@chaffjs/lang-en 0.25.0 → 0.26.0

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 (93) hide show
  1. package/dist/definitions.d.ts.map +1 -1
  2. package/dist/definitions.js +15 -5
  3. package/dist/definitions.js.map +1 -1
  4. package/dist/example-name.d.ts +7 -0
  5. package/dist/example-name.d.ts.map +1 -0
  6. package/dist/example-name.js +33 -0
  7. package/dist/example-name.js.map +1 -0
  8. package/dist/structure.d.ts.map +1 -1
  9. package/dist/structure.js +19 -1
  10. package/dist/structure.js.map +1 -1
  11. package/lexicons/action-due-label.yaml +8 -0
  12. package/lexicons/action-section-heading.yaml +9 -0
  13. package/lexicons/amount-bound-label.yaml +14 -0
  14. package/lexicons/amount-change-word.yaml +33 -0
  15. package/lexicons/amount-range-word.yaml +6 -0
  16. package/lexicons/amount-short-scale.yaml +9 -0
  17. package/lexicons/article-sound.yaml +3 -0
  18. package/lexicons/code-global.yaml +27 -0
  19. package/lexicons/common-acronym.yaml +19 -0
  20. package/lexicons/company-form.yaml +14 -11
  21. package/lexicons/count-counter.yaml +5 -0
  22. package/lexicons/count-hedge.yaml +32 -1
  23. package/lexicons/count-member.yaml +25 -0
  24. package/lexicons/day-shift-mark.yaml +12 -0
  25. package/lexicons/deadline-passed-word.yaml +10 -0
  26. package/lexicons/default-header.yaml +7 -0
  27. package/lexicons/default-phrase.yaml +8 -0
  28. package/lexicons/deprecation-negated.yaml +10 -0
  29. package/lexicons/deprecation-word.yaml +8 -0
  30. package/lexicons/doubled-nado.yaml +15 -0
  31. package/lexicons/due-date-label.yaml +10 -0
  32. package/lexicons/duty-marker.yaml +9 -0
  33. package/lexicons/duty-negation.yaml +7 -0
  34. package/lexicons/duty-terms-elsewhere.yaml +10 -0
  35. package/lexicons/earlier-section-word.yaml +8 -0
  36. package/lexicons/extra-word.yaml +5 -0
  37. package/lexicons/fact-change-from.yaml +6 -0
  38. package/lexicons/fact-change-not.yaml +13 -0
  39. package/lexicons/fact-change-to.yaml +6 -0
  40. package/lexicons/fact-condition-from.yaml +10 -0
  41. package/lexicons/fact-condition-to.yaml +9 -0
  42. package/lexicons/fact-event.yaml +14 -0
  43. package/lexicons/fact-label-lead.yaml +9 -0
  44. package/lexicons/fact-name-joiner.yaml +6 -0
  45. package/lexicons/fact-separator.yaml +5 -0
  46. package/lexicons/install-command.yaml +20 -0
  47. package/lexicons/issue-date-label.yaml +9 -0
  48. package/lexicons/item-done-word.yaml +9 -0
  49. package/lexicons/leg-column.yaml +15 -0
  50. package/lexicons/leg-mark.yaml +25 -0
  51. package/lexicons/line-amount-column.yaml +8 -0
  52. package/lexicons/local-time-mark.yaml +7 -0
  53. package/lexicons/meeting-date-label.yaml +8 -0
  54. package/lexicons/minutes-title-word.yaml +7 -0
  55. package/lexicons/nado-closed-list.yaml +22 -0
  56. package/lexicons/number-word.yaml +45 -0
  57. package/lexicons/ordinal-frame.yaml +15 -0
  58. package/lexicons/parameter-header.yaml +11 -0
  59. package/lexicons/partial-listing-marker.yaml +15 -0
  60. package/lexicons/per-unit-mark.yaml +25 -0
  61. package/lexicons/period-aside.yaml +46 -0
  62. package/lexicons/period-label.yaml +19 -0
  63. package/lexicons/person-not-after.yaml +42 -1
  64. package/lexicons/phone-label.yaml +14 -0
  65. package/lexicons/place-name-joiner.yaml +10 -0
  66. package/lexicons/place-word.yaml +29 -0
  67. package/lexicons/prerequisite-opener.yaml +10 -0
  68. package/lexicons/quantity-column.yaml +9 -0
  69. package/lexicons/rate-unit.yaml +18 -0
  70. package/lexicons/share-column-rate.yaml +15 -0
  71. package/lexicons/share-column.yaml +12 -0
  72. package/lexicons/share-rest.yaml +8 -0
  73. package/lexicons/stated-deadline.yaml +17 -0
  74. package/lexicons/stay-check-in-column.yaml +10 -0
  75. package/lexicons/stay-check-out-column.yaml +10 -0
  76. package/lexicons/stay-night.yaml +6 -0
  77. package/lexicons/stay-nights-column.yaml +7 -0
  78. package/lexicons/stay-pair-joiner.yaml +9 -0
  79. package/lexicons/summed-counter.yaml +17 -0
  80. package/lexicons/tax-included-word.yaml +8 -0
  81. package/lexicons/tax-label.yaml +9 -0
  82. package/lexicons/timed-duty.yaml +14 -0
  83. package/lexicons/unit-price-column.yaml +9 -0
  84. package/lexicons/unnamed-speaker.yaml +58 -0
  85. package/lexicons/us-state-code.yaml +1 -1
  86. package/lexicons/very-copula.yaml +5 -0
  87. package/lexicons/warning-action.yaml +20 -0
  88. package/lexicons/warning-label.yaml +8 -0
  89. package/lexicons/word-suffix.yaml +3 -0
  90. package/package.json +1 -1
  91. package/src/definitions.ts +13 -5
  92. package/src/example-name.ts +36 -0
  93. package/src/structure.ts +24 -1
@@ -0,0 +1,17 @@
1
+ # Ways of giving a duty its time besides a number of days (concrete-deadline): a date or day ("by the fifth business day",
2
+ # "on or before March 1"), or at once ("immediately"). A sentence holding one has a deadline, and obligation-without-deadline
3
+ # leaves it. "by" alone is not here: "by email" says how, not when.
4
+ id: stated-deadline
5
+ language: en
6
+ entries:
7
+ - pattern: by the
8
+ - pattern: on or before
9
+ - pattern: prior to
10
+ - pattern: before
11
+ - pattern: immediately
12
+ - pattern: at the end of
13
+ - pattern: on the … day
14
+ - pattern: at the time
15
+ - pattern: within the
16
+ # "timely" leaves the time to the reader, but it speaks of time: vague-deadline's kind, not a duty with none.
17
+ - pattern: timely
@@ -0,0 +1,10 @@
1
+ # Headings of the first-day column of a stay table (duration-mismatch). With a check-out and a nights column, each row's nights are compared with its dates.
2
+ id: stay-check-in-column
3
+ language: en
4
+ entries:
5
+ - pattern: "Check-in"
6
+ - pattern: "Check in"
7
+ - pattern: "Check-in date"
8
+ - pattern: "Arrive"
9
+ - pattern: "Arrival"
10
+ - pattern: "Arrival date"
@@ -0,0 +1,10 @@
1
+ # Headings of the leaving-day column of a stay table (duration-mismatch).
2
+ id: stay-check-out-column
3
+ language: en
4
+ entries:
5
+ - pattern: "Check-out"
6
+ - pattern: "Check out"
7
+ - pattern: "Check-out date"
8
+ - pattern: "Depart"
9
+ - pattern: "Departure"
10
+ - pattern: "Departure date"
@@ -0,0 +1,6 @@
1
+ # Units of a number of nights (duration-mismatch). A count in one of them is compared with the days from check-in to check-out.
2
+ id: stay-night
3
+ language: en
4
+ entries:
5
+ - pattern: "night"
6
+ - pattern: "nights"
@@ -0,0 +1,7 @@
1
+ # Headings of the nights column of a stay table (duration-mismatch).
2
+ id: stay-nights-column
3
+ language: en
4
+ entries:
5
+ - pattern: "Nights"
6
+ - pattern: "No. of nights"
7
+ - pattern: "Number of nights"
@@ -0,0 +1,9 @@
1
+ # What may stand between a number of nights and a number of days ("2 days, 1 night"), where the days are one more than the nights (duration-mismatch).
2
+ id: stay-pair-joiner
3
+ language: en
4
+ entries:
5
+ - pattern: ","
6
+ - pattern: "/"
7
+ - pattern: "and"
8
+ - pattern: ", and"
9
+ - pattern: "&"
@@ -0,0 +1,17 @@
1
+ # Counters that a total line may add up ("16 credits", "4 sessions"). total-mismatch reads a number directly followed by one
2
+ # of these as a quantity in that unit. The singular and the plural share a group, so "1 credit" and "3 credits" add up.
3
+ id: summed-counter
4
+ language: en
5
+ entries:
6
+ - pattern: credits
7
+ group: credit
8
+ - pattern: credit
9
+ group: credit
10
+ - pattern: units
11
+ group: unit
12
+ - pattern: unit
13
+ group: unit
14
+ - pattern: sessions
15
+ group: session
16
+ - pattern: session
17
+ group: session
@@ -0,0 +1,8 @@
1
+ # Words that mark a tax line as tax included in the total, not added to it. Such a line is not compared (tax-mismatch).
2
+ id: tax-included-word
3
+ language: en
4
+ entries:
5
+ - pattern: included
6
+ - pattern: includes
7
+ - pattern: incl.
8
+ - pattern: inclusive
@@ -0,0 +1,9 @@
1
+ # Words that mark a tax line. A list or table line starting with one and giving a rate (10%) is compared with that rate of the subtotal above it (tax-mismatch).
2
+ id: tax-label
3
+ language: en
4
+ entries:
5
+ - pattern: Sales tax
6
+ - pattern: Consumption tax
7
+ - pattern: Tax
8
+ - pattern: VAT
9
+ - pattern: GST
@@ -0,0 +1,14 @@
1
+ # Acts a duty binds to a point in time (pay, deliver, return, notify). After a duty marker (shall, will) in a sentence that
2
+ # states no limit, nobody can say by when the act is due (obligation-without-deadline). "Indemnify" and "keep" are not here:
3
+ # they bind for as long as the contract runs, not at a time.
4
+ id: timed-duty
5
+ language: en
6
+ entries:
7
+ - pattern: pay
8
+ - pattern: remit
9
+ - pattern: refund
10
+ - pattern: invoice
11
+ - pattern: deliver
12
+ - pattern: submit
13
+ - pattern: return
14
+ - pattern: notify
@@ -0,0 +1,9 @@
1
+ # Headings of the unit price column of an itemised table (line-amount-mismatch).
2
+ id: unit-price-column
3
+ language: en
4
+ entries:
5
+ - pattern: Unit price
6
+ - pattern: Unit cost
7
+ - pattern: Price
8
+ # A rate is a price per a period of time, and the header does not say which (group: a unit no quantity word counts).
9
+ - { pattern: Rate, group: unstated }
@@ -0,0 +1,58 @@
1
+ # Words that quote someone without a name (quote-unnamed-speaker). group says what each word does:
2
+ # role: a role written in place of a speaker's name. A role that is one person (the CEO of Acme) is not one.
3
+ # after-role: a word that may follow a role at the head of the speaker's phrase (an analyst at Acme, a source close to).
4
+ # A role followed by anything else (an official statement) is not the speaker.
5
+ # honorific: a title before a name (Ms Mori): the phrase names its speaker.
6
+ # cue: words that give a quotation to its speaker, besides quote-attribution's (told reporters, noted an analyst).
7
+ # anaphor: a word right before the role that points back at someone (the analyst, our spokesperson): not reported.
8
+ id: unnamed-speaker
9
+ language: en
10
+ entries:
11
+ - { pattern: analyst, group: role }
12
+ - { pattern: analysts, group: role }
13
+ - { pattern: expert, group: role }
14
+ - { pattern: experts, group: role }
15
+ - { pattern: source, group: role }
16
+ - { pattern: sources, group: role }
17
+ - { pattern: official, group: role }
18
+ - { pattern: officials, group: role }
19
+ - { pattern: insider, group: role }
20
+ - { pattern: insiders, group: role }
21
+ - { pattern: spokesperson, group: role }
22
+ - { pattern: spokesman, group: role }
23
+ - { pattern: spokeswoman, group: role }
24
+ - { pattern: observer, group: role }
25
+ - { pattern: observers, group: role }
26
+ - { pattern: at, group: after-role }
27
+ - { pattern: with, group: after-role }
28
+ - { pattern: from, group: after-role }
29
+ - { pattern: for, group: after-role }
30
+ - { pattern: in, group: after-role }
31
+ - { pattern: of, group: after-role }
32
+ - { pattern: on, group: after-role }
33
+ - { pattern: close, group: after-role }
34
+ - { pattern: familiar, group: after-role }
35
+ - { pattern: who, group: after-role }
36
+ - { pattern: Mr, group: honorific }
37
+ - { pattern: Mrs, group: honorific }
38
+ - { pattern: Ms, group: honorific }
39
+ - { pattern: Dr, group: honorific }
40
+ - { pattern: Prof, group: honorific }
41
+ - { pattern: say, group: cue }
42
+ - { pattern: told, group: cue }
43
+ - { pattern: tells, group: cue }
44
+ - { pattern: noted, group: cue }
45
+ - { pattern: notes, group: cue }
46
+ - { pattern: added, group: cue }
47
+ - { pattern: adds, group: cue }
48
+ - { pattern: explained, group: cue }
49
+ - { pattern: the, group: anaphor }
50
+ - { pattern: this, group: anaphor }
51
+ - { pattern: that, group: anaphor }
52
+ - { pattern: these, group: anaphor }
53
+ - { pattern: those, group: anaphor }
54
+ - { pattern: our, group: anaphor }
55
+ - { pattern: its, group: anaphor }
56
+ - { pattern: their, group: anaphor }
57
+ - { pattern: his, group: anaphor }
58
+ - { pattern: her, group: anaphor }
@@ -1,5 +1,5 @@
1
1
  # US postal codes of the states, the district and the territories, read only inside an address
2
- # (, CA 94720). Matched as written.
2
+ # (, CA 94720) or after a place name (Chicago, IL). Matched as written.
3
3
  id: us-state-code
4
4
  language: en
5
5
  entries:
@@ -0,0 +1,5 @@
1
+ # Words after a noun that make it an adjective, for very-adjective. English tags its adjectives as adjectives, so it has
2
+ # none; Japanese lists な and だ here (便利な), whose adjectival nouns are tagged as nouns.
3
+ id: very-copula
4
+ language: en
5
+ entries: []
@@ -0,0 +1,20 @@
1
+ # Words that make a sentence of a warning tell the reader what to do, matched as whole words ("Do not unplug the drive",
2
+ # "Make sure the export finished"). A sentence that opens with its verb ("Export today's orders first") is an action too,
3
+ # read from the parts of speech.
4
+ id: warning-action
5
+ language: en
6
+ entries:
7
+ - pattern: do not
8
+ - pattern: don't
9
+ - pattern: never
10
+ - pattern: always
11
+ - pattern: make sure
12
+ - pattern: be sure
13
+ - pattern: ensure
14
+ - pattern: avoid
15
+ - pattern: must
16
+ - pattern: should
17
+ - pattern: need to
18
+ - pattern: have to
19
+ - pattern: please
20
+ - pattern: before you
@@ -0,0 +1,8 @@
1
+ # Words a warning opens with, standing alone ("Warning:", "**Caution**", "> [!WARNING]"). warning-without-action reads the
2
+ # paragraph after one. "Note" and "Tip" are left out: they give information, not a risk.
3
+ id: warning-label
4
+ language: en
5
+ entries:
6
+ - pattern: warning
7
+ - pattern: caution
8
+ - pattern: danger
@@ -11,6 +11,9 @@ entries:
11
11
  instead_of: ""
12
12
  - pattern: ies
13
13
  instead_of: "y"
14
+ # The Greek plural: "pneumoconioses" is "pneumoconiosis".
15
+ - pattern: ses
16
+ instead_of: "sis"
14
17
  - pattern: ied
15
18
  instead_of: "y"
16
19
  - pattern: ed
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@chaffjs/lang-en",
3
- "version": "0.25.0",
3
+ "version": "0.26.0",
4
4
  "description": "English language adapter for chaff",
5
5
  "license": "MIT",
6
6
  "author": "isamu",
@@ -1,5 +1,6 @@
1
1
  import type { Mention } from "chaffjs/plugin";
2
2
  import { CHAPTER_DEPTH, PART_DEPTH } from "./depth.ts";
3
+ import { readsAsExample } from "./example-name.ts";
3
4
 
4
5
  // Definitions in English documents, and how far a definition holds.
5
6
 
@@ -69,12 +70,18 @@ const namesInList = (inside: string): QuotedName[] => {
69
70
  const nameListDefinitions = (text: string): Mention[] =>
70
71
  [...text.matchAll(BRACKET)].flatMap((match) => {
71
72
  const inside = match.groups?.["inside"] ?? "";
72
- const bare = BARE.test(match[0]) ? { form: "bare" } : {};
73
+ const isBare = BARE.test(match[0]);
74
+ const bare = isBare ? { form: "bare" } : {};
73
75
  const scope = APPLIES.test(text) ? { scope: "local" } : {};
74
- return namesInList(inside).map((name, index) => {
76
+ const names = namesInList(inside);
77
+ // A phrase first makes the bracket a list of examples; a phrase later is an example beside a name (("Cheers", "I hope
78
+ // this helps")), or an alias in sentence case (("Provider", "service provider")). The name is kept, the phrase is not.
79
+ if (isBare && readsAsExample(names[0]?.term ?? "")) return [];
80
+ return names.flatMap((name, index) => {
81
+ if (isBare && readsAsExample(name.term)) return [];
75
82
  const start = index === 0 ? match.index : match.index + 1 + name.start;
76
83
  const end = index === 0 ? match.index + match[0].length : match.index + 1 + name.end;
77
- return { start, end, attrs: { term: name.term, ...scope, ...(index === 0 ? { placement: "inline" } : {}), ...bare } };
84
+ return [{ start, end, attrs: { term: name.term, ...scope, ...(index === 0 ? { placement: "inline" } : {}), ...bare } }];
78
85
  });
79
86
  });
80
87
 
@@ -82,12 +89,13 @@ const patternDefinitions = (text: string): Mention[] =>
82
89
  DEFINITIONS.flatMap((pattern) =>
83
90
  [...text.matchAll(pattern)].flatMap((match) => {
84
91
  const term = match.groups?.["term"];
85
- if (term === undefined) return [];
92
+ const isBare = BARE.test(match[0]);
93
+ if (term === undefined || (isBare && readsAsExample(term))) return [];
86
94
  const end = match.index + match[0].length;
87
95
  const inline = match[0].startsWith("(");
88
96
  const namesAParty = inline && APPLIES.test(text);
89
97
  const scope = isPointer(text, end) || namesAParty ? { scope: "local" } : {};
90
- const bare = BARE.test(match[0]) ? { form: "bare" } : {};
98
+ const bare = isBare ? { form: "bare" } : {};
91
99
  return [{ start: match.index, end, attrs: { term, ...scope, ...(inline ? { placement: "inline" } : {}), ...bare } }];
92
100
  }),
93
101
  );
@@ -0,0 +1,36 @@
1
+ // Whether a name quoted in bare brackets (("The key point is")) reads as a quoted example rather than a defined term.
2
+
3
+ /**
4
+ * Words a capitalised name keeps in lower case: articles, conjunctions, prepositions, determiners and the particles of
5
+ * personal and company names (Terms of Use, Fees under this Agreement, Banco de Chile, Ludwig van Beethoven).
6
+ */
7
+ const MINOR_WORDS: ReadonlySet<string> = new Set(
8
+ [
9
+ "a an the and or but nor so yet",
10
+ "of in on at to for with as by from per via vs into onto upon under over between among within without about after",
11
+ "before against along across around through throughout during until toward towards beneath beside besides beyond",
12
+ "near off out up down than like unless whether arising",
13
+ "this that these those each every any all such its their his her our your",
14
+ "de del della di da du des la le van von der den ter ten y e",
15
+ ]
16
+ .join(" ")
17
+ .split(" "),
18
+ );
19
+
20
+ const WORD = /[\p{L}\p{N}][\p{L}\p{N}'’-]*/gu;
21
+ const LOWER_START = /^\p{Ll}/u;
22
+ /** A capital inside the word (iOS, eBay) makes it a name, whatever its first letter. */
23
+ const HAS_UPPER = /\p{Lu}/u;
24
+ /** A statute names a party in lower case after an article: ("the seller"), ("a sub-processor"). */
25
+ const ARTICLE_LEAD = /^(?:the|a|an)\s/u;
26
+
27
+ /**
28
+ * A defined term is one word ("Seller"), a capitalised name ("Use of Customer Data"), or a statute's lower-case party
29
+ * after an article ("the annual period"). Several words with one in lower case that a name would capitalise ("The key
30
+ * point is", "silently fails", "sections 44 and 45") quote an example instead.
31
+ */
32
+ export const readsAsExample = (name: string): boolean => {
33
+ const words = [...name.matchAll(WORD)].map((match) => match[0]);
34
+ if (words.length < 2 || ARTICLE_LEAD.test(name)) return false;
35
+ return words.some((word) => LOWER_START.test(word) && !HAS_UPPER.test(word) && !MINOR_WORDS.has(word));
36
+ };
package/src/structure.ts CHANGED
@@ -274,12 +274,35 @@ const currencyBefore = (text: string, at: number): string | undefined => {
274
274
  return CURRENCIES.find((currency) => text.startsWith(currency, end - currency.length));
275
275
  };
276
276
 
277
+ /** A count of days or months is written "six (6)" or "one thousand (1,000)": never "zero (0)", never "two million (2,000,000)". */
278
+ const FEWEST_COUNTED = 1;
279
+ const MOST_COUNTED = 1000;
280
+
281
+ const countsPeriods = (weight: number | undefined): boolean => weight !== undefined && weight >= FEWEST_COUNTED && weight <= MOST_COUNTED;
282
+
283
+ const FIGURE_WORDS: ReadonlySet<string> = new Set((LEXICONS["number-word"] ?? []).filter((entry) => countsPeriods(entry.weight)).map((entry) => entry.pattern));
284
+ const OPEN_BRACKET = " (";
285
+ const NOT_WORD_LETTER = /[^\p{L}-]/u;
286
+ const BRACKET_REACH = 30;
287
+
288
+ /** The word right before " (" is a number word, or ends with one: "six (", "forty-five (". */
289
+ const followsNumberWord = (text: string, start: number): boolean => {
290
+ const before = text.slice(Math.max(0, start - BRACKET_REACH), start);
291
+ if (!before.endsWith(OPEN_BRACKET)) return false;
292
+ const word = before.slice(0, -OPEN_BRACKET.length).split(NOT_WORD_LETTER).at(-1) ?? "";
293
+ return FIGURE_WORDS.has(word.toLowerCase().split("-").at(-1) ?? "");
294
+ };
295
+
296
+ /** "six (6) months": a figure in brackets after a number in words, with the unit after the bracket. */
297
+ const unitAfterBracket = (text: string, start: number, end: number): string | undefined =>
298
+ text[end] === ")" && followsNumberWord(text, start) ? unitAfter(text, end + 1) : undefined;
299
+
277
300
  const quantities = (text: string): Mention[] =>
278
301
  [...text.matchAll(NUMBER_RUN)].flatMap((match) => {
279
302
  const digits = withoutTrailingPunctuation(match[0]);
280
303
  const value = Number(digits.replace(/,/gu, ""));
281
304
  const end = match.index + digits.length;
282
- const unit = unitAfter(text, end) ?? currencyBefore(text, match.index);
305
+ const unit = unitAfter(text, end) ?? unitAfterBracket(text, match.index, end) ?? currencyBefore(text, match.index);
283
306
  return unit === undefined || Number.isNaN(value) || isWordChar(text[match.index - 1]) ? [] : [{ start: match.index, end, attrs: { value, unit } }];
284
307
  });
285
308