO,EAAE,CAAC,GAAG,GAAG,CAAC,OAAO,EAAE,KAAK,CAAC,EAAE,CAAC;IAC9D,CAAC,EACD,EAAE,IAAI,EAAE,EAAE,EAAE,OAAO,EAAE,EAAE,EAAE,CAC1B,CAAC;IACF,OAAO,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,IAAI,CAAC,OAAO,CAAC;SAChC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC;SAC3C,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE;QACf,MAAM,EAAE,GAAG,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;QACvC,OAAO,EAAE,KAAK,SAAS;YACrB,CAAC,CAAC,EAAE;YACJ,CAAC,CAAC;gBACE;oBACE,IAAI,EAAE,sBAAsB;oBAC5B,QAAQ,EAAE,SAAkB;oBAC5B,IAAI,EAAE,CAAC;oBACP,MAAM,EAAE,CAAC;oBACT,KAAK,EAAE,EAAE,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,IAAI,EAAE,IAAI,EAAE;oBACzC,MAAM,EAAE,EAAE,KAAK,EAAE,GAAG,CAAC,MAAM,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,EAAE;iBAC3E;aACF,CAAC;IACR,CAAC,CAAC,CAAC;AACP,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,MAAM,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAC1D,MAAM,OAAO,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC;IACpH,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,OAAO,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC;IACpG,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,KAAK,EAAE,EAAE,CAAC,GAAG,GAAG,CAAC,KAAK,CAAC,MAAM,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC;IAC1F,MAAM,KAAK,GAAG,GAAG,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,OAAO,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC;IAChG,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,IAAI,OAAO,CAAC,KAAK;QAAE,OAAO,EAAE,CAAC;IAC7D,OAAO;QACL;YACE,IAAI,EAAE,SAAS;YACf,QAAQ,EAAE,MAAM;YAChB,IAAI,EAAE,CAAC;YACP,MAAM,EAAE,CAAC;YACT,KAAK,EAAE,KAAK,CAAC,IAAI,CAAC,IAAI,EAAE;YACxB,MAAM,EAAE;gBACN,IAAI,EAAE,SAAS,CACb,KAAK,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC,EACjD,GAAG,CAAC,QAAQ,CACb;gBACD,KAAK,EAAE,KAAK,CAAC,MAAM;gBACnB,OAAO,EAAE,KAAK;gBACd,KAAK,EAAE,OAAO,CAAC,KAAK;gBACpB,MAAM,EAAE,KAAK,CAAC,IAAI,CAAC,KAAK;aACzB;SACF;KACF,CAAC;AACJ,CAAC,CAAC;AAEF;;;;;;;;GAQG;AACH;;;GAGG;AACH,MAAM,YAAY,GAAG,CAAC,IAAY,EAAE,IAAY,EAAW,EAAE,CAAC,IAAI,MAAM,CAAC,MAAM,IAAI,CAAC,OAAO,CAAC,sBAAsB,EAAE,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAElJ,MAAM,CAAC,MAAM,cAAc,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAClE,MAAM,KAAK,GAAG,CAAC,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,OAAO,EAAE,IAAI,EAAE,KAAK,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,CAAC;IACrJ,MAAM,IAAI,GAAG,GAAG,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACpF,MAAM,IAAI,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;IAClF,MAAM,OAAO,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,EAAE,IAAI,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;IACpF,oCAAoC;IACpC,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IACzD,MAAM,eAAe,GAAG,IAAI,CAAC,MAAM,IAAI,OAAO,CAAC,MAAM,CAAC;IACtD,MAAM,GAAG,GAAG,eAAe,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,OAAO,CAAC;IAC7C,IAAI,GAAG,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK;QAAE,OAAO,EAAE,CAAC;IAC1C,MAAM,MAAM,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC;IAC3F,MAAM,IAAI,GAAG,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE;QAC9C,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CAAC,WAAW,EAAE,CAAC;QACzC,MAAM,OAAO,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;QAChE,OAAO,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,OAAO,EAAE,CAAC,CAAC;IAC9D,CAAC,CAAC,CAAC;IACH,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;QACxB,IAAI,EAAE,yBAAyB;QAC/B,QAAQ,EAAE,MAAM;QAChB,IAAI,EAAE,CAAC;QACP,MAAM,EAAE,CAAC;QACT,KAAK,EAAE,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;QAC/B,MAAM,EAAE,EAAE,OAAO,EAAE,GAAG,CAAC,OAAO,EAAE,KAAK,EAAE,GAAG,CAAC,MAAM,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE;KAC3G,CAAC,CAAC,CAAC;AACN,CAAC,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"middle-dot.d.ts","sourceRoot":"","sources":["../../src/detectors/middle-dot.ts"],"names":[],"mappings":"AAKA;;;GAGG;AACH,eAAO,MAAM,gBAAgB,GAAI,MAAM,MAAM,KAAG,MAAyE,CAAC"}
|
|
@@ -0,0 +1,9 @@
|
|
|
1
|
+
const MIDDLE_DOT = /・/gu;
|
|
2
|
+
/** 文頭か行頭(前に空白だけ)の「・」。前に並べる項目が無いので、箇条書きの印。 */
|
|
3
|
+
const BULLET_DOT = /(^|\n)[^\S\n]*・/gu;
|
|
4
|
+
/**
|
|
5
|
+
* 項目を並べる中黒の数。箇条書きの印の「・」は数えない。
|
|
6
|
+
* 「・ 水分を補給すること」が何行続いても、どこまでが 1 つの項目かは改行で読める。
|
|
7
|
+
*/
|
|
8
|
+
export const parallelDotCount = (text) => [...text.replace(BULLET_DOT, "$1").matchAll(MIDDLE_DOT)].length;
|
|
9
|
+
//# sourceMappingURL=middle-dot.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"middle-dot.js","sourceRoot":"","sources":["../../src/detectors/middle-dot.ts"],"names":[],"mappings":"AAAA,MAAM,UAAU,GAAG,KAAK,CAAC;AAEzB,6CAA6C;AAC7C,MAAM,UAAU,GAAG,mBAAmB,CAAC;AAEvC;;;GAGG;AACH,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,IAAY,EAAU,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,UAAU,EAAE,IAAI,CAAC,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC,CAAC,MAAM,CAAC"}
|
|
@@ -0,0 +1,16 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* 大文字で始まる語のすぐ後ろの、I・V・X だけで書いたローマ数字(Engineer II、Leopold III、World War II)。名前に付く
|
|
3
|
+
* 番号で、略語ではない。語の前に文字が無ければ(文の頭、箇条書きの番号の後ろ)、大文字は名前の印ではなく文頭の書き方なので
|
|
4
|
+
* 外さない(Start IV fluids、1. Give IV fluids)。
|
|
5
|
+
* CD、MD、DC、CI、MIX のように C・D・L・M を含む数字は略語であることが多いので、この位置でも略語のまま数える。
|
|
6
|
+
*/
|
|
7
|
+
type Span = {
|
|
8
|
+
readonly start: number;
|
|
9
|
+
readonly end: number;
|
|
10
|
+
};
|
|
11
|
+
/** 文の中の、名前の後ろの番号の範囲(数字だけ)。 */
|
|
12
|
+
export declare const nameNumeralSpans: (text: string) => Span[];
|
|
13
|
+
/** 文の中の、1 語で立つ番号の範囲(数字だけ)。 */
|
|
14
|
+
export declare const loneNumeralSpans: (text: string) => Span[];
|
|
15
|
+
export {};
|
|
16
|
+
//# sourceMappingURL=name-numeral.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"name-numeral.d.ts","sourceRoot":"","sources":["../../src/detectors/name-numeral.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,KAAK,IAAI,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAA;CAAE,CAAC;AAM7D,8BAA8B;AAC9B,eAAO,MAAM,gBAAgB,GAAI,MAAM,MAAM,KAAG,IAAI,EAO9C,CAAC;AAUP,8BAA8B;AAC9B,eAAO,MAAM,gBAAgB,GAAI,MAAM,MAAM,KAAG,IAAI,EAIhD,CAAC"}
|
|
@@ -0,0 +1,30 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* 大文字で始まる語のすぐ後ろの、I・V・X だけで書いたローマ数字(Engineer II、Leopold III、World War II)。名前に付く
|
|
3
|
+
* 番号で、略語ではない。語の前に文字が無ければ(文の頭、箇条書きの番号の後ろ)、大文字は名前の印ではなく文頭の書き方なので
|
|
4
|
+
* 外さない(Start IV fluids、1. Give IV fluids)。
|
|
5
|
+
* CD、MD、DC、CI、MIX のように C・D・L・M を含む数字は略語であることが多いので、この位置でも略語のまま数える。
|
|
6
|
+
*/
|
|
7
|
+
const SMALL_NUMERAL = "(?=[IVX])X{0,3}(?:IX|IV|V?I{0,3})";
|
|
8
|
+
const NAMED_NUMERAL = new RegExp(String.raw `(?<![\p{L}\p{N}_])\p{Lu}\p{Ll}+\s+(?<numeral>${SMALL_NUMERAL})(?![\p{L}\p{N}_&])`, "gu");
|
|
9
|
+
const LETTER = /\p{L}/u;
|
|
10
|
+
/** 文の中の、名前の後ろの番号の範囲(数字だけ)。 */
|
|
11
|
+
export const nameNumeralSpans = (text) => [...text.matchAll(NAMED_NUMERAL)]
|
|
12
|
+
.filter((match) => LETTER.test(text.slice(0, match.index)))
|
|
13
|
+
.map((match) => {
|
|
14
|
+
const numeral = match.groups?.["numeral"] ?? "";
|
|
15
|
+
const end = match.index + match[0].length;
|
|
16
|
+
return { start: end - numeral.length, end };
|
|
17
|
+
});
|
|
18
|
+
/**
|
|
19
|
+
* 文の頭か文の終わりの . の後ろに、. を付けて 1 語で立つ数字(Survey of Cool Stars. VII. The Inner Disk、II. Background)。
|
|
20
|
+
* 続き物や区切りの番号で、略語ではない。日本語の文分けは英文の . で切らないので、文の中の . の後ろも見る。
|
|
21
|
+
* C・D・L・M を含む数字(CD.)は略語でもあるので外さない。
|
|
22
|
+
*/
|
|
23
|
+
// 前置きは後ろ読みにしない。長さの決まらない後ろ読みは、長い空白の上で位置ごとに遡って二乗の時間がかかる。
|
|
24
|
+
const LONE_NUMERAL = new RegExp(String.raw `(?<lead>^\s*|[.。]\s+)(?<numeral>${SMALL_NUMERAL})(?=\.(?:\s|$))`, "gu");
|
|
25
|
+
/** 文の中の、1 語で立つ番号の範囲(数字だけ)。 */
|
|
26
|
+
export const loneNumeralSpans = (text) => [...text.matchAll(LONE_NUMERAL)].map((match) => {
|
|
27
|
+
const start = match.index + (match.groups?.["lead"] ?? "").length;
|
|
28
|
+
return { start, end: start + (match.groups?.["numeral"] ?? "").length };
|
|
29
|
+
});
|
|
30
|
+
//# sourceMappingURL=name-numeral.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"name-numeral.js","sourceRoot":"","sources":["../../src/detectors/name-numeral.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAIH,MAAM,aAAa,GAAG,mCAAmC,CAAC;AAC1D,MAAM,aAAa,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,gDAAgD,aAAa,qBAAqB,EAAE,IAAI,CAAC,CAAC;AACrI,MAAM,MAAM,GAAG,QAAQ,CAAC;AAExB,8BAA8B;AAC9B,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,IAAY,EAAU,EAAE,CACvD,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,aAAa,CAAC,CAAC;KAC9B,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC;KAC1D,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE;IACb,MAAM,OAAO,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC;IAChD,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAC1C,OAAO,EAAE,KAAK,EAAE,GAAG,GAAG,OAAO,CAAC,MAAM,EAAE,GAAG,EAAE,CAAC;AAC9C,CAAC,CAAC,CAAC;AAEP;;;;GAIG;AACH,uDAAuD;AACvD,MAAM,YAAY,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,mCAAmC,aAAa,iBAAiB,EAAE,IAAI,CAAC,CAAC;AAEnH,8BAA8B;AAC9B,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,IAAY,EAAU,EAAE,CACvD,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE;IAC7C,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,GAAG,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;IAClE,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,KAAK,GAAG,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC;AAC1E,CAAC,CAAC,CAAC"}
|
|
@@ -0,0 +1,8 @@
|
|
|
1
|
+
import type { Token } from "../plugin.ts";
|
|
2
|
+
/**
|
|
3
|
+
* 語(添字)が、大文字で書いた名前の一部か。文の途中で大文字で始まり、隣の語も文の途中の大文字(Procedures Guide、
|
|
4
|
+
* the Location Object、Learning & Development)。解析器はこういう語を動詞と読むことがあるが、述語ではない。
|
|
5
|
+
* 文頭の語は隣に数えない。Select Save (if applicable) の Save は述語のまま。
|
|
6
|
+
*/
|
|
7
|
+
export declare const isNamePart: (tokens: readonly Token[], index: number) => boolean;
|
|
8
|
+
//# sourceMappingURL=name-part.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"name-part.d.ts","sourceRoot":"","sources":["../../src/detectors/name-part.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,KAAK,EAAE,MAAM,cAAc,CAAC;AAc1C;;;;GAIG;AACH,eAAO,MAAM,UAAU,GAAI,QAAQ,SAAS,KAAK,EAAE,EAAE,OAAO,MAAM,KAAG,OAIpE,CAAC"}
|
|
@@ -0,0 +1,20 @@
|
|
|
1
|
+
const LETTER = /\p{L}/u;
|
|
2
|
+
const CAPITAL_HEAD = /^\p{Lu}/u;
|
|
3
|
+
/** 文字を含む語の添字。&、括弧、句読点は名前の並びを切らない(Learning & Development、Guide (PAPPG))。 */
|
|
4
|
+
const wordIndexes = (tokens) => tokens.flatMap((token, index) => (LETTER.test(token.surface) ? [index] : []));
|
|
5
|
+
/** 文の最初の語より後ろで、大文字で始まる語。文頭の大文字は位置のしるしで、名前のしるしではない。 */
|
|
6
|
+
const isCapitalisedAt = (tokens, words, at) => {
|
|
7
|
+
const index = words[at];
|
|
8
|
+
return at > 0 && index !== undefined && CAPITAL_HEAD.test(tokens[index]?.surface ?? "");
|
|
9
|
+
};
|
|
10
|
+
/**
|
|
11
|
+
* 語(添字)が、大文字で書いた名前の一部か。文の途中で大文字で始まり、隣の語も文の途中の大文字(Procedures Guide、
|
|
12
|
+
* the Location Object、Learning & Development)。解析器はこういう語を動詞と読むことがあるが、述語ではない。
|
|
13
|
+
* 文頭の語は隣に数えない。Select Save (if applicable) の Save は述語のまま。
|
|
14
|
+
*/
|
|
15
|
+
export const isNamePart = (tokens, index) => {
|
|
16
|
+
const words = wordIndexes(tokens);
|
|
17
|
+
const at = words.indexOf(index);
|
|
18
|
+
return isCapitalisedAt(tokens, words, at) && (isCapitalisedAt(tokens, words, at - 1) || isCapitalisedAt(tokens, words, at + 1));
|
|
19
|
+
};
|
|
20
|
+
//# sourceMappingURL=name-part.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"name-part.js","sourceRoot":"","sources":["../../src/detectors/name-part.ts"],"names":[],"mappings":"AAEA,MAAM,MAAM,GAAG,QAAQ,CAAC;AACxB,MAAM,YAAY,GAAG,UAAU,CAAC;AAEhC,2EAA2E;AAC3E,MAAM,WAAW,GAAG,CAAC,MAAwB,EAAY,EAAE,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;AAE1I,sDAAsD;AACtD,MAAM,eAAe,GAAG,CAAC,MAAwB,EAAE,KAAwB,EAAE,EAAU,EAAW,EAAE;IAClG,MAAM,KAAK,GAAG,KAAK,CAAC,EAAE,CAAC,CAAC;IACxB,OAAO,EAAE,GAAG,CAAC,IAAI,KAAK,KAAK,SAAS,IAAI,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,OAAO,IAAI,EAAE,CAAC,CAAC;AAC1F,CAAC,CAAC;AAEF;;;;GAIG;AACH,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,MAAwB,EAAE,KAAa,EAAW,EAAE;IAC7E,MAAM,KAAK,GAAG,WAAW,CAAC,MAAM,CAAC,CAAC;IAClC,MAAM,EAAE,GAAG,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;IAChC,OAAO,eAAe,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,eAAe,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,GAAG,CAAC,CAAC,IAAI,eAAe,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC;AAClI,CAAC,CAAC"}
|
|
@@ -0,0 +1,8 @@
|
|
|
1
|
+
import type { Token } from "../plugin.ts";
|
|
2
|
+
/**
|
|
3
|
+
* 漢字の連なりの語(covering: tokens の添字)が 1 つの名前か。辞書が固有名詞と読む 1 語(日本銀行)か、人の姓と名。
|
|
4
|
+
* 名前は書き手が変えられない。固有名詞に普通の語が続けば(武蔵野美術大学造形構想学部、東京都知事選挙管理委員会事務局)、
|
|
5
|
+
* 名前から組み立てた語で、どこで切れるかは読み手に見えない。名前が続けば(日本銀行日本電信電話)、並び。どちらも名前とは言わない。
|
|
6
|
+
*/
|
|
7
|
+
export declare const isOneName: (tokens: readonly Token[], covering: readonly number[]) => boolean;
|
|
8
|
+
//# sourceMappingURL=name-run.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"name-run.d.ts","sourceRoot":"","sources":["../../src/detectors/name-run.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,KAAK,EAAE,MAAM,cAAc,CAAC;AAO1C;;;;GAIG;AACH,eAAO,MAAM,SAAS,GAAI,QAAQ,SAAS,KAAK,EAAE,EAAE,UAAU,SAAS,MAAM,EAAE,KAAG,OAGjF,CAAC"}
|
|
@@ -0,0 +1,13 @@
|
|
|
1
|
+
const nameType = (token) => token?.features?.["NameType"];
|
|
2
|
+
/** 人の姓と名(田中太郎)。姓が 2 つ続けば(田中山田)、名前の並び。 */
|
|
3
|
+
const isFullName = (named) => named.length === 2 && nameType(named[0]) === "Sur" && nameType(named[1]) === "Giv";
|
|
4
|
+
/**
|
|
5
|
+
* 漢字の連なりの語(covering: tokens の添字)が 1 つの名前か。辞書が固有名詞と読む 1 語(日本銀行)か、人の姓と名。
|
|
6
|
+
* 名前は書き手が変えられない。固有名詞に普通の語が続けば(武蔵野美術大学造形構想学部、東京都知事選挙管理委員会事務局)、
|
|
7
|
+
* 名前から組み立てた語で、どこで切れるかは読み手に見えない。名前が続けば(日本銀行日本電信電話)、並び。どちらも名前とは言わない。
|
|
8
|
+
*/
|
|
9
|
+
export const isOneName = (tokens, covering) => {
|
|
10
|
+
const named = covering.map((index) => tokens[index]);
|
|
11
|
+
return (named.length === 1 && named[0]?.pos === "PROPN") || isFullName(named);
|
|
12
|
+
};
|
|
13
|
+
//# sourceMappingURL=name-run.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"name-run.js","sourceRoot":"","sources":["../../src/detectors/name-run.ts"],"names":[],"mappings":"AAEA,MAAM,QAAQ,GAAG,CAAC,KAAwB,EAAsB,EAAE,CAAC,KAAK,EAAE,QAAQ,EAAE,CAAC,UAAU,CAAC,CAAC;AAEjG,yCAAyC;AACzC,MAAM,UAAU,GAAG,CAAC,KAAqC,EAAW,EAAE,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,KAAK,IAAI,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,KAAK,CAAC;AAE1J;;;;GAIG;AACH,MAAM,CAAC,MAAM,SAAS,GAAG,CAAC,MAAwB,EAAE,QAA2B,EAAW,EAAE;IAC1F,MAAM,KAAK,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC;IACrD,OAAO,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,EAAE,GAAG,KAAK,OAAO,CAAC,IAAI,UAAU,CAAC,KAAK,CAAC,CAAC;AAChF,CAAC,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"orthography.d.ts","sourceRoot":"","sources":["../../src/detectors/orthography.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,
|
|
1
|
+
{"version":3,"file":"orthography.d.ts","sourceRoot":"","sources":["../../src/detectors/orthography.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAA2B,MAAM,cAAc,CAAC;AAKtE,6DAA6D;AAC7D,eAAO,MAAM,aAAa,EAAE,QAc3B,CAAC;AA0BF;;;GAGG;AACH,eAAO,MAAM,YAAY,EAAE,QAiC1B,CAAC"}
|
|
@@ -1,5 +1,6 @@
|
|
|
1
1
|
import { latinBoundaries, minorityStyle, occurrencesOutside } from "../orthography.js";
|
|
2
|
-
import {
|
|
2
|
+
import { isWithinAny, quotedSpans } from "../quoted-span.js";
|
|
3
|
+
import { digitRunAround, isNumberName, sequenceLabelStarts } from "../number-name.js";
|
|
3
4
|
/** チームが chaff.yaml の prefer に書いた「使わない書き方」。書いていなければ何も言わない。 */
|
|
4
5
|
export const preferredTerm = (doc, options) => {
|
|
5
6
|
const pairs = (options.lexicon ?? []).flatMap((entry) => (entry.instead_of === undefined ? [] : [{ avoid: entry.pattern, use: entry.instead_of }]));
|
|
@@ -24,20 +25,31 @@ const digitBeside = (boundary) => {
|
|
|
24
25
|
return boundary.spaced ? boundary.offset + 1 : boundary.offset;
|
|
25
26
|
};
|
|
26
27
|
/** 番号・識別子として書かれた数の境目は、空け方の好みではないので数えない(number-name.ts)。 */
|
|
27
|
-
const isCounted = (sentence, boundary) => {
|
|
28
|
+
const isCounted = (sentence, boundary, sequence, topUnits) => {
|
|
28
29
|
if (boundary.kind === "letter")
|
|
29
30
|
return true;
|
|
30
31
|
const run = digitRunAround(sentence.text, digitBeside(boundary));
|
|
31
|
-
return run === undefined || !isNumberName(sentence.text, run, sentence.tokens, sentence.span.start);
|
|
32
|
+
return run === undefined || !isNumberName(sentence.text, run, sentence.tokens, sentence.span.start, sequence, topUnits);
|
|
32
33
|
};
|
|
34
|
+
/**
|
|
35
|
+
* 鉤括弧で引いた題名や発言(「…ガイドライン ver. 1.1」)の中の境目。空け方は引いた元のもので、書き手の書き方ではない。
|
|
36
|
+
* 括弧は日本語の字でも英数字でもないので、境目はまるごと括弧の中か外にある。
|
|
37
|
+
*/
|
|
38
|
+
const isQuoted = (quoted, boundary) => isWithinAny(quoted, { start: boundary.offset, end: boundary.offset });
|
|
33
39
|
/**
|
|
34
40
|
* 日本語と英字・数字のあいだを、空けるか詰めるか。文書の中で混ざっていたら、少ないほうを指摘する。
|
|
35
41
|
* どちらが正しいかは決めない。決めるのはチームで、chaff はそろっているかだけを見る。
|
|
36
42
|
*/
|
|
37
43
|
export const latinSpacing = (doc, options) => {
|
|
38
|
-
|
|
39
|
-
|
|
40
|
-
|
|
44
|
+
// 覆った文(prose)で探す。コードの中の「1 件」は並びに入れない。
|
|
45
|
+
const sequence = sequenceLabelStarts(doc.prose ?? doc.source);
|
|
46
|
+
const topUnits = new Set((doc.lexicons["prefecture-unit"] ?? []).map((entry) => entry.pattern));
|
|
47
|
+
const located = doc.sentences.flatMap((sentence) => {
|
|
48
|
+
const quoted = quotedSpans(sentence.text);
|
|
49
|
+
return latinBoundaries(sentence.text, doc.source.slice(sentence.span.start, sentence.span.end))
|
|
50
|
+
.filter((boundary) => !isQuoted(quoted, boundary) && isCounted(sentence, boundary, sequence, topUnits))
|
|
51
|
+
.map((boundary) => ({ sentence, ...boundary }));
|
|
52
|
+
});
|
|
41
53
|
return KINDS.flatMap((kind) => {
|
|
42
54
|
const ofKind = located.filter((entry) => entry.kind === kind);
|
|
43
55
|
const minority = minorityStyle(ofKind);
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"orthography.js","sourceRoot":"","sources":["../../src/detectors/orthography.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,eAAe,EAAE,aAAa,EAAE,kBAAkB,EAAmC,MAAM,mBAAmB,CAAC;AACxH,OAAO,EAAE,cAAc,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;
|
|
1
|
+
{"version":3,"file":"orthography.js","sourceRoot":"","sources":["../../src/detectors/orthography.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,eAAe,EAAE,aAAa,EAAE,kBAAkB,EAAmC,MAAM,mBAAmB,CAAC;AACxH,OAAO,EAAE,WAAW,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAC7D,OAAO,EAAE,cAAc,EAAE,YAAY,EAAE,mBAAmB,EAAE,MAAM,mBAAmB,CAAC;AAEtF,6DAA6D;AAC7D,MAAM,CAAC,MAAM,aAAa,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IACjE,MAAM,KAAK,GAAG,CAAC,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,OAAO,EAAE,GAAG,EAAE,KAAK,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,CAAC;IACpJ,MAAM,IAAI,GAAG,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CAC9C,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,kBAAkB,CAAC,QAAQ,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CACvH,CAAC;IACF,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,IAAI,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK;QAAE,OAAO,EAAE,CAAC;IAChE,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;QACxB,IAAI,EAAE,EAAE;QACR,QAAQ,EAAE,SAAS;QACnB,IAAI,EAAE,CAAC;QACP,MAAM,EAAE,CAAC;QACT,KAAK,EAAE,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;QAC/B,MAAM,EAAE,EAAE,OAAO,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,EAAE,SAAS,EAAE,GAAG,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,EAAE,IAAI,CAAC,MAAM,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,GAAG,CAAC,EAAE,EAAE;KACjJ,CAAC,CAAC,CAAC;AACN,CAAC,CAAC;AAEF,MAAM,SAAS,GAA0C,EAAE,MAAM,EAAE,IAAI,EAAE,cAAc,EAAE,OAAO,EAAE,aAAa,EAAE,MAAM,EAAE,CAAC;AAC1H,MAAM,KAAK,GAA2B,CAAC,QAAQ,EAAE,cAAc,EAAE,aAAa,CAAC,CAAC;AAIhF,0DAA0D;AAC1D,MAAM,WAAW,GAAG,CAAC,QAAkB,EAAU,EAAE;IACjD,IAAI,QAAQ,CAAC,IAAI,KAAK,aAAa;QAAE,OAAO,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC;IAChE,OAAO,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC;AACjE,CAAC,CAAC;AAEF,2DAA2D;AAC3D,MAAM,SAAS,GAAG,CAAC,QAAkB,EAAE,QAAkB,EAAE,QAA6B,EAAE,QAA6B,EAAW,EAAE;IAClI,IAAI,QAAQ,CAAC,IAAI,KAAK,QAAQ;QAAE,OAAO,IAAI,CAAC;IAC5C,MAAM,GAAG,GAAG,cAAc,CAAC,QAAQ,CAAC,IAAI,EAAE,WAAW,CAAC,QAAQ,CAAC,CAAC,CAAC;IACjE,OAAO,GAAG,KAAK,SAAS,IAAI,CAAC,YAAY,CAAC,QAAQ,CAAC,IAAI,EAAE,GAAG,EAAE,QAAQ,CAAC,MAAM,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAC;AAC1H,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,QAAQ,GAAG,CAAC,MAAuB,EAAE,QAAkB,EAAW,EAAE,CAAC,WAAW,CAAC,MAAM,EAAE,EAAE,KAAK,EAAE,QAAQ,CAAC,MAAM,EAAE,GAAG,EAAE,QAAQ,CAAC,MAAM,EAAE,CAAC,CAAC;AAEjJ;;;GAGG;AACH,MAAM,CAAC,MAAM,YAAY,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAChE,sCAAsC;IACtC,MAAM,QAAQ,GAAG,mBAAmB,CAAC,GAAG,CAAC,KAAK,IAAI,GAAG,CAAC,MAAM,CAAC,CAAC;IAC9D,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,iBAAiB,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC;IAChG,MAAM,OAAO,GAAc,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE;QAC5D,MAAM,MAAM,GAAG,WAAW,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;QAC1C,OAAO,eAAe,CAAC,QAAQ,CAAC,IAAI,EAAE,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;aAC5F,MAAM,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,QAAQ,CAAC,MAAM,EAAE,QAAQ,CAAC,IAAI,SAAS,CAAC,QAAQ,EAAE,QAAQ,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAC;aACtG,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,GAAG,QAAQ,EAAE,CAAC,CAAC,CAAC;IACpD,CAAC,CAAC,CAAC;IACH,OAAO,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE;QAC5B,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,IAAI,CAAC,CAAC;QAC9D,MAAM,QAAQ,GAAG,aAAa,CAAC,MAAM,CAAC,CAAC;QACvC,IAAI,QAAQ,KAAK,SAAS;YAAE,OAAO,EAAE,CAAC;QACtC,MAAM,GAAG,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAC;QAChE,IAAI,GAAG,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK;YAAE,OAAO,EAAE,CAAC;QAC1C,OAAO,GAAG,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;YACzB,IAAI,EAAE,EAAE;YACR,QAAQ,EAAE,SAAS;YACnB,IAAI,EAAE,CAAC;YACP,MAAM,EAAE,CAAC;YACT,KAAK,EAAE,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;YACjC,MAAM,EAAE;gBACN,IAAI,EAAE,SAAS,CAAC,IAAI,CAAC;gBACrB,KAAK,EAAE,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,QAAQ;gBACrC,KAAK,EAAE,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK;gBAC/B,KAAK,EAAE,GAAG,CAAC,MAAM;gBACjB,EAAE,EAAE,MAAM,CAAC,MAAM;gBACjB,KAAK,EAAE,OAAO,CAAC,KAAK;gBACpB,MAAM,EAAE,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,MAAM;aACjD;SACF,CAAC,CAAC,CAAC;IACN,CAAC,CAAC,CAAC;AACL,CAAC,CAAC"}
|
|
@@ -0,0 +1,14 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* 日付や時刻の数字の代わりに置く、同じ大文字の繰り返し(令和YY年MM月DD日 HH:MM、MM/DD/YYYY、YYYY-MM-DD)。
|
|
3
|
+
* 定型文の書式で、略語ではない。繰り返しの語が、日付・時刻の単位の語(年、月、日)の前にあるか、
|
|
4
|
+
* 区切り(/、:、-、.)で別の繰り返しの語か数字と繋がっているときだけ認める。単位の語は言語パッケージの語彙表から読む。
|
|
5
|
+
* AI時代の AI や CI/CD の CD は、同じ文字の繰り返しではないので略語のまま数える。
|
|
6
|
+
*/
|
|
7
|
+
type Span = {
|
|
8
|
+
readonly start: number;
|
|
9
|
+
readonly end: number;
|
|
10
|
+
};
|
|
11
|
+
/** 文の中の、日付・時刻の書式に置いた大文字の範囲。units は日付・時刻の単位の語。 */
|
|
12
|
+
export declare const placeholderSpans: (text: string, units: readonly string[]) => Span[];
|
|
13
|
+
export {};
|
|
14
|
+
//# sourceMappingURL=placeholder.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"placeholder.d.ts","sourceRoot":"","sources":["../../src/detectors/placeholder.ts"],"names":[],"mappings":"AAEA;;;;;GAKG;AAEH,KAAK,IAAI,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAA;CAAE,CAAC;AA2B7D,iDAAiD;AACjD,eAAO,MAAM,gBAAgB,GAAI,MAAM,MAAM,EAAE,OAAO,SAAS,MAAM,EAAE,KAAG,IAAI,EAA6D,CAAC"}
|
|
@@ -0,0 +1,24 @@
|
|
|
1
|
+
import { escapeRegExp } from "../orthography.js";
|
|
2
|
+
const REPEATED = /^(?<letter>[A-Z])\k<letter>+$/u;
|
|
3
|
+
const PART = String.raw `(?:[A-Z]{2,}|\d+)`;
|
|
4
|
+
const SEPARATOR = "[//::.-]";
|
|
5
|
+
const JOINED = new RegExp(String.raw `(?<![A-Za-z0-9_&])${PART}(?:${SEPARATOR}${PART})+(?![A-Za-z0-9_&])`, "gu");
|
|
6
|
+
const CAPITALS = /[A-Z]+/gu;
|
|
7
|
+
const isRepeated = (word) => REPEATED.test(word);
|
|
8
|
+
const spanOf = (match) => ({ start: match.index, end: match.index + match[0].length });
|
|
9
|
+
/** 区切りで繋いだ一続きのうち、大文字の部品がどれも同じ文字の繰り返しのもの(数字だけの 2025/01/31 は略語と関係がない)。 */
|
|
10
|
+
const joinedSpans = (text) => [...text.matchAll(JOINED)]
|
|
11
|
+
.filter((match) => {
|
|
12
|
+
const capitals = match[0].match(CAPITALS) ?? [];
|
|
13
|
+
return capitals.length > 0 && capitals.every(isRepeated);
|
|
14
|
+
})
|
|
15
|
+
.map(spanOf);
|
|
16
|
+
const beforeUnitSpans = (text, units) => {
|
|
17
|
+
if (units.length === 0)
|
|
18
|
+
return [];
|
|
19
|
+
const pattern = new RegExp(String.raw `(?<![A-Za-z0-9_&])[A-Z]{2,}(?=${units.map(escapeRegExp).join("|")})`, "gu");
|
|
20
|
+
return [...text.matchAll(pattern)].filter((match) => isRepeated(match[0])).map(spanOf);
|
|
21
|
+
};
|
|
22
|
+
/** 文の中の、日付・時刻の書式に置いた大文字の範囲。units は日付・時刻の単位の語。 */
|
|
23
|
+
export const placeholderSpans = (text, units) => [...joinedSpans(text), ...beforeUnitSpans(text, units)];
|
|
24
|
+
//# sourceMappingURL=placeholder.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"placeholder.js","sourceRoot":"","sources":["../../src/detectors/placeholder.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AAWjD,MAAM,QAAQ,GAAG,gCAAgC,CAAC;AAClD,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAA,mBAAmB,CAAC;AAC3C,MAAM,SAAS,GAAG,UAAU,CAAC;AAC7B,MAAM,MAAM,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,qBAAqB,IAAI,MAAM,SAAS,GAAG,IAAI,qBAAqB,EAAE,IAAI,CAAC,CAAC;AAChH,MAAM,QAAQ,GAAG,UAAU,CAAC;AAE5B,MAAM,UAAU,GAAG,CAAC,IAAY,EAAW,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAElE,MAAM,MAAM,GAAG,CAAC,KAAsB,EAAQ,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,GAAG,EAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC;AAE9G,wEAAwE;AACxE,MAAM,WAAW,GAAG,CAAC,IAAY,EAAU,EAAE,CAC3C,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;KACvB,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE;IAChB,MAAM,QAAQ,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC;IAChD,OAAO,QAAQ,CAAC,MAAM,GAAG,CAAC,IAAI,QAAQ,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC;AAC3D,CAAC,CAAC;KACD,GAAG,CAAC,MAAM,CAAC,CAAC;AAEjB,MAAM,eAAe,GAAG,CAAC,IAAY,EAAE,KAAwB,EAAU,EAAE;IACzE,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAClC,MAAM,OAAO,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,iCAAiC,KAAK,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;IAClH,OAAO,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;AACzF,CAAC,CAAC;AAEF,iDAAiD;AACjD,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,IAAY,EAAE,KAAwB,EAAU,EAAE,CAAC,CAAC,GAAG,WAAW,CAAC,IAAI,CAAC,EAAE,GAAG,eAAe,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"polite-word.d.ts","sourceRoot":"","sources":["../../src/detectors/polite-word.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,YAAY,EAAE,KAAK,EAAE,MAAM,cAAc,CAAC;AAWxD,sDAAsD;AACtD,eAAO,MAAM,YAAY,GAAI,OAAO,KAAK,EAAE,SAAS,SAAS,YAAY,EAAE,KAAG,OACsD,CAAC"}
|
|
@@ -0,0 +1,10 @@
|
|
|
1
|
+
/** 語彙表の語が解析器で一語になったときの、その語。二語以上(「ございます」)は読みで照らさない。 */
|
|
2
|
+
const singleWord = (entry) => (entry.tokens?.length === 1 ? entry.tokens[0] : undefined);
|
|
3
|
+
/**
|
|
4
|
+
* 読みと品詞が同じなら、表記が違っても同じ語の同じ形。解析器は「下さい」の原形を「下さる」、「ください」の原形を「くださる」と
|
|
5
|
+
* 別の字で返すが、読み(クダサイ)と品詞は同じ。品詞も比べるのは、読みだけだと「升(マス)」が「ます」に当たるから。
|
|
6
|
+
*/
|
|
7
|
+
const sameReading = (token, word) => word?.reading !== undefined && token.reading === word.reading && token.pos === word.pos;
|
|
8
|
+
/** 語が丁寧語の語彙表の語か。書いた形か原形が語彙表の語と同じか、語彙表の一語と読みと品詞が同じ。 */
|
|
9
|
+
export const isPoliteWord = (token, entries) => entries.some((entry) => token.surface === entry.pattern || token.lemma === entry.pattern || sameReading(token, singleWord(entry)));
|
|
10
|
+
//# sourceMappingURL=polite-word.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"polite-word.js","sourceRoot":"","sources":["../../src/detectors/polite-word.ts"],"names":[],"mappings":"AAEA,sDAAsD;AACtD,MAAM,UAAU,GAAG,CAAC,KAAmB,EAAqB,EAAE,CAAC,CAAC,KAAK,CAAC,MAAM,EAAE,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC;AAE1H;;;GAGG;AACH,MAAM,WAAW,GAAG,CAAC,KAAY,EAAE,IAAuB,EAAW,EAAE,CAAC,IAAI,EAAE,OAAO,KAAK,SAAS,IAAI,KAAK,CAAC,OAAO,KAAK,IAAI,CAAC,OAAO,IAAI,KAAK,CAAC,GAAG,KAAK,IAAI,CAAC,GAAG,CAAC;AAEhK,sDAAsD;AACtD,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,KAAY,EAAE,OAAgC,EAAW,EAAE,CACtF,OAAO,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,KAAK,KAAK,CAAC,OAAO,IAAI,KAAK,CAAC,KAAK,KAAK,KAAK,CAAC,OAAO,IAAI,WAAW,CAAC,KAAK,EAAE,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC"}
|
|
@@ -0,0 +1,7 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* 文から、見出しとは違う形を引用した部分を抜いた文字列。
|
|
3
|
+
* 「Not “datacentre”.」の “datacentre” は語の書き方として挙げたもの(言及)で、見出し「data centre」を繰り返してはいない。
|
|
4
|
+
* 見出しそのままの引用(「利用規約への同意」について説明します)は見出しを指しているので残す。
|
|
5
|
+
*/
|
|
6
|
+
export declare const withoutQuotedVariants: (sentence: string, heading: string) => string;
|
|
7
|
+
//# sourceMappingURL=quoted-variant.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"quoted-variant.d.ts","sourceRoot":"","sources":["../../src/detectors/quoted-variant.ts"],"names":[],"mappings":"AAaA;;;;GAIG;AACH,eAAO,MAAM,qBAAqB,GAAI,UAAU,MAAM,EAAE,SAAS,MAAM,KAAG,MAC6E,CAAC"}
|
|
@@ -0,0 +1,18 @@
|
|
|
1
|
+
// 引用符の中身。英語の ’ と ' はアポストロフィ(organisation’s, don't)にも使われるので、字のあとに続くものは閉じとみなさない。
|
|
2
|
+
const QUOTATIONS = [
|
|
3
|
+
/“([^“”\n]*)”/gu,
|
|
4
|
+
/‘((?:[^‘’\n]|’(?=\p{L}))*)’(?!\p{L})/gu,
|
|
5
|
+
/"([^"\n]*)"/gu,
|
|
6
|
+
/(?<![\p{L}\p{N}])'((?:[^'\n]|'(?=\p{L}))*)'(?![\p{L}\p{N}])/gu,
|
|
7
|
+
/「([^「」\n]*)」/gu,
|
|
8
|
+
/『([^『』\n]*)』/gu,
|
|
9
|
+
];
|
|
10
|
+
// 大文字小文字は畳まない。表記の手引きは大文字小文字だけ違う形も引用して退ける(Not “Open Source software”)。
|
|
11
|
+
const sameText = (left, right) => left.trim() === right.trim();
|
|
12
|
+
/**
|
|
13
|
+
* 文から、見出しとは違う形を引用した部分を抜いた文字列。
|
|
14
|
+
* 「Not “datacentre”.」の “datacentre” は語の書き方として挙げたもの(言及)で、見出し「data centre」を繰り返してはいない。
|
|
15
|
+
* 見出しそのままの引用(「利用規約への同意」について説明します)は見出しを指しているので残す。
|
|
16
|
+
*/
|
|
17
|
+
export const withoutQuotedVariants = (sentence, heading) => QUOTATIONS.reduce((text, quotation) => text.replace(quotation, (whole, inner) => (sameText(inner, heading) ? whole : " ")), sentence);
|
|
18
|
+
//# sourceMappingURL=quoted-variant.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"quoted-variant.js","sourceRoot":"","sources":["../../src/detectors/quoted-variant.ts"],"names":[],"mappings":"AAAA,+EAA+E;AAC/E,MAAM,UAAU,GAAsB;IACpC,gBAAgB;IAChB,wCAAwC;IACxC,eAAe;IACf,+DAA+D;IAC/D,gBAAgB;IAChB,gBAAgB;CACjB,CAAC;AAEF,sEAAsE;AACtE,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAE,KAAa,EAAW,EAAE,CAAC,IAAI,CAAC,IAAI,EAAE,KAAK,KAAK,CAAC,IAAI,EAAE,CAAC;AAExF;;;;GAIG;AACH,MAAM,CAAC,MAAM,qBAAqB,GAAG,CAAC,QAAgB,EAAE,OAAe,EAAU,EAAE,CACjF,UAAU,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,SAAS,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE,CAAC,KAAa,EAAE,KAAa,EAAE,EAAE,CAAC,CAAC,QAAQ,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,QAAQ,CAAC,CAAC"}
|
|
@@ -1,12 +1,12 @@
|
|
|
1
|
-
import type { Sentence, Span, Token } from "../plugin.ts";
|
|
1
|
+
import type { LexiconEntry, Sentence, Span, Token } from "../plugin.ts";
|
|
2
2
|
export type Register = "polite" | "plain";
|
|
3
3
|
/**
|
|
4
4
|
* 文末の語の調子。preceding は文末の語より前の語。
|
|
5
5
|
* 文末に述語(動詞・形容詞・助動詞)が無ければ、ですます調でもである調でもないので undefined。
|
|
6
6
|
* 「以下の通り。」「円錐形の麦わら帽子。」のような名詞で終わる文を数えると、そのままである調の少数派になっていた。
|
|
7
|
-
*
|
|
7
|
+
* 書いた形でも原形でも読みでも当てる(polite-word.ts)。「ください」の原形は「くださる」で、原形だけを見ると丁寧な文末を見落とす。
|
|
8
8
|
*/
|
|
9
|
-
export declare const registerOf: (ending: readonly Token[], preceding: readonly Token[], polite: readonly
|
|
9
|
+
export declare const registerOf: (ending: readonly Token[], preceding: readonly Token[], polite: readonly LexiconEntry[]) => Register | undefined;
|
|
10
10
|
/**
|
|
11
11
|
* 「!」「?」で切れた文が、間を置かずに助詞で始まる次の文へ続くか。「作っていた!!という人」「導入しませんか?が断られた」は、
|
|
12
12
|
* 分割器が切っても一つの文で、切れ目の手前は書き手の文末ではない。「。」で切れた文は続きとは見ない。
|
|
@@ -14,7 +14,9 @@ export declare const registerOf: (ending: readonly Token[], preceding: readonly
|
|
|
14
14
|
export declare const continuesInto: (sentence: Sentence, next: Sentence | undefined) => boolean;
|
|
15
15
|
/** 位置 offset を含む一番外側の箇条書きの始まり。本文なら undefined。入れ子の項目は外側の箇条書きと一緒に見る。 */
|
|
16
16
|
export declare const outermostList: (offset: number, lists: readonly Span[]) => number | undefined;
|
|
17
|
-
/**
|
|
17
|
+
/** 文が入っているまとまりの始まり。一番外側の箇条書き、それが無ければ番号で始まる段落の並び(runs)。本文なら undefined。 */
|
|
18
|
+
export declare const groupOf: (offset: number, lists: readonly Span[], runs: readonly Span[]) => number | undefined;
|
|
19
|
+
/** group: 文が入っているまとまり(groupOf の値)。本文なら undefined。 */
|
|
18
20
|
export type Judged = {
|
|
19
21
|
readonly register: Register;
|
|
20
22
|
readonly group: number | undefined;
|
|
@@ -25,7 +27,7 @@ export type Slip<T extends Judged> = {
|
|
|
25
27
|
readonly count: number;
|
|
26
28
|
};
|
|
27
29
|
/**
|
|
28
|
-
* 調子が混ざった文を、judged
|
|
30
|
+
* 調子が混ざった文を、judged の順で。本文は本文どうし、箇条書きと番号で始まる段落の並びは 1 つずつ、その中で揃っているかを見る。
|
|
29
31
|
* ですます調の本文に常体の箇条書きを置くのはよくある書き方で、箇条書きが丸ごと揃っていれば混在ではない。
|
|
30
32
|
*/
|
|
31
33
|
export declare const slipsOf: <T extends Judged>(judged: readonly T[], limit: number) => Slip<T>[];
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"register.d.ts","sourceRoot":"","sources":["../../src/detectors/register.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"register.d.ts","sourceRoot":"","sources":["../../src/detectors/register.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,YAAY,EAAE,QAAQ,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,cAAc,CAAC;AAExE,MAAM,MAAM,QAAQ,GAAG,QAAQ,GAAG,OAAO,CAAC;AAqB1C;;;;;GAKG;AACH,eAAO,MAAM,UAAU,GAAI,QAAQ,SAAS,KAAK,EAAE,EAAE,WAAW,SAAS,KAAK,EAAE,EAAE,QAAQ,SAAS,YAAY,EAAE,KAAG,QAAQ,GAAG,SAI9H,CAAC;AAIF;;;GAGG;AACH,eAAO,MAAM,aAAa,GAAI,UAAU,QAAQ,EAAE,MAAM,QAAQ,GAAG,SAAS,KAAG,OACkD,CAAC;AAElI,sEAAsE;AACtE,eAAO,MAAM,aAAa,GAAI,QAAQ,MAAM,EAAE,OAAO,SAAS,IAAI,EAAE,KAAG,MAAM,GAAG,SAG6C,CAAC;AAE9H,yEAAyE;AACzE,eAAO,MAAM,OAAO,GAAI,QAAQ,MAAM,EAAE,OAAO,SAAS,IAAI,EAAE,EAAE,MAAM,SAAS,IAAI,EAAE,KAAG,MAAM,GAAG,SACpC,CAAC;AAE9D,qDAAqD;AACrD,MAAM,MAAM,MAAM,GAAG;IAAE,QAAQ,CAAC,QAAQ,EAAE,QAAQ,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,GAAG,SAAS,CAAA;CAAE,CAAC;AAEzF,uCAAuC;AACvC,MAAM,MAAM,IAAI,CAAC,CAAC,SAAS,MAAM,IAAI;IAAE,QAAQ,CAAC,KAAK,EAAE,CAAC,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC;AAoBnF;;;GAGG;AACH,eAAO,MAAM,OAAO,GAAI,CAAC,SAAS,MAAM,EAAE,QAAQ,SAAS,CAAC,EAAE,EAAE,OAAO,MAAM,KAAG,IAAI,CAAC,CAAC,CAAC,EAGtF,CAAC"}
|