@stll/anonymize 2.0.0-alpha.1 → 2.0.0

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 (156) hide show
  1. package/README.md +1 -1
  2. package/dist/index.d.mts +2 -2
  3. package/dist/index.mjs +1 -1
  4. package/dist/native-node.d.mts +18 -4
  5. package/dist/native-node.mjs +1 -1
  6. package/dist/native-node2.d.mts +2 -2
  7. package/dist/native-node2.mjs +106 -12731
  8. package/dist/native-node2.mjs.map +1 -1
  9. package/dist/native.d.mts +526 -714
  10. package/dist/native.mjs.map +1 -1
  11. package/dist/native2.d.mts +2 -2
  12. package/package.json +27 -13
  13. package/dist/address-boundaries.mjs +0 -197
  14. package/dist/address-boundaries.mjs.map +0 -1
  15. package/dist/address-jurisdiction-prefixes.mjs +0 -16
  16. package/dist/address-jurisdiction-prefixes.mjs.map +0 -1
  17. package/dist/address-stop-keywords.mjs +0 -148
  18. package/dist/address-stop-keywords.mjs.map +0 -1
  19. package/dist/address-stopwords.mjs +0 -84
  20. package/dist/address-stopwords.mjs.map +0 -1
  21. package/dist/address-unit-abbreviations.mjs +0 -15
  22. package/dist/address-unit-abbreviations.mjs.map +0 -1
  23. package/dist/allow-list.mjs +0 -196
  24. package/dist/allow-list.mjs.map +0 -1
  25. package/dist/clause-noun-heads.mjs +0 -79
  26. package/dist/clause-noun-heads.mjs.map +0 -1
  27. package/dist/common-words-en.mjs +0 -9887
  28. package/dist/common-words-en.mjs.map +0 -1
  29. package/dist/coreference-org-determiners.mjs +0 -19
  30. package/dist/coreference-org-determiners.mjs.map +0 -1
  31. package/dist/coreference.cs.mjs +0 -14
  32. package/dist/coreference.cs.mjs.map +0 -1
  33. package/dist/coreference.de.mjs +0 -14
  34. package/dist/coreference.de.mjs.map +0 -1
  35. package/dist/coreference.en.mjs +0 -14
  36. package/dist/coreference.en.mjs.map +0 -1
  37. package/dist/coreference.es.mjs +0 -22
  38. package/dist/coreference.es.mjs.map +0 -1
  39. package/dist/coreference.fr.mjs +0 -32
  40. package/dist/coreference.fr.mjs.map +0 -1
  41. package/dist/coreference.it.mjs +0 -27
  42. package/dist/coreference.it.mjs.map +0 -1
  43. package/dist/coreference.pl.mjs +0 -27
  44. package/dist/coreference.pl.mjs.map +0 -1
  45. package/dist/coreference.pt-br.mjs +0 -14
  46. package/dist/coreference.pt-br.mjs.map +0 -1
  47. package/dist/coreference.sk.mjs +0 -27
  48. package/dist/coreference.sk.mjs.map +0 -1
  49. package/dist/currencies.mjs +0 -231
  50. package/dist/currencies.mjs.map +0 -1
  51. package/dist/date-months.mjs +0 -618
  52. package/dist/date-months.mjs.map +0 -1
  53. package/dist/defined-term-heads.mjs +0 -15
  54. package/dist/defined-term-heads.mjs.map +0 -1
  55. package/dist/document-structure-headings.mjs +0 -90
  56. package/dist/document-structure-headings.mjs.map +0 -1
  57. package/dist/false-positive-shapes.mjs +0 -36
  58. package/dist/false-positive-shapes.mjs.map +0 -1
  59. package/dist/generic-roles.mjs +0 -244
  60. package/dist/generic-roles.mjs.map +0 -1
  61. package/dist/hotword-rules.mjs +0 -149
  62. package/dist/hotword-rules.mjs.map +0 -1
  63. package/dist/legal-form-leading-clauses.mjs +0 -23
  64. package/dist/legal-form-leading-clauses.mjs.map +0 -1
  65. package/dist/legal-forms.mjs +0 -2115
  66. package/dist/legal-forms.mjs.map +0 -1
  67. package/dist/legal-role-heads.cs.mjs +0 -48
  68. package/dist/legal-role-heads.cs.mjs.map +0 -1
  69. package/dist/legal-role-heads.de.mjs +0 -33
  70. package/dist/legal-role-heads.de.mjs.map +0 -1
  71. package/dist/legal-role-heads.en.mjs +0 -37
  72. package/dist/legal-role-heads.en.mjs.map +0 -1
  73. package/dist/legal-role-heads.es.mjs +0 -54
  74. package/dist/legal-role-heads.es.mjs.map +0 -1
  75. package/dist/legal-role-heads.fr.mjs +0 -72
  76. package/dist/legal-role-heads.fr.mjs.map +0 -1
  77. package/dist/legal-role-heads.it.mjs +0 -68
  78. package/dist/legal-role-heads.it.mjs.map +0 -1
  79. package/dist/legal-role-heads.pl.mjs +0 -84
  80. package/dist/legal-role-heads.pl.mjs.map +0 -1
  81. package/dist/legal-role-heads.pt-br.mjs +0 -63
  82. package/dist/legal-role-heads.pt-br.mjs.map +0 -1
  83. package/dist/legal-role-heads.sk.mjs +0 -80
  84. package/dist/legal-role-heads.sk.mjs.map +0 -1
  85. package/dist/manifest.mjs +0 -69
  86. package/dist/manifest.mjs.map +0 -1
  87. package/dist/names-exclusions.mjs +0 -223
  88. package/dist/names-exclusions.mjs.map +0 -1
  89. package/dist/names-first.mjs +0 -418
  90. package/dist/names-first.mjs.map +0 -1
  91. package/dist/names-nw-ar.mjs +0 -202
  92. package/dist/names-nw-ar.mjs.map +0 -1
  93. package/dist/names-nw-excluded-allcaps.mjs +0 -112
  94. package/dist/names-nw-excluded-allcaps.mjs.map +0 -1
  95. package/dist/names-nw-fil.mjs +0 -202
  96. package/dist/names-nw-fil.mjs.map +0 -1
  97. package/dist/names-nw-id.mjs +0 -210
  98. package/dist/names-nw-id.mjs.map +0 -1
  99. package/dist/names-nw-in.mjs +0 -526
  100. package/dist/names-nw-in.mjs.map +0 -1
  101. package/dist/names-nw-ja-latn.mjs +0 -260
  102. package/dist/names-nw-ja-latn.mjs.map +0 -1
  103. package/dist/names-nw-ko.mjs +0 -162
  104. package/dist/names-nw-ko.mjs.map +0 -1
  105. package/dist/names-nw-th.mjs +0 -188
  106. package/dist/names-nw-th.mjs.map +0 -1
  107. package/dist/names-nw-vi.mjs +0 -151
  108. package/dist/names-nw-vi.mjs.map +0 -1
  109. package/dist/names-nw-zh-latn.mjs +0 -197
  110. package/dist/names-nw-zh-latn.mjs.map +0 -1
  111. package/dist/names-surnames.mjs +0 -113
  112. package/dist/names-surnames.mjs.map +0 -1
  113. package/dist/names-title-tokens.mjs +0 -40
  114. package/dist/names-title-tokens.mjs.map +0 -1
  115. package/dist/organization-unit-heads.mjs +0 -20
  116. package/dist/organization-unit-heads.mjs.map +0 -1
  117. package/dist/person-stopwords.mjs +0 -211
  118. package/dist/person-stopwords.mjs.map +0 -1
  119. package/dist/section-headings.mjs +0 -64
  120. package/dist/section-headings.mjs.map +0 -1
  121. package/dist/sentence-verb-indicators.mjs +0 -232
  122. package/dist/sentence-verb-indicators.mjs.map +0 -1
  123. package/dist/signing-clauses.mjs +0 -102
  124. package/dist/signing-clauses.mjs.map +0 -1
  125. package/dist/stopwords.mjs +0 -9915
  126. package/dist/stopwords.mjs.map +0 -1
  127. package/dist/structural-single-cap-prefixes.mjs +0 -99
  128. package/dist/structural-single-cap-prefixes.mjs.map +0 -1
  129. package/dist/triggers.cs.mjs +0 -569
  130. package/dist/triggers.cs.mjs.map +0 -1
  131. package/dist/triggers.de.mjs +0 -139
  132. package/dist/triggers.de.mjs.map +0 -1
  133. package/dist/triggers.en.mjs +0 -119
  134. package/dist/triggers.en.mjs.map +0 -1
  135. package/dist/triggers.es.mjs +0 -96
  136. package/dist/triggers.es.mjs.map +0 -1
  137. package/dist/triggers.fr.mjs +0 -275
  138. package/dist/triggers.fr.mjs.map +0 -1
  139. package/dist/triggers.global.mjs +0 -79
  140. package/dist/triggers.global.mjs.map +0 -1
  141. package/dist/triggers.hu.mjs +0 -41
  142. package/dist/triggers.hu.mjs.map +0 -1
  143. package/dist/triggers.it.mjs +0 -74
  144. package/dist/triggers.it.mjs.map +0 -1
  145. package/dist/triggers.pl.mjs +0 -271
  146. package/dist/triggers.pl.mjs.map +0 -1
  147. package/dist/triggers.pt-br.mjs +0 -193
  148. package/dist/triggers.pt-br.mjs.map +0 -1
  149. package/dist/triggers.ro.mjs +0 -59
  150. package/dist/triggers.ro.mjs.map +0 -1
  151. package/dist/triggers.sk.mjs +0 -555
  152. package/dist/triggers.sk.mjs.map +0 -1
  153. package/dist/triggers.sv.mjs +0 -58
  154. package/dist/triggers.sv.mjs.map +0 -1
  155. package/dist/year-words.mjs +0 -62
  156. package/dist/year-words.mjs.map +0 -1
@@ -1,84 +0,0 @@
1
- //#region src/data/address-stopwords.json
2
- var address_stopwords_default = {
3
- _comment: "Single-token English/legal words that collide with tiny entries in city dictionaries (GeoNames-based). When the deny-list pass would emit one of these as a standalone `address` hit, it is dropped — almost certainly a Purchase Price / European Union / Effective Time false positive, not a real place. Multi-token matches (e.g. 'Lake Union, WA') are unaffected because the filter only fires on single-token spans.",
4
- _it_comment: "Italian additions are conservative: only common-noun Italian words that collide with city dictionary entries. Real Italian cities are NOT listed; multi-token spans (`Via Trieste`, `41012 Modena`) are unaffected.",
5
- words: [
6
- "agenda",
7
- "agent",
8
- "amount",
9
- "basic",
10
- "brent",
11
- "central",
12
- "centro",
13
- "client",
14
- "code",
15
- "company",
16
- "consent",
17
- "contract",
18
- "customer",
19
- "default",
20
- "delivery",
21
- "event",
22
- "exhibit",
23
- "general",
24
- "gordon",
25
- "interest",
26
- "license",
27
- "memorandum",
28
- "notice",
29
- "owner",
30
- "party",
31
- "payment",
32
- "post",
33
- "price",
34
- "principal",
35
- "rate",
36
- "rent",
37
- "registro",
38
- "schedule",
39
- "se",
40
- "section",
41
- "server",
42
- "tenant",
43
- "term",
44
- "termination",
45
- "time",
46
- "union",
47
- "urbana",
48
- "ventosa",
49
- "vendor",
50
- "waiver",
51
- "amministrazione",
52
- "appaltante",
53
- "appalto",
54
- "articolo",
55
- "cliente",
56
- "concedente",
57
- "concessionario",
58
- "conduttore",
59
- "denominazione",
60
- "esecuzione",
61
- "fornitore",
62
- "indirizzo",
63
- "informazioni",
64
- "lavoratore",
65
- "lavoro",
66
- "locatore",
67
- "luogo",
68
- "magazzino",
69
- "mandante",
70
- "mandatario",
71
- "oggetto",
72
- "piano",
73
- "procedimento",
74
- "procedura",
75
- "responsabile",
76
- "sede",
77
- "sezione",
78
- "venditore"
79
- ]
80
- };
81
- //#endregion
82
- export { address_stopwords_default as default };
83
-
84
- //# sourceMappingURL=address-stopwords.mjs.map
@@ -1 +0,0 @@
1
- {"version":3,"file":"address-stopwords.mjs","names":[],"sources":["../src/data/address-stopwords.json"],"sourcesContent":[""],"mappings":""}
@@ -1,15 +0,0 @@
1
- //#region src/data/address-unit-abbreviations.json
2
- var address_unit_abbreviations_default = {
3
- _comment: "Dotted address unit abbreviations that should not terminate address seed expansion. Organised per language because abbreviations are locale-specific.",
4
- en: [
5
- "apt.",
6
- "bldg.",
7
- "fl.",
8
- "ste.",
9
- "unit."
10
- ]
11
- };
12
- //#endregion
13
- export { address_unit_abbreviations_default as default };
14
-
15
- //# sourceMappingURL=address-unit-abbreviations.mjs.map
@@ -1 +0,0 @@
1
- {"version":3,"file":"address-unit-abbreviations.mjs","names":[],"sources":["../src/data/address-unit-abbreviations.json"],"sourcesContent":[""],"mappings":""}
@@ -1,196 +0,0 @@
1
- //#region src/data/allow-list.json
2
- var allow_list_default = {
3
- _comment: "Words that should never be flagged as PII. Includes common English/Czech words that are also surnames in the name corpus.",
4
- words: [
5
- "absence",
6
- "acceptance",
7
- "account",
8
- "action",
9
- "address",
10
- "agreement",
11
- "amu",
12
- "annual",
13
- "applicable",
14
- "balance",
15
- "base",
16
- "benefit",
17
- "bic",
18
- "board",
19
- "bond",
20
- "bonus",
21
- "branch",
22
- "business",
23
- "capital",
24
- "case",
25
- "cause",
26
- "change",
27
- "chief",
28
- "cisco",
29
- "claim",
30
- "close",
31
- "code",
32
- "comfort",
33
- "common",
34
- "company",
35
- "condition",
36
- "confidential",
37
- "consent",
38
- "counsel",
39
- "credit",
40
- "čr",
41
- "čssr",
42
- "current",
43
- "damages",
44
- "day",
45
- "default",
46
- "dell",
47
- "demand",
48
- "deposit",
49
- "determination",
50
- "disability",
51
- "discharge",
52
- "dph",
53
- "effective",
54
- "equity",
55
- "eu",
56
- "event",
57
- "evidence",
58
- "exchange",
59
- "executive",
60
- "exercise",
61
- "expense",
62
- "fair",
63
- "federal",
64
- "fee",
65
- "fees",
66
- "final",
67
- "flood",
68
- "force",
69
- "full",
70
- "fund",
71
- "gdpr",
72
- "general",
73
- "good",
74
- "grace",
75
- "grant",
76
- "gross",
77
- "holding",
78
- "house",
79
- "income",
80
- "initial",
81
- "insurance",
82
- "interest",
83
- "law",
84
- "laws",
85
- "ldn",
86
- "lease",
87
- "letter",
88
- "letters",
89
- "lien",
90
- "loan",
91
- "loss",
92
- "major",
93
- "mark",
94
- "material",
95
- "matters",
96
- "measure",
97
- "measures",
98
- "merit",
99
- "mfčr",
100
- "minor",
101
- "mobile",
102
- "moral",
103
- "mpsv",
104
- "net",
105
- "no",
106
- "not",
107
- "note",
108
- "notice",
109
- "officer",
110
- "oil",
111
- "option",
112
- "order",
113
- "otter",
114
- "pacht",
115
- "page",
116
- "parent",
117
- "party",
118
- "patent",
119
- "payment",
120
- "pension",
121
- "period",
122
- "plan",
123
- "pledge",
124
- "power",
125
- "practice",
126
- "president",
127
- "prime",
128
- "principal",
129
- "property",
130
- "purchase",
131
- "rate",
132
- "republic",
133
- "follows",
134
- "lhůta",
135
- "read",
136
- "reason",
137
- "record",
138
- "release",
139
- "rent",
140
- "report",
141
- "reserve",
142
- "rest",
143
- "review",
144
- "right",
145
- "risk",
146
- "road",
147
- "rule",
148
- "salary",
149
- "sale",
150
- "seal",
151
- "secretary",
152
- "section",
153
- "seek",
154
- "senior",
155
- "separation",
156
- "server",
157
- "service",
158
- "severance",
159
- "share",
160
- "shares",
161
- "short",
162
- "single",
163
- "sleva",
164
- "smlouva",
165
- "sole",
166
- "special",
167
- "spory",
168
- "sr",
169
- "standard",
170
- "stock",
171
- "storage",
172
- "subject",
173
- "superior",
174
- "supplement",
175
- "surrender",
176
- "tax",
177
- "title",
178
- "škola",
179
- "tato",
180
- "term",
181
- "to",
182
- "total",
183
- "transfer",
184
- "trust",
185
- "value",
186
- "vice",
187
- "vote",
188
- "votes",
189
- "warrant",
190
- "will"
191
- ]
192
- };
193
- //#endregion
194
- export { allow_list_default as default };
195
-
196
- //# sourceMappingURL=allow-list.mjs.map
@@ -1 +0,0 @@
1
- {"version":3,"file":"allow-list.mjs","names":[],"sources":["../src/data/allow-list.json"],"sourcesContent":[""],"mappings":""}
@@ -1,79 +0,0 @@
1
- //#region src/data/clause-noun-heads.json
2
- var _comment = "Common contract/clause nouns that appear in legal prose between a sentence-verb and the company name (e.g. 'Vendor signed Agreement with Acme Inc.'). The legal-form trim skips these when scanning forward for the org's first Cap word, so the anchor lands on the real company name rather than the contract noun. Lowercased; matched case-insensitively. Organised per-language; at runtime the legal-form detector unions every language's list into a single set because match spans don't carry a language tag.";
3
- var en = [
4
- "agreement",
5
- "agreements",
6
- "contract",
7
- "contracts",
8
- "license",
9
- "licence",
10
- "lease",
11
- "memorandum",
12
- "notice",
13
- "exhibit",
14
- "schedule",
15
- "addendum",
16
- "amendment",
17
- "appendix",
18
- "attachment"
19
- ];
20
- var cs = [
21
- "smlouva",
22
- "smlouvy",
23
- "smlouvu",
24
- "smlouvou",
25
- "dohoda",
26
- "dohody",
27
- "dohodu",
28
- "dohodou",
29
- "licence",
30
- "licenci",
31
- "příloha",
32
- "přílohy",
33
- "přílohu",
34
- "dodatek",
35
- "dodatku",
36
- "článek",
37
- "oznámení",
38
- "podmínky",
39
- "předmět",
40
- "ustanovení"
41
- ];
42
- var de = [
43
- "vertrag",
44
- "vertrages",
45
- "vereinbarung",
46
- "vereinbarungen",
47
- "lizenz",
48
- "anlage",
49
- "anhang"
50
- ];
51
- var es = [
52
- "acuerdo",
53
- "acuerdos",
54
- "contrato",
55
- "contratos",
56
- "convenio",
57
- "convenios",
58
- "licencia",
59
- "licencias",
60
- "memorándum",
61
- "notificación",
62
- "anexo",
63
- "anexos",
64
- "adenda",
65
- "apéndice",
66
- "cláusula",
67
- "cláusulas"
68
- ];
69
- var clause_noun_heads_default = {
70
- _comment,
71
- en,
72
- cs,
73
- de,
74
- es
75
- };
76
- //#endregion
77
- export { _comment, cs, de, clause_noun_heads_default as default, en, es };
78
-
79
- //# sourceMappingURL=clause-noun-heads.mjs.map
@@ -1 +0,0 @@
1
- {"version":3,"file":"clause-noun-heads.mjs","names":[],"sources":["../src/data/clause-noun-heads.json"],"sourcesContent":[""],"mappings":""}