@chaffjs/lang-en 0.12.0 → 0.13.0

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.
Files changed (56) hide show
  1. package/dist/citation.d.ts +2 -0
  2. package/dist/citation.d.ts.map +1 -1
  3. package/dist/citation.js +37 -12
  4. package/dist/citation.js.map +1 -1
  5. package/dist/closing-quote.d.ts +6 -0
  6. package/dist/closing-quote.d.ts.map +1 -0
  7. package/dist/closing-quote.js +50 -0
  8. package/dist/closing-quote.js.map +1 -0
  9. package/dist/index.d.ts +1 -1
  10. package/dist/index.d.ts.map +1 -1
  11. package/dist/index.js +5 -2
  12. package/dist/index.js.map +1 -1
  13. package/dist/item-style.d.ts +30 -0
  14. package/dist/item-style.d.ts.map +1 -0
  15. package/dist/item-style.js +124 -0
  16. package/dist/item-style.js.map +1 -0
  17. package/dist/long-runs.d.ts +6 -0
  18. package/dist/long-runs.d.ts.map +1 -0
  19. package/dist/long-runs.js +6 -0
  20. package/dist/long-runs.js.map +1 -0
  21. package/dist/pos.d.ts.map +1 -1
  22. package/dist/pos.js +32 -5
  23. package/dist/pos.js.map +1 -1
  24. package/dist/proper-noun.d.ts +21 -0
  25. package/dist/proper-noun.d.ts.map +1 -1
  26. package/dist/proper-noun.js +37 -0
  27. package/dist/proper-noun.js.map +1 -1
  28. package/dist/quoted-stop.d.ts +13 -0
  29. package/dist/quoted-stop.d.ts.map +1 -0
  30. package/dist/quoted-stop.js +45 -0
  31. package/dist/quoted-stop.js.map +1 -0
  32. package/dist/sentence-split.d.ts +2 -0
  33. package/dist/sentence-split.d.ts.map +1 -1
  34. package/dist/sentence-split.js +3 -1
  35. package/dist/sentence-split.js.map +1 -1
  36. package/dist/structure.d.ts.map +1 -1
  37. package/dist/structure.js +24 -65
  38. package/dist/structure.js.map +1 -1
  39. package/lexicons/common-acronym.yaml +4 -0
  40. package/lexicons/date-time-unit.yaml +6 -0
  41. package/lexicons/emphasis-word.yaml +1 -0
  42. package/lexicons/fixed-phrase.yaml +14 -0
  43. package/lexicons/honorific.yaml +15 -0
  44. package/lexicons/http-method.yaml +14 -0
  45. package/lexicons/quantity-noun.yaml +6 -0
  46. package/package.json +1 -1
  47. package/src/citation.ts +41 -13
  48. package/src/closing-quote.ts +52 -0
  49. package/src/index.ts +5 -2
  50. package/src/item-style.ts +130 -0
  51. package/src/long-runs.ts +6 -0
  52. package/src/pos.ts +40 -5
  53. package/src/proper-noun.ts +43 -0
  54. package/src/quoted-stop.ts +52 -0
  55. package/src/sentence-split.ts +3 -1
  56. package/src/structure.ts +23 -67
@@ -14,4 +14,6 @@ export declare const citedDocumentBefore: (text: string, start: number) => strin
14
14
  * "of this Agreement" and "of the Agreement" are this document; "of the Master Agreement" is another.
15
15
  */
16
16
  export declare const citedDocumentAfter: (text: string, end: number) => string | undefined;
17
+ /** A hyphenated tag right after a reference ("Section 4.2.3 of [HTTP-CACHING]") or just before it ("[HTTP-CACHING], Section 4"). */
18
+ export declare const hyphenatedTagAround: (text: string, start: number, end: number) => string | undefined;
17
19
  //# sourceMappingURL=citation.d.ts.map
@@ -1 +1 @@
1
- {"version":3,"file":"citation.d.ts","sourceRoot":"","sources":["../src/citation.ts"],"names":[],"mappings":"AAiBA,MAAM,MAAM,UAAU,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;CAAE,CAAC;AAmB3E;;;GAGG;AACH,eAAO,MAAM,WAAW,GAAI,MAAM,MAAM,EAAE,QAAQ,OAAO,KAAG,UAAU,EAGtB,CAAC;AAWjD,6FAA6F;AAC7F,eAAO,MAAM,mBAAmB,GAAI,MAAM,MAAM,EAAE,OAAO,MAAM,KAAG,MAAM,GAAG,SACG,CAAC;AA2B/E;;;GAGG;AACH,eAAO,MAAM,kBAAkB,GAAI,MAAM,MAAM,EAAE,KAAK,MAAM,KAAG,MAAM,GAAG,SAYvE,CAAC"}
1
+ {"version":3,"file":"citation.d.ts","sourceRoot":"","sources":["../src/citation.ts"],"names":[],"mappings":"AAoBA,MAAM,MAAM,UAAU,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;CAAE,CAAC;AAmB3E;;;GAGG;AACH,eAAO,MAAM,WAAW,GAAI,MAAM,MAAM,EAAE,QAAQ,OAAO,KAAG,UAAU,EAGtB,CAAC;AAwBjD,6FAA6F;AAC7F,eAAO,MAAM,mBAAmB,GAAI,MAAM,MAAM,EAAE,OAAO,MAAM,KAAG,MAAM,GAAG,SAAiD,CAAC;AAoC7H;;;GAGG;AACH,eAAO,MAAM,kBAAkB,GAAI,MAAM,MAAM,EAAE,KAAK,MAAM,KAAG,MAAM,GAAG,SASvE,CAAC;AAEF,oIAAoI;AACpI,eAAO,MAAM,mBAAmB,GAAI,MAAM,MAAM,EAAE,OAAO,MAAM,EAAE,KAAK,MAAM,KAAG,MAAM,GAAG,SAIvF,CAAC"}
package/dist/citation.js CHANGED
@@ -3,8 +3,11 @@ const OF = /^,? of (?:the |that |those )?/u;
3
3
  /** "section 4(2)(a) (exception to liability …) of the Damages (Scotland) Act 2011": the gloss sits between the number and the name. */
4
4
  const GLOSS = /^ \([^()]{1,100}\)/u;
5
5
  const CONNECTOR = /^(?:,? (?:to|and|or)|,) /u;
6
- /** "9", "29(2)", "(g)", and a roman "V" for a list of Articles. The roman numeral must end the word: "VIII", not "Vendor". */
7
- const LISTED_NUMBERS = [/^\d{1,3}[A-Z]{0,2}(?:\([a-z0-9]{1,4}\))*/u, /^(?:\([a-z0-9]{1,4}\))+/u, /^[IVXLC]{1,7}\b/u];
6
+ /**
7
+ * "9", "29(2)", "(g)", and a roman "V" for a list of Articles. The roman numeral must end the word: "VIII", not "Vendor".
8
+ * A dotted "310.5" is not read: its "310" alone is not the section the list names.
9
+ */
10
+ const LISTED_NUMBERS = [/^\d{1,3}(?!\.?\d)[A-Z]{0,2}(?:\([a-z0-9]{1,4}\))*/u, /^(?:\([a-z0-9]{1,4}\))+/u, /^[IVXLC]{1,7}\b/u];
8
11
  const listedNumber = (text) => LISTED_NUMBERS.map((pattern) => pattern.exec(text)?.[0]).find((found) => found !== undefined);
9
12
  /**
10
13
  * A listed number is a reference only where the list goes on or ends: "Sections 1, 2 and 9 of" or "Section 3 and 4." —
@@ -39,11 +42,22 @@ export const listMembers = (rest, plural) => listed(rest)
39
42
  * word of a contract ("[Company]", "[Reserved]") is not taken for another document and still has to exist here.
40
43
  */
41
44
  const TAG = "(?<tag>[A-Z][A-Z0-9]{1,30})";
42
- const TAG_AFTER = new RegExp(`^\\[${TAG}\\]`, "u");
45
+ /**
46
+ * "[HTTP-CACHING]" is written like a contract's placeholder "[BUYER-1]". Only the document tells them apart, by listing
47
+ * the tag or not, so this tag is a candidate that the core checks against the document (attrs.citedTag).
48
+ */
49
+ const HYPHENATED_TAG = "(?<tag>[A-Z][A-Z0-9]{0,30}(?:-[A-Z0-9]{1,30}){1,4})";
50
+ const tagAfter = (tag) => new RegExp(`^\\[${tag}\\]`, "u");
43
51
  /** "[HTTP], Section 12.1": the tag written just before the reference. */
44
- const TAG_BEFORE = new RegExp(`\\[${TAG}\\],?\\s?$`, "u");
52
+ const tagBefore = (tag) => new RegExp(`\\[${tag}\\],?\\s?$`, "u");
53
+ const TAG_AFTER = tagAfter(TAG);
54
+ const TAG_BEFORE = tagBefore(TAG);
55
+ const HYPHENATED_AFTER = tagAfter(HYPHENATED_TAG);
56
+ const HYPHENATED_BEFORE = tagBefore(HYPHENATED_TAG);
57
+ const TAG_REACH = 40;
58
+ const tagEndingAt = (pattern, text, start) => pattern.exec(text.slice(Math.max(0, start - TAG_REACH), start))?.groups?.["tag"];
45
59
  /** The document cited by a tag just before a reference, as in "see [HTTP], Section 12.1". */
46
- export const citedDocumentBefore = (text, start) => TAG_BEFORE.exec(text.slice(Math.max(0, start - 40), start))?.groups?.["tag"];
60
+ export const citedDocumentBefore = (text, start) => tagEndingAt(TAG_BEFORE, text, start);
47
61
  const CAPITALISED = /^[A-Z][\w'’-]*/u;
48
62
  /**
49
63
  * How a document names itself. "Section 3 of the Agreement" in an agreement means this one.
@@ -69,24 +83,35 @@ const titleWords = (rest) => {
69
83
  }
70
84
  return words;
71
85
  };
86
+ /** What follows the "of" after a reference and its list, or undefined when no "of" follows. */
87
+ const afterOf = (text, end) => {
88
+ const rest = text.slice(end, end + 200);
89
+ const afterList = listEnd(rest);
90
+ const listed = afterList + (GLOSS.exec(rest.slice(afterList))?.[0].length ?? 0);
91
+ const of = OF.exec(rest.slice(listed));
92
+ return of === null ? undefined : rest.slice(listed + of[0].length);
93
+ };
72
94
  /**
73
95
  * The document named right after a reference, or undefined when the reference is into this document.
74
96
  * "of this Agreement" and "of the Agreement" are this document; "of the Master Agreement" is another.
75
97
  */
76
98
  export const citedDocumentAfter = (text, end) => {
77
- const rest = text.slice(end, end + 200);
78
- const afterList = listEnd(rest);
79
- const listed = afterList + (GLOSS.exec(rest.slice(afterList))?.[0].length ?? 0);
80
- const of = OF.exec(rest.slice(listed));
81
- if (of === null)
99
+ const named = afterOf(text, end);
100
+ if (named === undefined)
82
101
  return undefined;
83
- const tag = TAG_AFTER.exec(rest.slice(listed + of[0].length))?.groups?.["tag"];
102
+ const tag = TAG_AFTER.exec(named)?.groups?.["tag"];
84
103
  if (tag !== undefined)
85
104
  return tag;
86
- const words = titleWords(rest.slice(listed + of[0].length));
105
+ const words = titleWords(named);
87
106
  if (words.length === 0)
88
107
  return undefined;
89
108
  const name = words.join(" ");
90
109
  return words.length === 1 && SELF.has(name) ? undefined : name;
91
110
  };
111
+ /** A hyphenated tag right after a reference ("Section 4.2.3 of [HTTP-CACHING]") or just before it ("[HTTP-CACHING], Section 4"). */
112
+ export const hyphenatedTagAround = (text, start, end) => {
113
+ const named = afterOf(text, end);
114
+ const following = named === undefined ? undefined : HYPHENATED_AFTER.exec(named)?.groups?.["tag"];
115
+ return following ?? tagEndingAt(HYPHENATED_BEFORE, text, start);
116
+ };
92
117
  //# sourceMappingURL=citation.js.map
@@ -1 +1 @@
1
- {"version":3,"file":"citation.js","sourceRoot":"","sources":["../src/citation.ts"],"names":[],"mappings":"AAAA,2GAA2G;AAE3G,MAAM,EAAE,GAAG,gCAAgC,CAAC;AAC5C,uIAAuI;AACvI,MAAM,KAAK,GAAG,qBAAqB,CAAC;AACpC,MAAM,SAAS,GAAG,2BAA2B,CAAC;AAC9C,8HAA8H;AAC9H,MAAM,cAAc,GAAG,CAAC,2CAA2C,EAAE,0BAA0B,EAAE,kBAAkB,CAAC,CAAC;AAErH,MAAM,YAAY,GAAG,CAAC,IAAY,EAAsB,EAAE,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC;AAEzJ;;;GAGG;AACH,MAAM,UAAU,GAAG,2CAA2C,CAAC;AAI/D,sGAAsG;AACtG,MAAM,MAAM,GAAG,CAAC,IAAY,EAA6C,EAAE;IACzE,MAAM,KAAK,GAAqC,EAAE,CAAC;IACnD,IAAI,GAAG,GAAG,CAAC,CAAC;IACZ,SAAS,CAAC;QACR,MAAM,SAAS,GAAG,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;QACvD,MAAM,MAAM,GAAG,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC;QACtG,IAAI,SAAS,KAAK,SAAS,IAAI,MAAM,KAAK,SAAS;YAAE,OAAO,KAAK,CAAC;QAClE,MAAM,KAAK,GAAG,GAAG,GAAG,SAAS,CAAC,MAAM,CAAC;QACrC,GAAG,GAAG,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC;QAC5B,KAAK,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC,CAAC;IAC3C,CAAC;AACH,CAAC,CAAC;AAEF,6GAA6G;AAC7G,MAAM,OAAO,GAAG,CAAC,IAAY,EAAU,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,IAAI,CAAC,CAAC;AAExE;;;GAGG;AACH,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,IAAY,EAAE,MAAe,EAAgB,EAAE,CACzE,MAAM,CAAC,IAAI,CAAC;KACT,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,IAAI,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC;KACrE,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;AAEjD;;;GAGG;AACH,MAAM,GAAG,GAAG,6BAA6B,CAAC;AAC1C,MAAM,SAAS,GAAG,IAAI,MAAM,CAAC,OAAO,GAAG,KAAK,EAAE,GAAG,CAAC,CAAC;AACnD,yEAAyE;AACzE,MAAM,UAAU,GAAG,IAAI,MAAM,CAAC,MAAM,GAAG,YAAY,EAAE,GAAG,CAAC,CAAC;AAE1D,6FAA6F;AAC7F,MAAM,CAAC,MAAM,mBAAmB,GAAG,CAAC,IAAY,EAAE,KAAa,EAAsB,EAAE,CACrF,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,GAAG,EAAE,CAAC,EAAE,KAAK,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,KAAK,CAAC,CAAC;AAE/E,MAAM,WAAW,GAAG,iBAAiB,CAAC;AACtC;;;GAGG;AACH,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,CAAC,WAAW,EAAE,UAAU,EAAE,OAAO,EAAE,QAAQ,CAAC,CAAC,CAAC;AACnE,wEAAwE;AACxE,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC;AACpD,MAAM,SAAS,GAAG,CAAC,CAAC;AAEpB,qHAAqH;AACrH,MAAM,UAAU,GAAG,CAAC,IAAY,EAAY,EAAE;IAC5C,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAC/B,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,MAAM,CAAC,MAAM,IAAI,KAAK,CAAC,MAAM,GAAG,SAAS,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QAClF,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;QAClC,MAAM,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;QAC1C,MAAM,KAAK,GAAG,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QACxH,IAAI,IAAI,KAAK,SAAS,IAAI,CAAC,KAAK;YAAE,MAAM;QACxC,KAAK,CAAC,IAAI,CAAC,IAAI,IAAI,KAAK,CAAC,CAAC;QAC1B,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM;YAAE,MAAM;IAC9D,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,kBAAkB,GAAG,CAAC,IAAY,EAAE,GAAW,EAAsB,EAAE;IAClF,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,GAAG,GAAG,GAAG,CAAC,CAAC;IACxC,MAAM,SAAS,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IAChC,MAAM,MAAM,GAAG,SAAS,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC,CAAC;IAChF,MAAM,EAAE,GAAG,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;IACvC,IAAI,EAAE,KAAK,IAAI;QAAE,OAAO,SAAS,CAAC;IAClC,MAAM,GAAG,GAAG,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,KAAK,CAAC,CAAC;IAC/E,IAAI,GAAG,KAAK,SAAS;QAAE,OAAO,GAAG,CAAC;IAClC,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC;IAC5D,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IACzC,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC7B,OAAO,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;AACjE,CAAC,CAAC"}
1
+ {"version":3,"file":"citation.js","sourceRoot":"","sources":["../src/citation.ts"],"names":[],"mappings":"AAAA,2GAA2G;AAE3G,MAAM,EAAE,GAAG,gCAAgC,CAAC;AAC5C,uIAAuI;AACvI,MAAM,KAAK,GAAG,qBAAqB,CAAC;AACpC,MAAM,SAAS,GAAG,2BAA2B,CAAC;AAC9C;;;GAGG;AACH,MAAM,cAAc,GAAG,CAAC,oDAAoD,EAAE,0BAA0B,EAAE,kBAAkB,CAAC,CAAC;AAE9H,MAAM,YAAY,GAAG,CAAC,IAAY,EAAsB,EAAE,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC;AAEzJ;;;GAGG;AACH,MAAM,UAAU,GAAG,2CAA2C,CAAC;AAI/D,sGAAsG;AACtG,MAAM,MAAM,GAAG,CAAC,IAAY,EAA6C,EAAE;IACzE,MAAM,KAAK,GAAqC,EAAE,CAAC;IACnD,IAAI,GAAG,GAAG,CAAC,CAAC;IACZ,SAAS,CAAC;QACR,MAAM,SAAS,GAAG,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;QACvD,MAAM,MAAM,GAAG,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC;QACtG,IAAI,SAAS,KAAK,SAAS,IAAI,MAAM,KAAK,SAAS;YAAE,OAAO,KAAK,CAAC;QAClE,MAAM,KAAK,GAAG,GAAG,GAAG,SAAS,CAAC,MAAM,CAAC;QACrC,GAAG,GAAG,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC;QAC5B,KAAK,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC,CAAC;IAC3C,CAAC;AACH,CAAC,CAAC;AAEF,6GAA6G;AAC7G,MAAM,OAAO,GAAG,CAAC,IAAY,EAAU,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,IAAI,CAAC,CAAC;AAExE;;;GAGG;AACH,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,IAAY,EAAE,MAAe,EAAgB,EAAE,CACzE,MAAM,CAAC,IAAI,CAAC;KACT,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,IAAI,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC;KACrE,GAAG,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;AAEjD;;;GAGG;AACH,MAAM,GAAG,GAAG,6BAA6B,CAAC;AAC1C;;;GAGG;AACH,MAAM,cAAc,GAAG,qDAAqD,CAAC;AAC7E,MAAM,QAAQ,GAAG,CAAC,GAAW,EAAU,EAAE,CAAC,IAAI,MAAM,CAAC,OAAO,GAAG,KAAK,EAAE,GAAG,CAAC,CAAC;AAC3E,yEAAyE;AACzE,MAAM,SAAS,GAAG,CAAC,GAAW,EAAU,EAAE,CAAC,IAAI,MAAM,CAAC,MAAM,GAAG,YAAY,EAAE,GAAG,CAAC,CAAC;AAClF,MAAM,SAAS,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC;AAChC,MAAM,UAAU,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC;AAClC,MAAM,gBAAgB,GAAG,QAAQ,CAAC,cAAc,CAAC,CAAC;AAClD,MAAM,iBAAiB,GAAG,SAAS,CAAC,cAAc,CAAC,CAAC;AACpD,MAAM,SAAS,GAAG,EAAE,CAAC;AAErB,MAAM,WAAW,GAAG,CAAC,OAAe,EAAE,IAAY,EAAE,KAAa,EAAsB,EAAE,CACvF,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,GAAG,SAAS,CAAC,EAAE,KAAK,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,KAAK,CAAC,CAAC;AAEnF,6FAA6F;AAC7F,MAAM,CAAC,MAAM,mBAAmB,GAAG,CAAC,IAAY,EAAE,KAAa,EAAsB,EAAE,CAAC,WAAW,CAAC,UAAU,EAAE,IAAI,EAAE,KAAK,CAAC,CAAC;AAE7H,MAAM,WAAW,GAAG,iBAAiB,CAAC;AACtC;;;GAGG;AACH,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,CAAC,WAAW,EAAE,UAAU,EAAE,OAAO,EAAE,QAAQ,CAAC,CAAC,CAAC;AACnE,wEAAwE;AACxE,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC;AACpD,MAAM,SAAS,GAAG,CAAC,CAAC;AAEpB,qHAAqH;AACrH,MAAM,UAAU,GAAG,CAAC,IAAY,EAAY,EAAE;IAC5C,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAC/B,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,MAAM,CAAC,MAAM,IAAI,KAAK,CAAC,MAAM,GAAG,SAAS,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QAClF,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;QAClC,MAAM,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;QAC1C,MAAM,KAAK,GAAG,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QACxH,IAAI,IAAI,KAAK,SAAS,IAAI,CAAC,KAAK;YAAE,MAAM;QACxC,KAAK,CAAC,IAAI,CAAC,IAAI,IAAI,KAAK,CAAC,CAAC;QAC1B,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM;YAAE,MAAM;IAC9D,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC,CAAC;AAEF,+FAA+F;AAC/F,MAAM,OAAO,GAAG,CAAC,IAAY,EAAE,GAAW,EAAsB,EAAE;IAChE,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,GAAG,GAAG,GAAG,CAAC,CAAC;IACxC,MAAM,SAAS,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IAChC,MAAM,MAAM,GAAG,SAAS,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC,CAAC;IAChF,MAAM,EAAE,GAAG,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;IACvC,OAAO,EAAE,KAAK,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;AACrE,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,kBAAkB,GAAG,CAAC,IAAY,EAAE,GAAW,EAAsB,EAAE;IAClF,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;IACjC,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC1C,MAAM,GAAG,GAAG,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,MAAM,EAAE,CAAC,KAAK,CAAC,CAAC;IACnD,IAAI,GAAG,KAAK,SAAS;QAAE,OAAO,GAAG,CAAC;IAClC,MAAM,KAAK,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC;IAChC,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IACzC,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC7B,OAAO,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;AACjE,CAAC,CAAC;AAEF,oIAAoI;AACpI,MAAM,CAAC,MAAM,mBAAmB,GAAG,CAAC,IAAY,EAAE,KAAa,EAAE,GAAW,EAAsB,EAAE;IAClG,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;IACjC,MAAM,SAAS,GAAG,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,gBAAgB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,MAAM,EAAE,CAAC,KAAK,CAAC,CAAC;IAClG,OAAO,SAAS,IAAI,WAAW,CAAC,iBAAiB,EAAE,IAAI,EAAE,KAAK,CAAC,CAAC;AAClE,CAAC,CAAC"}
@@ -0,0 +1,6 @@
1
+ import type { Span } from "chaffjs/plugin";
2
+ /** 文頭の閉じ引用符の並びの長さ。後ろに字が続けば開き引用符かアポストロフィなので 0。 */
3
+ export declare const closingRunLength: (sentence: string) => number;
4
+ /** 文の span の並び(text 上の位置)で、文頭に取り残された閉じ引用符を前の文へ戻したもの。 */
5
+ export declare const reattachClosingQuotes: (text: string, spans: readonly Span[]) => Span[];
6
+ //# sourceMappingURL=closing-quote.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"closing-quote.d.ts","sourceRoot":"","sources":["../src/closing-quote.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,IAAI,EAAE,MAAM,gBAAgB,CAAC;AAmB3C,iDAAiD;AACjD,eAAO,MAAM,gBAAgB,GAAI,UAAU,MAAM,KAAG,MAGnD,CAAC;AAoBF,wDAAwD;AACxD,eAAO,MAAM,qBAAqB,GAAI,MAAM,MAAM,EAAE,OAAO,SAAS,IAAI,EAAE,KAAG,IAAI,EAOzE,CAAC"}
@@ -0,0 +1,50 @@
1
+ import { NEXT_SENTENCE_HEAD } from "./quoted-stop.js";
2
+ /**
3
+ * sentence-splitter は疑問符・感嘆符の後の曲がった閉じ引用符(“Is it done?” Nobody …)と直線の一重引用符の手前で文を切る。
4
+ * 次の文が「”」で始まり、前の文は引用符が開いたまま終わる。
5
+ *
6
+ * 前の文が句点で終わり、次の文がそこから間を置かずに閉じ引用符(閉じ括弧も)で始まるとき、それを前の文の末尾へ戻す。
7
+ * 戻した後に空白と大文字が続けば文は二つのまま、何も続かなければ前の文だけ、
8
+ * 小文字・数字・句読点が続けば引用は文の途中にあるので一つの文につなぐ(直線の二重引用符で分割器がそうするのと同じ)。
9
+ * 開き引用符(“ ‘)は閉じの分類に入らない。語頭のアポストロフィ(’Tis ’90s)は後ろに字が続くので閉じと読まない。
10
+ */
11
+ const CLOSING_RUN = /^[\p{Pf}\p{Pe}"']+/u;
12
+ const WORD_CHARACTER = /^[\p{L}\p{N}]/u;
13
+ const STOP_AT_END = /[.?!]$/u;
14
+ // 語頭のアポストロフィの後の大文字(’Tis)も次の文の頭。
15
+ const NEXT_SENTENCE = new RegExp(String.raw `^(?:${NEXT_SENTENCE_HEAD}|\s+’\p{Lu})`, "u");
16
+ const LEADING_SPACE = /^\s*/u;
17
+ /** 文頭の閉じ引用符の並びの長さ。後ろに字が続けば開き引用符かアポストロフィなので 0。 */
18
+ export const closingRunLength = (sentence) => {
19
+ const run = CLOSING_RUN.exec(sentence)?.[0].length ?? 0;
20
+ return run > 0 && !WORD_CHARACTER.test(sentence.slice(run)) ? run : 0;
21
+ };
22
+ const follows = (text, previous, span) => previous.end === span.start && STOP_AT_END.test(text.slice(previous.start, previous.end));
23
+ /** previous と span を置き換える文。閉じ引用符が戻らなければ undefined。 */
24
+ const reattached = (text, previous, span) => {
25
+ const run = follows(text, previous, span) ? closingRunLength(text.slice(span.start, span.end)) : 0;
26
+ if (run === 0)
27
+ return undefined;
28
+ const quoteEnd = span.start + run;
29
+ const rest = text.slice(quoteEnd, span.end);
30
+ if (rest.trim() === "")
31
+ return [{ start: previous.start, end: quoteEnd }];
32
+ if (!NEXT_SENTENCE.test(rest))
33
+ return [{ start: previous.start, end: span.end }];
34
+ const next = quoteEnd + (LEADING_SPACE.exec(rest)?.[0].length ?? 0);
35
+ return [
36
+ { start: previous.start, end: quoteEnd },
37
+ { start: next, end: span.end },
38
+ ];
39
+ };
40
+ /** 文の span の並び(text 上の位置)で、文頭に取り残された閉じ引用符を前の文へ戻したもの。 */
41
+ export const reattachClosingQuotes = (text, spans) => spans.reduce((sentences, span) => {
42
+ const previous = sentences.at(-1);
43
+ const replaced = previous === undefined ? undefined : reattached(text, previous, span);
44
+ if (replaced === undefined)
45
+ sentences.push(span);
46
+ else
47
+ sentences.splice(-1, 1, ...replaced);
48
+ return sentences;
49
+ }, []);
50
+ //# sourceMappingURL=closing-quote.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"closing-quote.js","sourceRoot":"","sources":["../src/closing-quote.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,kBAAkB,EAAE,MAAM,kBAAkB,CAAC;AAEtD;;;;;;;;GAQG;AACH,MAAM,WAAW,GAAG,qBAAqB,CAAC;AAC1C,MAAM,cAAc,GAAG,gBAAgB,CAAC;AACxC,MAAM,WAAW,GAAG,SAAS,CAAC;AAC9B,gCAAgC;AAChC,MAAM,aAAa,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,OAAO,kBAAkB,cAAc,EAAE,GAAG,CAAC,CAAC;AACzF,MAAM,aAAa,GAAG,OAAO,CAAC;AAE9B,iDAAiD;AACjD,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,QAAgB,EAAU,EAAE;IAC3D,MAAM,GAAG,GAAG,WAAW,CAAC,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC;IACxD,OAAO,GAAG,GAAG,CAAC,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;AACxE,CAAC,CAAC;AAEF,MAAM,OAAO,GAAG,CAAC,IAAY,EAAE,QAAc,EAAE,IAAU,EAAW,EAAE,CACpE,QAAQ,CAAC,GAAG,KAAK,IAAI,CAAC,KAAK,IAAI,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,KAAK,EAAE,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC;AAE5F,sDAAsD;AACtD,MAAM,UAAU,GAAG,CAAC,IAAY,EAAE,QAAc,EAAE,IAAU,EAAsB,EAAE;IAClF,MAAM,GAAG,GAAG,OAAO,CAAC,IAAI,EAAE,QAAQ,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,gBAAgB,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACnG,IAAI,GAAG,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IAChC,MAAM,QAAQ,GAAG,IAAI,CAAC,KAAK,GAAG,GAAG,CAAC;IAClC,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC;IAC5C,IAAI,IAAI,CAAC,IAAI,EAAE,KAAK,EAAE;QAAE,OAAO,CAAC,EAAE,KAAK,EAAE,QAAQ,CAAC,KAAK,EAAE,GAAG,EAAE,QAAQ,EAAE,CAAC,CAAC;IAC1E,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,CAAC,EAAE,KAAK,EAAE,QAAQ,CAAC,KAAK,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,CAAC,CAAC;IACjF,MAAM,IAAI,GAAG,QAAQ,GAAG,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC,CAAC;IACpE,OAAO;QACL,EAAE,KAAK,EAAE,QAAQ,CAAC,KAAK,EAAE,GAAG,EAAE,QAAQ,EAAE;QACxC,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE;KAC/B,CAAC;AACJ,CAAC,CAAC;AAEF,wDAAwD;AACxD,MAAM,CAAC,MAAM,qBAAqB,GAAG,CAAC,IAAY,EAAE,KAAsB,EAAU,EAAE,CACpF,KAAK,CAAC,MAAM,CAAS,CAAC,SAAS,EAAE,IAAI,EAAE,EAAE;IACvC,MAAM,QAAQ,GAAG,SAAS,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAClC,MAAM,QAAQ,GAAG,QAAQ,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,UAAU,CAAC,IAAI,EAAE,QAAQ,EAAE,IAAI,CAAC,CAAC;IACvF,IAAI,QAAQ,KAAK,SAAS;QAAE,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;;QAC5C,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,GAAG,QAAQ,CAAC,CAAC;IAC1C,OAAO,SAAS,CAAC;AACnB,CAAC,EAAE,EAAE,CAAC,CAAC"}
package/dist/index.d.ts CHANGED
@@ -1,7 +1,7 @@
1
1
  import type { LanguageAdapter } from "chaffjs/plugin";
2
2
  /**
3
3
  * 英語は sentence-splitter の既定にほぼ任せる。"Dr." "e.g." "U.S." "$3.50" を
4
- * いずれも文末と誤認しない。前処理は、行の途中の番号を箇条書きと読ませないことだけ。spec §7.2。
4
+ * いずれも文末と誤認しない。前処理は行の途中の番号を箇条書きと読ませること、後処理は閉じ引用符の内側で閉じた文を切ることと、文頭に取り残された閉じ引用符を前の文へ戻すこと。spec §7.2。
5
5
  */
6
6
  export declare const adapter: LanguageAdapter;
7
7
  //# sourceMappingURL=index.d.ts.map
@@ -1 +1 @@
1
- {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAKA,OAAO,KAAK,EAAgB,eAAe,EAA0B,MAAM,gBAAgB,CAAC;AAoB5F;;;GAGG;AACH,eAAO,MAAM,OAAO,EAAE,eA4BrB,CAAC"}
1
+ {"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAOA,OAAO,KAAK,EAAgB,eAAe,EAA0B,MAAM,gBAAgB,CAAC;AAoB5F;;;GAGG;AACH,eAAO,MAAM,OAAO,EAAE,eA6BrB,CAAC"}
package/dist/index.js CHANGED
@@ -1,6 +1,8 @@
1
1
  import { loadLexicons } from "./lexicons.js";
2
2
  import { unmarkNumberStops } from "./number-stop.js";
3
3
  import { sentenceSpans } from "./sentence-split.js";
4
+ import { splitAtQuotedStops } from "./quoted-stop.js";
5
+ import { reattachClosingQuotes } from "./closing-quote.js";
4
6
  import { structure } from "./structure.js";
5
7
  import { isReady, prepare, tokenize } from "./pos.js";
6
8
  // chaff からは型だけを取る。実行時の値依存を作らない。アダプタは単体で動く。
@@ -21,7 +23,7 @@ const withTokens = (sentence) => {
21
23
  };
22
24
  /**
23
25
  * 英語は sentence-splitter の既定にほぼ任せる。"Dr." "e.g." "U.S." "$3.50" を
24
- * いずれも文末と誤認しない。前処理は、行の途中の番号を箇条書きと読ませないことだけ。spec §7.2。
26
+ * いずれも文末と誤認しない。前処理は行の途中の番号を箇条書きと読ませること、後処理は閉じ引用符の内側で閉じた文を切ることと、文頭に取り残された閉じ引用符を前の文へ戻すこと。spec §7.2。
25
27
  */
26
28
  export const adapter = {
27
29
  kind: "language",
@@ -50,7 +52,8 @@ export const adapter = {
50
52
  lexicons: loadLexicons(),
51
53
  structure,
52
54
  segment: (text) => {
53
- const sentences = sentenceSpans(unmarkNumberStops(text)).map((span) => ({ span, text: text.slice(span.start, span.end) }));
55
+ const quotedStops = sentenceSpans(unmarkNumberStops(text)).flatMap((span) => splitAtQuotedStops(text, span));
56
+ const sentences = reattachClosingQuotes(text, quotedStops).map((span) => ({ span, text: text.slice(span.start, span.end) }));
54
57
  return { sentences: isReady() ? sentences.map(withTokens) : sentences };
55
58
  },
56
59
  };
package/dist/index.js.map CHANGED
@@ -1 +1 @@
1
- {"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAC7C,OAAO,EAAE,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AACrD,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,SAAS,EAAE,MAAM,gBAAgB,CAAC;AAC3C,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,MAAM,UAAU,CAAC;AAGtD,2CAA2C;AAE3C,MAAM,YAAY,GAAG,UAAU,CAAC;AAChC,MAAM,SAAS,GAAG,MAAM,CAAC;AAEzB;;;GAGG;AACH,MAAM,UAAU,GAAG,CAAC,QAAkB,EAAY,EAAE;IAClD,MAAM,MAAM,GAAG,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IACvC,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,QAAQ,CAAC;IAC1C,OAAO;QACL,GAAG,QAAQ;QACX,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,KAAK,EAAE,IAAI,EAAE,EAAE,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,GAAG,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;KAClJ,CAAC;AACJ,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,OAAO,GAAoB;IACtC,IAAI,EAAE,UAAU;IAChB,EAAE,EAAE,IAAI;IACR,UAAU,EAAE,CAAC;IACb,YAAY,EAAE;QACZ,aAAa,EAAE,IAAI;QACnB,SAAS,EAAE,IAAI;QACf,6CAA6C;QAC7C,GAAG,EAAE,IAAI;QACT,KAAK,EAAE,IAAI;QACX,UAAU,EAAE,MAAM;KACnB;IACD,OAAO,EAAE,CAAC,IAAkB,EAAiB,EAAE;QAC7C,0CAA0C;QAC1C,IAAI,IAAI,CAAC,GAAG;YAAE,OAAO,EAAE,CAAC;QACxB,OAAO,OAAO,CAAC,OAAO,EAAE,CAAC;IAC3B,CAAC;IACD,MAAM,EAAE,CAAC,MAAc,EAAU,EAAE;QACjC,MAAM,KAAK,GAAG,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC;QACrD,IAAI,KAAK,KAAK,CAAC;YAAE,OAAO,CAAC,CAAC;QAC1B,OAAO,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC,CAAC,MAAM,GAAG,KAAK,CAAC;IAC3D,CAAC;IACD,QAAQ,EAAE,YAAY,EAAE;IACxB,SAAS;IACT,OAAO,EAAE,CAAC,IAAY,EAAgB,EAAE;QACtC,MAAM,SAAS,GAAe,aAAa,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC;QACvI,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,SAAS,EAAE,CAAC;IAC1E,CAAC;CACF,CAAC"}
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAC7C,OAAO,EAAE,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AACrD,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,kBAAkB,EAAE,MAAM,kBAAkB,CAAC;AACtD,OAAO,EAAE,qBAAqB,EAAE,MAAM,oBAAoB,CAAC;AAC3D,OAAO,EAAE,SAAS,EAAE,MAAM,gBAAgB,CAAC;AAC3C,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,MAAM,UAAU,CAAC;AAGtD,2CAA2C;AAE3C,MAAM,YAAY,GAAG,UAAU,CAAC;AAChC,MAAM,SAAS,GAAG,MAAM,CAAC;AAEzB;;;GAGG;AACH,MAAM,UAAU,GAAG,CAAC,QAAkB,EAAY,EAAE;IAClD,MAAM,MAAM,GAAG,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IACvC,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,QAAQ,CAAC;IAC1C,OAAO;QACL,GAAG,QAAQ;QACX,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,KAAK,EAAE,IAAI,EAAE,EAAE,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,GAAG,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;KAClJ,CAAC;AACJ,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,OAAO,GAAoB;IACtC,IAAI,EAAE,UAAU;IAChB,EAAE,EAAE,IAAI;IACR,UAAU,EAAE,CAAC;IACb,YAAY,EAAE;QACZ,aAAa,EAAE,IAAI;QACnB,SAAS,EAAE,IAAI;QACf,6CAA6C;QAC7C,GAAG,EAAE,IAAI;QACT,KAAK,EAAE,IAAI;QACX,UAAU,EAAE,MAAM;KACnB;IACD,OAAO,EAAE,CAAC,IAAkB,EAAiB,EAAE;QAC7C,0CAA0C;QAC1C,IAAI,IAAI,CAAC,GAAG;YAAE,OAAO,EAAE,CAAC;QACxB,OAAO,OAAO,CAAC,OAAO,EAAE,CAAC;IAC3B,CAAC;IACD,MAAM,EAAE,CAAC,MAAc,EAAU,EAAE;QACjC,MAAM,KAAK,GAAG,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC;QACrD,IAAI,KAAK,KAAK,CAAC;YAAE,OAAO,CAAC,CAAC;QAC1B,OAAO,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC,CAAC,MAAM,GAAG,KAAK,CAAC;IAC3D,CAAC;IACD,QAAQ,EAAE,YAAY,EAAE;IACxB,SAAS;IACT,OAAO,EAAE,CAAC,IAAY,EAAgB,EAAE;QACtC,MAAM,WAAW,GAAG,aAAa,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,kBAAkB,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;QAC7G,MAAM,SAAS,GAAe,qBAAqB,CAAC,IAAI,EAAE,WAAW,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC;QACzI,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,SAAS,EAAE,CAAC;IAC1E,CAAC;CACF,CAAC"}
@@ -0,0 +1,30 @@
1
+ import type { NumberedLine, NumberingContext } from "chaffjs/plugin";
2
+ /**
3
+ * An amendment inserts a subsection between two others and numbers it "(A1)" or "(2A)". It is a subsection, written
4
+ * outside the sequence: it has no ordinal, so "(1)" after "(A1)" is still the first.
5
+ */
6
+ export declare const INSERTED = "\\d{1,3}[A-Z]{1,2}|[A-Z]{1,2}\\d{1,3}";
7
+ export declare const IS_INSERTED: RegExp;
8
+ export type Style = "letter" | "roman" | "digit" | "capital" | "capital-roman";
9
+ /** 開いている項目の書き方。"(ii)" はローマ数字、"(b)" は英字、"(B)" は大文字の英字。 */
10
+ export declare const styleOfOpen: (open: NumberedLine) => Style | undefined;
11
+ /** "(b)" と "(B)" は 2 番目、"(ii)" と "(II)" も 2 番目。二文字以上の英字("(aa)")は並びが決まらないので付けない。 */
12
+ export declare const ordinalOf: (raw: string, style: Style) => number | undefined;
13
+ /**
14
+ * "(i)" is a roman numeral right under "(a)" or "(1)", or when a roman list is already open; the letter i otherwise.
15
+ * A capital "(A)" is an item only where a US regulation puts it, under a roman item; undefined elsewhere.
16
+ */
17
+ export declare const styleOf: (raw: string, context: NumberingContext) => Style | undefined;
18
+ /**
19
+ * A sibling has the depth of the open item written the same way. A new way of numbering goes one deeper than whatever
20
+ * is open, and so does a first item right under a capital, or a first capital right under a roman item.
21
+ */
22
+ export declare const depthFor: (style: Style, context: NumberingContext, ordinal?: number) => number;
23
+ /**
24
+ * 参照の "(A)" や "(I)" は、ローマ数字の "(iii)" の次に書かれたときだけ番地の続き。木も、大文字はローマ数字の下でしか読まない。
25
+ * "section 5(A)" の "(A)" は番地に入れず、これまでどおり "section 5" を引く。"(h)(i)" や番号のすぐ後の "(i)" は、木と同じく英字と読む。
26
+ */
27
+ export declare const continuesAddress: (part: string, before: readonly string[]) => boolean;
28
+ /** 番号だけの行 "(5)" は、開いている "(4)" の次のときだけ項目。本文は次の行から始まる。 */
29
+ export declare const continuesOpen: (style: Style, ordinal: number | undefined, context: NumberingContext) => boolean;
30
+ //# sourceMappingURL=item-style.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"item-style.d.ts","sourceRoot":"","sources":["../src/item-style.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,YAAY,EAAE,gBAAgB,EAAE,MAAM,gBAAgB,CAAC;AAGrE;;;GAGG;AACH,eAAO,MAAM,QAAQ,0CAA0C,CAAC;AAChE,eAAO,MAAM,WAAW,QAAuC,CAAC;AAOhE,MAAM,MAAM,KAAK,GAAG,QAAQ,GAAG,OAAO,GAAG,OAAO,GAAG,SAAS,GAAG,eAAe,CAAC;AAW/E,yDAAyD;AACzD,eAAO,MAAM,WAAW,GAAI,MAAM,YAAY,KAAG,KAAK,GAAG,SAMxD,CAAC;AAIF,mFAAmF;AACnF,eAAO,MAAM,SAAS,GAAI,KAAK,MAAM,EAAE,OAAO,KAAK,KAAG,MAAM,GAAG,SAI9D,CAAC;AAiCF;;;GAGG;AACH,eAAO,MAAM,OAAO,GAAI,KAAK,MAAM,EAAE,SAAS,gBAAgB,KAAG,KAAK,GAAG,SAOxE,CAAC;AAiBF;;;GAGG;AACH,eAAO,MAAM,QAAQ,GAAI,OAAO,KAAK,EAAE,SAAS,gBAAgB,EAAE,UAAU,MAAM,KAAG,MAKpF,CAAC;AAEF;;;GAGG;AACH,eAAO,MAAM,gBAAgB,GAAI,MAAM,MAAM,EAAE,QAAQ,SAAS,MAAM,EAAE,KAAG,OAK1E,CAAC;AAEF,wDAAwD;AACxD,eAAO,MAAM,aAAa,GAAI,OAAO,KAAK,EAAE,SAAS,MAAM,GAAG,SAAS,EAAE,SAAS,gBAAgB,KAAG,OAGpG,CAAC"}
@@ -0,0 +1,124 @@
1
+ import { parseRoman } from "./roman.js";
2
+ /**
3
+ * An amendment inserts a subsection between two others and numbers it "(A1)" or "(2A)". It is a subsection, written
4
+ * outside the sequence: it has no ordinal, so "(1)" after "(A1)" is still the first.
5
+ */
6
+ export const INSERTED = "\\d{1,3}[A-Z]{1,2}|[A-Z]{1,2}\\d{1,3}";
7
+ export const IS_INSERTED = new RegExp(`^(?:${INSERTED})$`, "u");
8
+ const MULTI_ROMAN = /^(?:ii|iii|iv|vi|vii|viii|ix)$/u;
9
+ const AMBIGUOUS = /^[ivx]$/u;
10
+ const DIGITS = /^\d+$/u;
11
+ const CAPITALS = /^[A-Z]+$/u;
12
+ const ONE_CAPITAL = /^[A-Z]$/u;
13
+ /** 大文字で書いた同じ並び。"(B)" は英字の、"(II)" はローマ数字の大文字。 */
14
+ const CAPITAL_OF = { letter: "capital", roman: "capital-roman" };
15
+ /** 小文字にした番号を形で読む。一文字の "i" は、開いたときに付けた並びの位置で見分ける。ローマ数字なら 1、英字なら 9。 */
16
+ const shapeOf = (lower, ordinal) => {
17
+ if (MULTI_ROMAN.test(lower))
18
+ return "roman";
19
+ return AMBIGUOUS.test(lower) && ordinal === parseRoman(lower) ? "roman" : "letter";
20
+ };
21
+ /** 開いている項目の書き方。"(ii)" はローマ数字、"(b)" は英字、"(B)" は大文字の英字。 */
22
+ export const styleOfOpen = (open) => {
23
+ const inner = /^\((?<n>[A-Za-z0-9]{1,5})\)$/u.exec(open.label)?.groups?.["n"];
24
+ if (inner === undefined)
25
+ return undefined;
26
+ if (DIGITS.test(inner) || IS_INSERTED.test(inner))
27
+ return "digit";
28
+ const shape = shapeOf(inner.toLowerCase(), open.ordinal);
29
+ return CAPITALS.test(inner) ? CAPITAL_OF[shape] : shape;
30
+ };
31
+ const LETTER_BEFORE_A = "a".charCodeAt(0) - 1;
32
+ /** "(b)" と "(B)" は 2 番目、"(ii)" と "(II)" も 2 番目。二文字以上の英字("(aa)")は並びが決まらないので付けない。 */
33
+ export const ordinalOf = (raw, style) => {
34
+ if (style === "digit")
35
+ return IS_INSERTED.test(raw) ? undefined : Number(raw);
36
+ if (style === "roman" || style === "capital-roman")
37
+ return parseRoman(raw);
38
+ return raw.length === 1 ? raw.toLowerCase().charCodeAt(0) - LETTER_BEFORE_A : undefined;
39
+ };
40
+ const styles = (context) => context.open.map(styleOfOpen);
41
+ /** "(h)" の次の "(i)"、"(H)" の次の "(I)" は英字。同じ書き方で開いている英字の次の文字なら、ローマ数字とは読まない。 */
42
+ const follows = (raw, style, context) => context.open.some((open, index) => styles(context)[index] === style && open.number.charCodeAt(0) + 1 === raw.charCodeAt(0));
43
+ /**
44
+ * 英字の並びは (a) から始まるので、"(a)" や "(1)" のすぐ下に来た "(i)" は一段深いローマ数字。
45
+ * 米国の規則は (a)(1)(i) の順に下る。見出しのすぐ下の "(i)" は、どちらとも決まらないので英字。
46
+ */
47
+ const OPENS_ROMAN = new Set(["letter", "digit"]);
48
+ /** "(I)" は米国の法典で (i) の一段下の大文字ローマ数字。"(H)" の次なら英字。 */
49
+ const capitalRoman = (raw, open, context) => {
50
+ const lower = raw.toLowerCase();
51
+ if (MULTI_ROMAN.test(lower))
52
+ return open.includes("capital-roman");
53
+ if (!AMBIGUOUS.test(lower) || follows(raw, "capital", context))
54
+ return false;
55
+ return open.includes("capital-roman") || (raw === "I" && open.at(-1) === "roman");
56
+ };
57
+ /**
58
+ * 米国の規則は (a)(1)(i)(A) と下る。大文字の "(A)" は、開いているローマ数字のすぐ下で始まるか、開いている大文字の続きのときだけ項目。
59
+ * 英国の法令や契約書が一段目に使う "(A)" は、何の下とも決まらないので本文のまま読む。
60
+ */
61
+ const capitalStyleOf = (raw, context) => {
62
+ const open = styles(context);
63
+ if (capitalRoman(raw, open, context))
64
+ return "capital-roman";
65
+ if (!ONE_CAPITAL.test(raw))
66
+ return undefined;
67
+ return open.includes("capital") || (raw === "A" && open.at(-1) === "roman") ? "capital" : undefined;
68
+ };
69
+ /**
70
+ * "(i)" is a roman numeral right under "(a)" or "(1)", or when a roman list is already open; the letter i otherwise.
71
+ * A capital "(A)" is an item only where a US regulation puts it, under a roman item; undefined elsewhere.
72
+ */
73
+ export const styleOf = (raw, context) => {
74
+ if (DIGITS.test(raw) || IS_INSERTED.test(raw))
75
+ return "digit";
76
+ if (CAPITALS.test(raw))
77
+ return capitalStyleOf(raw, context);
78
+ if (MULTI_ROMAN.test(raw))
79
+ return "roman";
80
+ if (!AMBIGUOUS.test(raw) || follows(raw, "letter", context))
81
+ return "letter";
82
+ const open = styles(context);
83
+ return open.includes("roman") || OPENS_ROMAN.has(open.at(-1)) ? "roman" : "letter";
84
+ };
85
+ /** The nearest open item written the same way: "(b)" closes "(i)" and sits beside "(a)". */
86
+ const siblingOf = (style, context) => {
87
+ const open = styles(context);
88
+ return context.open.findLast((_item, index) => open[index] === style);
89
+ };
90
+ const IS_CAPITAL = new Set(["capital", "capital-roman"]);
91
+ /**
92
+ * "(A)" の下の "(1)" や "(i)"、"(i)" の下の "(A)" は、上に開いている同じ書き方の項目の続きではなく、一段深い並びの始まり。
93
+ * 米国の規則は (a)(1)(i)(A) の下を、もう一度 (1) や (i) で数える。
94
+ */
95
+ const startsBelow = (style, ordinal, innermost) => ordinal === 1 && style !== innermost && (IS_CAPITAL.has(innermost) || (IS_CAPITAL.has(style) && innermost === "roman"));
96
+ /**
97
+ * A sibling has the depth of the open item written the same way. A new way of numbering goes one deeper than whatever
98
+ * is open, and so does a first item right under a capital, or a first capital right under a roman item.
99
+ */
100
+ export const depthFor = (style, context, ordinal) => {
101
+ const innermost = context.open.at(-1);
102
+ const deeper = (innermost?.depth ?? 0) + 1;
103
+ if (innermost !== undefined && startsBelow(style, ordinal, styleOfOpen(innermost)))
104
+ return deeper;
105
+ return siblingOf(style, context)?.depth ?? deeper;
106
+ };
107
+ /**
108
+ * 参照の "(A)" や "(I)" は、ローマ数字の "(iii)" の次に書かれたときだけ番地の続き。木も、大文字はローマ数字の下でしか読まない。
109
+ * "section 5(A)" の "(A)" は番地に入れず、これまでどおり "section 5" を引く。"(h)(i)" や番号のすぐ後の "(i)" は、木と同じく英字と読む。
110
+ */
111
+ export const continuesAddress = (part, before) => {
112
+ if (!CAPITALS.test(part))
113
+ return true;
114
+ const [previous, earlier] = [before.at(-1) ?? "", before.at(-2) ?? ""];
115
+ if (MULTI_ROMAN.test(previous))
116
+ return true;
117
+ return AMBIGUOUS.test(previous) && earlier !== "" && earlier.charCodeAt(0) + 1 !== previous.charCodeAt(0);
118
+ };
119
+ /** 番号だけの行 "(5)" は、開いている "(4)" の次のときだけ項目。本文は次の行から始まる。 */
120
+ export const continuesOpen = (style, ordinal, context) => {
121
+ const previous = siblingOf(style, context)?.ordinal;
122
+ return ordinal !== undefined && previous !== undefined && ordinal === previous + 1;
123
+ };
124
+ //# sourceMappingURL=item-style.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"item-style.js","sourceRoot":"","sources":["../src/item-style.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,UAAU,EAAE,MAAM,YAAY,CAAC;AAExC;;;GAGG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAG,uCAAuC,CAAC;AAChE,MAAM,CAAC,MAAM,WAAW,GAAG,IAAI,MAAM,CAAC,OAAO,QAAQ,IAAI,EAAE,GAAG,CAAC,CAAC;AAChE,MAAM,WAAW,GAAG,iCAAiC,CAAC;AACtD,MAAM,SAAS,GAAG,UAAU,CAAC;AAC7B,MAAM,MAAM,GAAG,QAAQ,CAAC;AACxB,MAAM,QAAQ,GAAG,WAAW,CAAC;AAC7B,MAAM,WAAW,GAAG,UAAU,CAAC;AAI/B,gDAAgD;AAChD,MAAM,UAAU,GAAgD,EAAE,MAAM,EAAE,SAAS,EAAE,KAAK,EAAE,eAAe,EAAE,CAAC;AAE9G,qEAAqE;AACrE,MAAM,OAAO,GAAG,CAAC,KAAa,EAAE,OAA2B,EAAsB,EAAE;IACjF,IAAI,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC;QAAE,OAAO,OAAO,CAAC;IAC5C,OAAO,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,OAAO,KAAK,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC;AACrF,CAAC,CAAC;AAEF,yDAAyD;AACzD,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,IAAkB,EAAqB,EAAE;IACnE,MAAM,KAAK,GAAG,+BAA+B,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC;IAC9E,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC1C,IAAI,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC;QAAE,OAAO,OAAO,CAAC;IAClE,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,WAAW,EAAE,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC;IACzD,OAAO,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;AAC1D,CAAC,CAAC;AAEF,MAAM,eAAe,GAAG,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;AAE9C,mFAAmF;AACnF,MAAM,CAAC,MAAM,SAAS,GAAG,CAAC,GAAW,EAAE,KAAY,EAAsB,EAAE;IACzE,IAAI,KAAK,KAAK,OAAO;QAAE,OAAO,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;IAC9E,IAAI,KAAK,KAAK,OAAO,IAAI,KAAK,KAAK,eAAe;QAAE,OAAO,UAAU,CAAC,GAAG,CAAC,CAAC;IAC3E,OAAO,GAAG,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,WAAW,EAAE,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,eAAe,CAAC,CAAC,CAAC,SAAS,CAAC;AAC1F,CAAC,CAAC;AAEF,MAAM,MAAM,GAAG,CAAC,OAAyB,EAAyB,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC;AAEnG,4EAA4E;AAC5E,MAAM,OAAO,GAAG,CAAC,GAAW,EAAE,KAA2B,EAAE,OAAyB,EAAW,EAAE,CAC/F,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,CAAC,KAAK,KAAK,IAAI,IAAI,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,KAAK,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;AAE9H;;;GAGG;AACH,MAAM,WAAW,GAAmC,IAAI,GAAG,CAAC,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC,CAAC;AAEjF,oDAAoD;AACpD,MAAM,YAAY,GAAG,CAAC,GAAW,EAAE,IAAoC,EAAE,OAAyB,EAAW,EAAE;IAC7G,MAAM,KAAK,GAAG,GAAG,CAAC,WAAW,EAAE,CAAC;IAChC,IAAI,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC,QAAQ,CAAC,eAAe,CAAC,CAAC;IACnE,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,OAAO,CAAC,GAAG,EAAE,SAAS,EAAE,OAAO,CAAC;QAAE,OAAO,KAAK,CAAC;IAC7E,OAAO,IAAI,CAAC,QAAQ,CAAC,eAAe,CAAC,IAAI,CAAC,GAAG,KAAK,GAAG,IAAI,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,OAAO,CAAC,CAAC;AACpF,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,cAAc,GAAG,CAAC,GAAW,EAAE,OAAyB,EAAqB,EAAE;IACnF,MAAM,IAAI,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC;IAC7B,IAAI,YAAY,CAAC,GAAG,EAAE,IAAI,EAAE,OAAO,CAAC;QAAE,OAAO,eAAe,CAAC;IAC7D,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,SAAS,CAAC;IAC7C,OAAO,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,KAAK,GAAG,IAAI,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,OAAO,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC;AACtG,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,OAAO,GAAG,CAAC,GAAW,EAAE,OAAyB,EAAqB,EAAE;IACnF,IAAI,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,OAAO,CAAC;IAC9D,IAAI,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,cAAc,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;IAC5D,IAAI,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,OAAO,CAAC;IAC1C,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,OAAO,CAAC,GAAG,EAAE,QAAQ,EAAE,OAAO,CAAC;QAAE,OAAO,QAAQ,CAAC;IAC7E,MAAM,IAAI,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC;IAC7B,OAAO,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,WAAW,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC;AACrF,CAAC,CAAC;AAEF,4FAA4F;AAC5F,MAAM,SAAS,GAAG,CAAC,KAAY,EAAE,OAAyB,EAA4B,EAAE;IACtF,MAAM,IAAI,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC;IAC7B,OAAO,OAAO,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,KAAK,CAAC,CAAC;AACxE,CAAC,CAAC;AAEF,MAAM,UAAU,GAAmC,IAAI,GAAG,CAAC,CAAC,SAAS,EAAE,eAAe,CAAC,CAAC,CAAC;AAEzF;;;GAGG;AACH,MAAM,WAAW,GAAG,CAAC,KAAY,EAAE,OAA2B,EAAE,SAA4B,EAAW,EAAE,CACvG,OAAO,KAAK,CAAC,IAAI,KAAK,KAAK,SAAS,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,SAAS,KAAK,OAAO,CAAC,CAAC,CAAC;AAE1H;;;GAGG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAG,CAAC,KAAY,EAAE,OAAyB,EAAE,OAAgB,EAAU,EAAE;IAC5F,MAAM,SAAS,GAAG,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IACtC,MAAM,MAAM,GAAG,CAAC,SAAS,EAAE,KAAK,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;IAC3C,IAAI,SAAS,KAAK,SAAS,IAAI,WAAW,CAAC,KAAK,EAAE,OAAO,EAAE,WAAW,CAAC,SAAS,CAAC,CAAC;QAAE,OAAO,MAAM,CAAC;IAClG,OAAO,SAAS,CAAC,KAAK,EAAE,OAAO,CAAC,EAAE,KAAK,IAAI,MAAM,CAAC;AACpD,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,IAAY,EAAE,MAAyB,EAAW,EAAE;IACnF,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,IAAI,CAAC;IACtC,MAAM,CAAC,QAAQ,EAAE,OAAO,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,EAAE,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;IACvE,IAAI,WAAW,CAAC,IAAI,CAAC,QAAQ,CAAC;QAAE,OAAO,IAAI,CAAC;IAC5C,OAAO,SAAS,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,OAAO,KAAK,EAAE,IAAI,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,KAAK,QAAQ,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC;AAC5G,CAAC,CAAC;AAEF,wDAAwD;AACxD,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,KAAY,EAAE,OAA2B,EAAE,OAAyB,EAAW,EAAE;IAC7G,MAAM,QAAQ,GAAG,SAAS,CAAC,KAAK,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC;IACpD,OAAO,OAAO,KAAK,SAAS,IAAI,QAAQ,KAAK,SAAS,IAAI,OAAO,KAAK,QAAQ,GAAG,CAAC,CAAC;AACrF,CAAC,CAAC"}
@@ -0,0 +1,6 @@
1
+ /**
2
+ * 空白の無いまま limit 文字を超える並びを、同じ長さの空白で覆う。語ではなく(base64、ハッシュ、区切りの無い記号の列)、
3
+ * 解析器(wink)はその長さの二乗で遅くなる。長さを変えないので、残りの語の位置はそのまま。
4
+ */
5
+ export declare const blankLongRuns: (text: string, limit: number) => string;
6
+ //# sourceMappingURL=long-runs.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"long-runs.d.ts","sourceRoot":"","sources":["../src/long-runs.ts"],"names":[],"mappings":"AAAA;;;GAGG;AACH,eAAO,MAAM,aAAa,GAAI,MAAM,MAAM,EAAE,OAAO,MAAM,KAAG,MACmC,CAAC"}
@@ -0,0 +1,6 @@
1
+ /**
2
+ * 空白の無いまま limit 文字を超える並びを、同じ長さの空白で覆う。語ではなく(base64、ハッシュ、区切りの無い記号の列)、
3
+ * 解析器(wink)はその長さの二乗で遅くなる。長さを変えないので、残りの語の位置はそのまま。
4
+ */
5
+ export const blankLongRuns = (text, limit) => text.replace(new RegExp(`\\S{${String(limit + 1)},}`, "gu"), (run) => " ".repeat(run.length));
6
+ //# sourceMappingURL=long-runs.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"long-runs.js","sourceRoot":"","sources":["../src/long-runs.ts"],"names":[],"mappings":"AAAA;;;GAGG;AACH,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,IAAY,EAAE,KAAa,EAAU,EAAE,CACnE,IAAI,CAAC,OAAO,CAAC,IAAI,MAAM,CAAC,OAAO,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,CAAC,EAAE,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC"}
package/dist/pos.d.ts.map CHANGED
@@ -1 +1 @@
1
- {"version":3,"file":"pos.d.ts","sourceRoot":"","sources":["../src/pos.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,KAAK,EAAE,MAAM,gBAAgB,CAAC;AA+E5C,eAAO,MAAM,IAAI,GAAI,KAAK,MAAM,KAAG,MAAiE,CAAC;AAyFrG,eAAO,MAAM,OAAO,QAAO,IAE1B,CAAC;AAEF,eAAO,MAAM,OAAO,QAAO,OAAoC,CAAC;AAwBhE,eAAO,MAAM,QAAQ,GAAI,MAAM,MAAM,KAAG,KAAK,EAAE,GAAG,SAIjD,CAAC"}
1
+ {"version":3,"file":"pos.d.ts","sourceRoot":"","sources":["../src/pos.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,KAAK,EAAE,MAAM,gBAAgB,CAAC;AAgF5C,eAAO,MAAM,IAAI,GAAI,KAAK,MAAM,KAAG,MAAiE,CAAC;AA4GrG,eAAO,MAAM,OAAO,QAAO,IAI1B,CAAC;AAEF,eAAO,MAAM,OAAO,QAAO,OAAoC,CAAC;AAoChE,eAAO,MAAM,QAAQ,GAAI,MAAM,MAAM,KAAG,KAAK,EAAE,GAAG,SAKjD,CAAC"}
package/dist/pos.js CHANGED
@@ -1,6 +1,7 @@
1
1
  import { createRequire } from "node:module";
2
2
  import { loadLexicons } from "./lexicons.js";
3
- import { properNounChecked } from "./proper-noun.js";
3
+ import { blankLongRuns } from "./long-runs.js";
4
+ import { lowercasedAt, properNounChecked, rereadAt, sentenceInitialCommonWord } from "./proper-noun.js";
4
5
  import { isStativeParticiple, stativeVocabulary } from "./stative-participle.js";
5
6
  const require = createRequire(import.meta.url);
6
7
  const isRecord = (value) => typeof value === "object" && value !== null;
@@ -126,18 +127,34 @@ const determinerFeatures = (entry) => {
126
127
  return OBJECT_OR_POSSESSIVE.has(entry.value.toLowerCase()) ? {} : { features: { Poss: "Yes" } };
127
128
  return entry.pos === "DT" && ARTICLES.has(entry.value.toLowerCase()) ? { features: { PronType: "Art" } } : {};
128
129
  };
130
+ /** 複数形の名詞は UD の Number=Plur。数の語がそれを数えていれば(five minutes)、one of のような言い回しではなく量。 */
131
+ const PLURAL_TAG = new Set(["NNS", "NNPS"]);
132
+ const nounOrDeterminerFeatures = (entry) => (PLURAL_TAG.has(entry.pos) ? { features: { Number: "Plur" } } : determinerFeatures(entry));
129
133
  /** 過去分詞は VerbForm=Part。Based on the review, のような分詞の導入句を、命令形の並び(fix the parser, ship it)と見分ける。 */
130
134
  const featuresOf = (tagged, at) => {
131
135
  const entry = tagged[at];
132
136
  if (entry === undefined)
133
137
  return {};
134
138
  if (entry.pos !== "VBN")
135
- return determinerFeatures(entry);
139
+ return nounOrDeterminerFeatures(entry);
136
140
  return { features: isPassive(tagged, at) ? { VerbForm: "Part", Voice: "Pass" } : { VerbForm: "Part" } };
137
141
  };
138
- const state = { ready: undefined };
142
+ const isTags = (value) => Array.isArray(value) && value.every((tag) => typeof tag === "string");
143
+ const buildVocabulary = () => {
144
+ const words = createRequire(require.resolve("wink-pos-tagger"))("wink-lexicon/src/lexicon.js");
145
+ if (!isRecord(words))
146
+ throw new Error("wink-lexicon の語彙が object ではありません");
147
+ return (word) => {
148
+ const tags = Object.hasOwn(words, word) ? words[word] : undefined;
149
+ return isTags(tags) ? tags : undefined;
150
+ };
151
+ };
152
+ const state = { ready: undefined, vocabulary: () => undefined };
139
153
  export const prepare = () => {
140
- state.ready ??= build();
154
+ if (state.ready !== undefined)
155
+ return;
156
+ state.ready = build();
157
+ state.vocabulary = buildVocabulary();
141
158
  };
142
159
  export const isReady = () => state.ready !== undefined;
143
160
  /**
@@ -158,10 +175,20 @@ const locate = (text, tagged) => tagged.reduce((acc, entry, at) => {
158
175
  };
159
176
  return { tokens: [...acc.tokens, token], cursor: end };
160
177
  }, { tokens: [], cursor: 0 }).tokens;
178
+ /** 英語の語はこれより長くならない。超える並びは語として読まない。 */
179
+ const RUN_LIMIT = 1000;
180
+ const tagged = (tagger, text) => toArray(callMethod(tagger, "tagSentence", [text])).filter(isTagged);
181
+ /** 文頭で大文字になっただけの普通の語を、小文字で書いたときの品詞に戻す。前後の語による判断も効くよう、文ごと解析し直す。 */
182
+ const withSentenceInitialCase = (tagger, text, entries) => {
183
+ const at = sentenceInitialCommonWord(entries, state.vocabulary);
184
+ const lowered = lowercasedAt(text, entries[at]?.value);
185
+ return lowered === undefined ? entries : rereadAt(entries, at, tagged(tagger, lowered));
186
+ };
161
187
  export const tokenize = (text) => {
162
188
  const tagger = state.ready;
163
189
  if (tagger === undefined)
164
190
  return undefined;
165
- return locate(text, toArray(callMethod(tagger, "tagSentence", [text])).filter(isTagged));
191
+ const words = blankLongRuns(text, RUN_LIMIT);
192
+ return locate(words, withSentenceInitialCase(tagger, words, tagged(tagger, words)));
166
193
  };
167
194
  //# sourceMappingURL=pos.js.map
package/dist/pos.js.map CHANGED
@@ -1 +1 @@
1
- {"version":3,"file":"pos.js","sourceRoot":"","sources":["../src/pos.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAE5C,OAAO,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAC7C,OAAO,EAAE,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AACrD,OAAO,EAAE,mBAAmB,EAAE,iBAAiB,EAAE,MAAM,yBAAyB,CAAC;AAEjF,MAAM,OAAO,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAO/C,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAoC,EAAE,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC;AAEnH,MAAM,UAAU,GAAG,CAAC,KAAc,EAAqD,EAAE,CAAC,OAAO,KAAK,KAAK,UAAU,CAAC;AAEtH,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAmB,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,OAAO,KAAK,CAAC,OAAO,CAAC,KAAK,QAAQ,IAAI,OAAO,KAAK,CAAC,KAAK,CAAC,KAAK,QAAQ,CAAC;AAEhJ,MAAM,OAAO,GAAG,CAAC,KAAc,EAAsB,EAAE,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;AAE5F,kEAAkE;AAClE,MAAM,UAAU,GAAG,CAAC,KAA8B,EAAE,IAAY,EAAE,IAAwB,EAAW,EAAE;IACrG,MAAM,MAAM,GAAY,KAAK,CAAC,IAAI,CAAC,CAAC;IACpC,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,qBAAqB,IAAI,WAAW,CAAC,CAAC;IAC/E,OAAO,OAAO,CAAC,KAAK,CAAC,MAAM,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC;AAC5C,CAAC,CAAC;AAEF,MAAM,KAAK,GAAG,GAAW,EAAE;IACzB,MAAM,OAAO,GAAY,OAAO,CAAC,iBAAiB,CAAC,CAAC;IACpD,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAC;IAChF,MAAM,MAAM,GAAY,OAAO,EAAE,CAAC;IAClC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAC;IAC7E,OAAO,MAAM,CAAC;AAChB,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,MAAM,GAAqC;IAC/C,EAAE,EAAE,OAAO;IACX,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,OAAO;IACZ,IAAI,EAAE,OAAO;IACb,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,IAAI,EAAE,MAAM;IACZ,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,KAAK;CACX,CAAC;AAEF,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE7E,MAAM,CAAC,MAAM,IAAI,GAAG,CAAC,GAAW,EAAU,EAAE,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC;AAErG;;;GAGG;AACH;;;;;;GAMG;AACH,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC;AAE3D,MAAM,EAAE,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;AAE9E,MAAM,IAAI,GAAG,CAAC,KAAa,EAAW,EAAE,CAAC,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC;AAE/H,8BAA8B;AAC9B,MAAM,QAAQ,GAAG,CAAC,MAAyB,EAAE,EAAU,EAAU,EAAE;IACjE,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,aAAa,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC;IACrF,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC;IAC3B,OAAO,KAAK,KAAK,SAAS,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AACxD,CAAC,CAAC;AAEF;;;;;GAKG;AACH,MAAM,eAAe,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC;AAE5D,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,QAAQ,CAAC,CAAC,CAAC;AAEvD,MAAM,WAAW,GAAG,CAAC,KAAa,EAAW,EAAE,CAAC,eAAe,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC;AAErI,wDAAwD;AACxD,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;AAE7E,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC;AAE5C,MAAM,gBAAgB,GAAG,CAAC,MAAyB,EAAE,EAAU,EAAW,EAAE;IAC1E,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,aAAa,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,KAAK,CAAC,CAAC,CAAC;IAC/E,MAAM,QAAQ,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC;IAC9B,IAAI,QAAQ,KAAK,SAAS,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,OAAO,KAAK,CAAC;IAC5E,6DAA6D;IAC7D,MAAM,UAAU,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,QAAQ,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,GAAG,KAAK,GAAG,CAAC,CAAC;IAChF,OAAO,UAAU,KAAK,SAAS,IAAI,WAAW,CAAC,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC;AACrE,CAAC,CAAC;AAEF,MAAM,QAAQ,GAAG,YAAY,EAAE,CAAC;AAEhC,MAAM,OAAO,GAAG,iBAAiB,CAAC,QAAQ,CAAC,CAAC;AAE5C,MAAM,SAAS,GAAG,CAAC,MAAyB,EAAE,EAAU,EAAW,EAAE;IACnE,IAAI,MAAM,CAAC,EAAE,CAAC,EAAE,GAAG,KAAK,KAAK;QAAE,OAAO,KAAK,CAAC;IAC5C,MAAM,EAAE,GAAG,QAAQ,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;IAChC,OAAO,EAAE,KAAK,CAAC,CAAC,IAAI,CAAC,gBAAgB,CAAC,MAAM,EAAE,EAAE,CAAC,IAAI,CAAC,mBAAmB,CAAC,MAAM,EAAE,EAAE,EAAE,OAAO,CAAC,CAAC;AACjG,CAAC,CAAC;AAEF,MAAM,OAAO,GAAG,CAAC,IAAY,EAAuB,EAAE,CAAC,IAAI,GAAG,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;AAEnI,MAAM,QAAQ,GAAG,OAAO,CAAC,SAAS,CAAC,CAAC;AAEpC,+EAA+E;AAC/E,MAAM,oBAAoB,GAAG,OAAO,CAAC,sBAAsB,CAAC,CAAC;AAE7D,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC;AAIhD,kFAAkF;AAClF,MAAM,kBAAkB,GAAG,CAAC,KAAa,EAAY,EAAE;IACrD,IAAI,cAAc,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC;QAAE,OAAO,oBAAoB,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;IACnI,OAAO,KAAK,CAAC,GAAG,KAAK,IAAI,IAAI,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;AAChH,CAAC,CAAC;AAEF,iGAAiG;AACjG,MAAM,UAAU,GAAG,CAAC,MAAyB,EAAE,EAAU,EAAY,EAAE;IACrE,MAAM,KAAK,GAAG,MAAM,CAAC,EAAE,CAAC,CAAC;IACzB,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,EAAE,CAAC;IACnC,IAAI,KAAK,CAAC,GAAG,KAAK,KAAK;QAAE,OAAO,kBAAkB,CAAC,KAAK,CAAC,CAAC;IAC1D,OAAO,EAAE,QAAQ,EAAE,SAAS,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,CAAC;AAC1G,CAAC,CAAC;AAEF,MAAM,KAAK,GAAkC,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC;AAElE,MAAM,CAAC,MAAM,OAAO,GAAG,GAAS,EAAE;IAChC,KAAK,CAAC,KAAK,KAAK,KAAK,EAAE,CAAC;AAC1B,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,OAAO,GAAG,GAAY,EAAE,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC;AAEhE;;;GAGG;AACH,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,MAAyB,EAAW,EAAE,CAClE,MAAM,CAAC,MAAM,CACX,CAAC,GAAG,EAAE,KAAK,EAAE,EAAE,EAAE,EAAE;IACjB,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IACpD,IAAI,KAAK,KAAK,CAAC,CAAC;QAAE,OAAO,GAAG,CAAC;IAC7B,MAAM,GAAG,GAAG,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC;IACvC,MAAM,KAAK,GAAG;QACZ,IAAI,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE;QACpB,OAAO,EAAE,KAAK,CAAC,KAAK;QACpB,GAAG,EAAE,iBAAiB,CAAC,KAAK,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QACpD,GAAG,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC;QAC5D,GAAG,UAAU,CAAC,MAAM,EAAE,EAAE,CAAC;KAC1B,CAAC;IACF,OAAO,EAAE,MAAM,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;AACzD,CAAC,EACD,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,CAC1B,CAAC,MAAM,CAAC;AAEX,MAAM,CAAC,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAuB,EAAE;IAC5D,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC;IAC3B,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC3C,OAAO,MAAM,CAAC,IAAI,EAAE,OAAO,CAAC,UAAU,CAAC,MAAM,EAAE,aAAa,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC;AAC3F,CAAC,CAAC"}
1
+ {"version":3,"file":"pos.js","sourceRoot":"","sources":["../src/pos.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAE5C,OAAO,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAC7C,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,YAAY,EAAE,iBAAiB,EAAE,QAAQ,EAAE,yBAAyB,EAAE,MAAM,kBAAkB,CAAC;AACxG,OAAO,EAAE,mBAAmB,EAAE,iBAAiB,EAAE,MAAM,yBAAyB,CAAC;AAEjF,MAAM,OAAO,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAO/C,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAoC,EAAE,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC;AAEnH,MAAM,UAAU,GAAG,CAAC,KAAc,EAAqD,EAAE,CAAC,OAAO,KAAK,KAAK,UAAU,CAAC;AAEtH,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAmB,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,OAAO,KAAK,CAAC,OAAO,CAAC,KAAK,QAAQ,IAAI,OAAO,KAAK,CAAC,KAAK,CAAC,KAAK,QAAQ,CAAC;AAEhJ,MAAM,OAAO,GAAG,CAAC,KAAc,EAAsB,EAAE,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;AAE5F,kEAAkE;AAClE,MAAM,UAAU,GAAG,CAAC,KAA8B,EAAE,IAAY,EAAE,IAAwB,EAAW,EAAE;IACrG,MAAM,MAAM,GAAY,KAAK,CAAC,IAAI,CAAC,CAAC;IACpC,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,qBAAqB,IAAI,WAAW,CAAC,CAAC;IAC/E,OAAO,OAAO,CAAC,KAAK,CAAC,MAAM,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC;AAC5C,CAAC,CAAC;AAEF,MAAM,KAAK,GAAG,GAAW,EAAE;IACzB,MAAM,OAAO,GAAY,OAAO,CAAC,iBAAiB,CAAC,CAAC;IACpD,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAC;IAChF,MAAM,MAAM,GAAY,OAAO,EAAE,CAAC;IAClC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAC;IAC7E,OAAO,MAAM,CAAC;AAChB,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,MAAM,GAAqC;IAC/C,EAAE,EAAE,OAAO;IACX,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,OAAO;IACZ,IAAI,EAAE,OAAO;IACb,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,IAAI,EAAE,MAAM;IACZ,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,KAAK;CACX,CAAC;AAEF,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE7E,MAAM,CAAC,MAAM,IAAI,GAAG,CAAC,GAAW,EAAU,EAAE,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC;AAErG;;;GAGG;AACH;;;;;;GAMG;AACH,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC;AAE3D,MAAM,EAAE,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;AAE9E,MAAM,IAAI,GAAG,CAAC,KAAa,EAAW,EAAE,CAAC,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC;AAE/H,8BAA8B;AAC9B,MAAM,QAAQ,GAAG,CAAC,MAAyB,EAAE,EAAU,EAAU,EAAE;IACjE,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,aAAa,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC;IACrF,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC;IAC3B,OAAO,KAAK,KAAK,SAAS,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AACxD,CAAC,CAAC;AAEF;;;;;GAKG;AACH,MAAM,eAAe,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC;AAE5D,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,QAAQ,CAAC,CAAC,CAAC;AAEvD,MAAM,WAAW,GAAG,CAAC,KAAa,EAAW,EAAE,CAAC,eAAe,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC;AAErI,wDAAwD;AACxD,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC;AAE7E,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC;AAE5C,MAAM,gBAAgB,GAAG,CAAC,MAAyB,EAAE,EAAU,EAAW,EAAE;IAC1E,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,aAAa,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,KAAK,CAAC,CAAC,CAAC;IAC/E,MAAM,QAAQ,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC;IAC9B,IAAI,QAAQ,KAAK,SAAS,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,OAAO,KAAK,CAAC;IAC5E,6DAA6D;IAC7D,MAAM,UAAU,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,QAAQ,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,GAAG,KAAK,GAAG,CAAC,CAAC;IAChF,OAAO,UAAU,KAAK,SAAS,IAAI,WAAW,CAAC,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC;AACrE,CAAC,CAAC;AAEF,MAAM,QAAQ,GAAG,YAAY,EAAE,CAAC;AAEhC,MAAM,OAAO,GAAG,iBAAiB,CAAC,QAAQ,CAAC,CAAC;AAE5C,MAAM,SAAS,GAAG,CAAC,MAAyB,EAAE,EAAU,EAAW,EAAE;IACnE,IAAI,MAAM,CAAC,EAAE,CAAC,EAAE,GAAG,KAAK,KAAK;QAAE,OAAO,KAAK,CAAC;IAC5C,MAAM,EAAE,GAAG,QAAQ,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC;IAChC,OAAO,EAAE,KAAK,CAAC,CAAC,IAAI,CAAC,gBAAgB,CAAC,MAAM,EAAE,EAAE,CAAC,IAAI,CAAC,mBAAmB,CAAC,MAAM,EAAE,EAAE,EAAE,OAAO,CAAC,CAAC;AACjG,CAAC,CAAC;AAEF,MAAM,OAAO,GAAG,CAAC,IAAY,EAAuB,EAAE,CAAC,IAAI,GAAG,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;AAEnI,MAAM,QAAQ,GAAG,OAAO,CAAC,SAAS,CAAC,CAAC;AAEpC,+EAA+E;AAC/E,MAAM,oBAAoB,GAAG,OAAO,CAAC,sBAAsB,CAAC,CAAC;AAE7D,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC;AAIhD,kFAAkF;AAClF,MAAM,kBAAkB,GAAG,CAAC,KAAa,EAAY,EAAE;IACrD,IAAI,cAAc,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC;QAAE,OAAO,oBAAoB,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC;IACnI,OAAO,KAAK,CAAC,GAAG,KAAK,IAAI,IAAI,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,QAAQ,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;AAChH,CAAC,CAAC;AAEF,iFAAiF;AACjF,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,CAAC;AAE5C,MAAM,wBAAwB,GAAG,CAAC,KAAa,EAAY,EAAE,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,MAAM,EAAE,MAAM,EAAE,EAAE,CAAC,CAAC,CAAC,kBAAkB,CAAC,KAAK,CAAC,CAAC,CAAC;AAEzJ,iGAAiG;AACjG,MAAM,UAAU,GAAG,CAAC,MAAyB,EAAE,EAAU,EAAY,EAAE;IACrE,MAAM,KAAK,GAAG,MAAM,CAAC,EAAE,CAAC,CAAC;IACzB,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,EAAE,CAAC;IACnC,IAAI,KAAK,CAAC,GAAG,KAAK,KAAK;QAAE,OAAO,wBAAwB,CAAC,KAAK,CAAC,CAAC;IAChE,OAAO,EAAE,QAAQ,EAAE,SAAS,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,CAAC;AAC1G,CAAC,CAAC;AAKF,MAAM,MAAM,GAAG,CAAC,KAAc,EAA8B,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,OAAO,GAAG,KAAK,QAAQ,CAAC,CAAC;AAErI,MAAM,eAAe,GAAG,GAAe,EAAE;IACvC,MAAM,KAAK,GAAY,aAAa,CAAC,OAAO,CAAC,OAAO,CAAC,iBAAiB,CAAC,CAAC,CAAC,6BAA6B,CAAC,CAAC;IACxG,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,kCAAkC,CAAC,CAAC;IAC1E,OAAO,CAAC,IAAI,EAAE,EAAE;QACd,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;QAClE,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,SAAS,CAAC;IACzC,CAAC,CAAC;AACJ,CAAC,CAAC;AAEF,MAAM,KAAK,GAA0D,EAAE,KAAK,EAAE,SAAS,EAAE,UAAU,EAAE,GAAG,EAAE,CAAC,SAAS,EAAE,CAAC;AAEvH,MAAM,CAAC,MAAM,OAAO,GAAG,GAAS,EAAE;IAChC,IAAI,KAAK,CAAC,KAAK,KAAK,SAAS;QAAE,OAAO;IACtC,KAAK,CAAC,KAAK,GAAG,KAAK,EAAE,CAAC;IACtB,KAAK,CAAC,UAAU,GAAG,eAAe,EAAE,CAAC;AACvC,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,OAAO,GAAG,GAAY,EAAE,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC;AAEhE;;;GAGG;AACH,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,MAAyB,EAAW,EAAE,CAClE,MAAM,CAAC,MAAM,CACX,CAAC,GAAG,EAAE,KAAK,EAAE,EAAE,EAAE,EAAE;IACjB,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IACpD,IAAI,KAAK,KAAK,CAAC,CAAC;QAAE,OAAO,GAAG,CAAC;IAC7B,MAAM,GAAG,GAAG,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC;IACvC,MAAM,KAAK,GAAG;QACZ,IAAI,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE;QACpB,OAAO,EAAE,KAAK,CAAC,KAAK;QACpB,GAAG,EAAE,iBAAiB,CAAC,KAAK,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QACpD,GAAG,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC;QAC5D,GAAG,UAAU,CAAC,MAAM,EAAE,EAAE,CAAC;KAC1B,CAAC;IACF,OAAO,EAAE,MAAM,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;AACzD,CAAC,EACD,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,CAC1B,CAAC,MAAM,CAAC;AAEX,sCAAsC;AACtC,MAAM,SAAS,GAAG,IAAI,CAAC;AAEvB,MAAM,MAAM,GAAG,CAAC,MAAc,EAAE,IAAY,EAAqB,EAAE,CAAC,OAAO,CAAC,UAAU,CAAC,MAAM,EAAE,aAAa,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;AAExI,kEAAkE;AAClE,MAAM,uBAAuB,GAAG,CAAC,MAAc,EAAE,IAAY,EAAE,OAA0B,EAAqB,EAAE;IAC9G,MAAM,EAAE,GAAG,yBAAyB,CAAC,OAAO,EAAE,KAAK,CAAC,UAAU,CAAC,CAAC;IAChE,MAAM,OAAO,GAAG,YAAY,CAAC,IAAI,EAAE,OAAO,CAAC,EAAE,CAAC,EAAE,KAAK,CAAC,CAAC;IACvD,OAAO,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,EAAE,EAAE,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;AAC1F,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAuB,EAAE;IAC5D,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC;IAC3B,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC3C,MAAM,KAAK,GAAG,aAAa,CAAC,IAAI,EAAE,SAAS,CAAC,CAAC;IAC7C,OAAO,MAAM,CAAC,KAAK,EAAE,uBAAuB,CAAC,MAAM,EAAE,KAAK,EAAE,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC;AACtF,CAAC,CAAC"}
@@ -3,4 +3,25 @@
3
3
  * 小文字だけで書いた語(linters、json、e.g. の e)は普通名詞、文字の無い語(—、$)は記号。大文字の無い文字(東京)の語は確かめようがないので触れない。
4
4
  */
5
5
  export declare const properNounChecked: (surface: string, pos: string) => string;
6
+ /** 解析器が返す 1 語。pos は Penn Treebank。 */
7
+ type TaggedWord = {
8
+ readonly value: string;
9
+ readonly pos: string;
10
+ readonly lemma?: string;
11
+ };
12
+ /**
13
+ * 文頭で大文字になっただけの普通の語の位置。無ければ -1。
14
+ * wink は大文字で始まる名詞と形容詞をすべて固有名詞にするので、Containers are … の Containers も Traditional servers … の Traditional も固有名詞になる。
15
+ * 文頭の大文字は英語の書き方で、名前の印ではない。解析器の語彙が小文字の形を普通の語として知っていれば、名前ではなくその語として読む。
16
+ * 語彙に無い語(Kubernetes、Congress)と、文の途中の大文字の語は名前のまま。
17
+ */
18
+ export declare const sentenceInitialCommonWord: (tagged: readonly TaggedWord[], tagsOf: (word: string) => readonly string[] | undefined) => number;
19
+ /** 文の中の最初の word を小文字にした文。word が文に無ければ undefined。 */
20
+ export declare const lowercasedAt: (text: string, word: string | undefined) => string | undefined;
21
+ /**
22
+ * at の語の品詞と原形を、小文字にして解析し直した結果(again)から取る。表層は元のまま。
23
+ * 解析し直した文の同じ位置に、同じ語の小文字が無ければ、語の切り方が変わったので元のまま。
24
+ */
25
+ export declare const rereadAt: (entries: readonly TaggedWord[], at: number, again: readonly TaggedWord[]) => readonly TaggedWord[];
26
+ export {};
6
27
  //# sourceMappingURL=proper-noun.d.ts.map
@@ -1 +1 @@
1
- {"version":3,"file":"proper-noun.d.ts","sourceRoot":"","sources":["../src/proper-noun.ts"],"names":[],"mappings":"AAYA;;;GAGG;AACH,eAAO,MAAM,iBAAiB,GAAI,SAAS,MAAM,EAAE,KAAK,MAAM,KAAG,MAIhE,CAAC"}
1
+ {"version":3,"file":"proper-noun.d.ts","sourceRoot":"","sources":["../src/proper-noun.ts"],"names":[],"mappings":"AAYA;;;GAGG;AACH,eAAO,MAAM,iBAAiB,GAAI,SAAS,MAAM,EAAE,KAAK,MAAM,KAAG,MAIhE,CAAC;AAEF,sCAAsC;AACtC,KAAK,UAAU,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,MAAM,CAAA;CAAE,CAAC;AAU5F;;;;;GAKG;AACH,eAAO,MAAM,yBAAyB,GAAI,QAAQ,SAAS,UAAU,EAAE,EAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,KAAK,SAAS,MAAM,EAAE,GAAG,SAAS,KAAG,MAKlI,CAAC;AAEF,oDAAoD;AACpD,eAAO,MAAM,YAAY,GAAI,MAAM,MAAM,EAAE,MAAM,MAAM,GAAG,SAAS,KAAG,MAAM,GAAG,SAI9E,CAAC;AAEF;;;GAGG;AACH,eAAO,MAAM,QAAQ,GAAI,SAAS,SAAS,UAAU,EAAE,EAAE,IAAI,MAAM,EAAE,OAAO,SAAS,UAAU,EAAE,KAAG,SAAS,UAAU,EAMtH,CAAC"}