@stll/anonymize 2.0.0-alpha.1 → 2.0.1

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
@@ -1,36 +0,0 @@
1
- //#region src/data/false-positive-shapes.json
2
- var _comment = "Language-keyed lexical markers used by false-positive shape guards.";
3
- var addressComponentTerms = { "cs": [
4
- "č.p.",
5
- "č.ev.",
6
- "č.",
7
- "sídliště"
8
- ] };
9
- var ambiguousStreetTypeTerms = { "fr": ["cours"] };
10
- var numberAbbrevPrefixes = {
11
- "cs": ["čís.", "č."],
12
- "de": ["nr."],
13
- "en": ["no.", "n."]
14
- };
15
- var documentHeadingOrdinalMarkers = {
16
- "cs": ["č.", "č"],
17
- "de": ["nr.", "nr"],
18
- "en": [
19
- "no.",
20
- "no",
21
- "n.",
22
- "n"
23
- ],
24
- "global": ["#"]
25
- };
26
- var false_positive_shapes_default = {
27
- _comment,
28
- addressComponentTerms,
29
- ambiguousStreetTypeTerms,
30
- numberAbbrevPrefixes,
31
- documentHeadingOrdinalMarkers
32
- };
33
- //#endregion
34
- export { _comment, addressComponentTerms, ambiguousStreetTypeTerms, false_positive_shapes_default as default, documentHeadingOrdinalMarkers, numberAbbrevPrefixes };
35
-
36
- //# sourceMappingURL=false-positive-shapes.mjs.map
@@ -1 +0,0 @@
1
- {"version":3,"file":"false-positive-shapes.mjs","names":[],"sources":["../src/data/false-positive-shapes.json"],"sourcesContent":[""],"mappings":""}
@@ -1,244 +0,0 @@
1
- //#region src/data/generic-roles.json
2
- var _comment = "Generic legal/business role terms that are not PII. Used to filter false positives from name detection.";
3
- var roles = [
4
- "acreedor",
5
- "acreedora",
6
- "adjudicatario",
7
- "adjudicataria",
8
- "arbeitgeber",
9
- "arbeitnehmer",
10
- "arrendador",
11
- "arrendadora",
12
- "arrendatario",
13
- "arrendataria",
14
- "auftraggeber",
15
- "auftragnehmer",
16
- "bonus",
17
- "borrower",
18
- "buyer",
19
- "capitán",
20
- "cedente",
21
- "cena",
22
- "cesionario",
23
- "cesionaria",
24
- "client",
25
- "cliente",
26
- "comprador",
27
- "compradora",
28
- "committee",
29
- "company",
30
- "companies",
31
- "contractor",
32
- "contratante",
33
- "contratista",
34
- "coronel",
35
- "customer",
36
- "darlehensgeber",
37
- "darlehensnehmer",
38
- "deudor",
39
- "deudora",
40
- "dlužník",
41
- "dlžník",
42
- "dodavatel",
43
- "dodávateľ",
44
- "dostawca",
45
- "dostawcą",
46
- "dostawcy",
47
- "empleado",
48
- "empleador",
49
- "empleadora",
50
- "empleada",
51
- "empresa",
52
- "empresario",
53
- "empresaria",
54
- "employee",
55
- "employer",
56
- "fax",
57
- "fiador",
58
- "fiadora",
59
- "gesellschaft",
60
- "guarantor",
61
- "kupující",
62
- "kupujúci",
63
- "käufer",
64
- "landlord",
65
- "licitador",
66
- "licitadora",
67
- "lender",
68
- "licensee",
69
- "licensor",
70
- "mieter",
71
- "nájemce",
72
- "nájomca",
73
- "objednatel",
74
- "objednávateľ",
75
- "odběratel",
76
- "odberateľ",
77
- "otorgante",
78
- "otorgado",
79
- "otorgada",
80
- "parte",
81
- "partes",
82
- "parties",
83
- "party",
84
- "plátce",
85
- "poskytovateľ",
86
- "prestador",
87
- "prestadora",
88
- "prestatario",
89
- "prestataria",
90
- "prijímateľ",
91
- "prodávající",
92
- "predávajúci",
93
- "pronajímatel",
94
- "prenajímateľ",
95
- "proveedor",
96
- "proveedora",
97
- "příjemce",
98
- "seller",
99
- "sociedad",
100
- "spoločnosť",
101
- "společnost",
102
- "strana",
103
- "strany",
104
- "supplier",
105
- "tenant",
106
- "vendedor",
107
- "vendedora",
108
- "vendor",
109
- "vermieter",
110
- "verkäufer",
111
- "věřitel",
112
- "veriteľ",
113
- "uchazeč",
114
- "ubytovateľ",
115
- "usporiadateľ",
116
- "zadavatel",
117
- "zamawiający",
118
- "zamawiającego",
119
- "zamawiającym",
120
- "zaměstnanec",
121
- "zamestnanec",
122
- "zaměstnavatel",
123
- "zamestnávateľ",
124
- "zhotovitel",
125
- "zhotoviteľ",
126
- "acquirente",
127
- "amministrazione",
128
- "appaltatore",
129
- "appaltatrice",
130
- "cedente",
131
- "cessionario",
132
- "cliente",
133
- "committente",
134
- "compratore",
135
- "compratrice",
136
- "concedente",
137
- "concessionario",
138
- "conduttore",
139
- "conduttrice",
140
- "creditore",
141
- "datore",
142
- "debitore",
143
- "denominazione",
144
- "dipendente",
145
- "fideiussore",
146
- "fornitore",
147
- "fornitrice",
148
- "garante",
149
- "lavoratore",
150
- "lavoratrice",
151
- "licenziante",
152
- "licenziatario",
153
- "locatario",
154
- "locataria",
155
- "locatore",
156
- "locatrice",
157
- "mandante",
158
- "mandatario",
159
- "mandataria",
160
- "mutuante",
161
- "mutuatario",
162
- "parte",
163
- "parti",
164
- "stazione",
165
- "subappaltatore",
166
- "venditore",
167
- "venditrice",
168
- "zleceniodawca",
169
- "zleceniodawcą",
170
- "zleceniodawcy",
171
- "zleceniobiorca",
172
- "zleceniobiorcą",
173
- "zleceniobiorcy",
174
- "wykonawca",
175
- "wykonawcą",
176
- "wykonawcy",
177
- "wynajmujący",
178
- "wynajmująca",
179
- "wynajmującym",
180
- "wynajmującej",
181
- "najemca",
182
- "najemcy",
183
- "sprzedawca",
184
- "sprzedawcy",
185
- "sprzedawcą",
186
- "sprzedający",
187
- "kupujący",
188
- "kupującego",
189
- "pracodawca",
190
- "pracodawcą",
191
- "pracownik",
192
- "pracownikiem",
193
- "pożyczkodawca",
194
- "pożyczkodawcą",
195
- "pożyczkobiorca",
196
- "pożyczkobiorcą",
197
- "odbiorca",
198
- "odbiorcą",
199
- "acheteur",
200
- "acheteuse",
201
- "acquéreur",
202
- "acquéreuse",
203
- "bailleresse",
204
- "bailleur",
205
- "cédant",
206
- "cédante",
207
- "cessionnaire",
208
- "concédant",
209
- "concédante",
210
- "concessionnaire",
211
- "donneur d'ordre",
212
- "emprunteur",
213
- "emprunteuse",
214
- "employé",
215
- "employée",
216
- "employeur",
217
- "employeuse",
218
- "fournisseur",
219
- "fournisseuse",
220
- "garant",
221
- "garante",
222
- "locataire",
223
- "mandant",
224
- "mandante",
225
- "mandataire",
226
- "preneur",
227
- "preneuse",
228
- "prestataire",
229
- "prêteur",
230
- "prêteuse",
231
- "salarié",
232
- "salariée",
233
- "société",
234
- "vendeur",
235
- "vendeuse"
236
- ];
237
- var generic_roles_default = {
238
- _comment,
239
- roles
240
- };
241
- //#endregion
242
- export { _comment, generic_roles_default as default, roles };
243
-
244
- //# sourceMappingURL=generic-roles.mjs.map
@@ -1 +0,0 @@
1
- {"version":3,"file":"generic-roles.mjs","names":[],"sources":["../src/data/generic-roles.json"],"sourcesContent":[""],"mappings":""}
@@ -1,149 +0,0 @@
1
- //#region src/data/hotword-rules.json
2
- var rules = [
3
- {
4
- "hotwords": [
5
- "rodné číslo",
6
- "r.č.",
7
- "birth number",
8
- "Geburtsnummer"
9
- ],
10
- "targetLabels": ["birth number"],
11
- "scoreAdjustment": .25,
12
- "proximityBefore": 60,
13
- "proximityAfter": 60
14
- },
15
- {
16
- "hotwords": [
17
- "narozen",
18
- "narozena",
19
- "nar.",
20
- "dat.nar.",
21
- "geboren",
22
- "born",
23
- "née"
24
- ],
25
- "targetLabels": ["date"],
26
- "scoreAdjustment": .15,
27
- "reclassifyTo": "date of birth",
28
- "proximityBefore": 60,
29
- "proximityAfter": 60
30
- },
31
- {
32
- "hotwords": [
33
- "IČO",
34
- "IČ",
35
- "identifikační číslo",
36
- "reg. č.",
37
- "Handelsregister",
38
- "company number"
39
- ],
40
- "targetLabels": ["registration number"],
41
- "scoreAdjustment": .2,
42
- "proximityBefore": 60,
43
- "proximityAfter": 60
44
- },
45
- {
46
- "hotwords": [
47
- "DIČ",
48
- "daňové identifikační číslo",
49
- "tax ID",
50
- "Steuernummer",
51
- "VAT"
52
- ],
53
- "targetLabels": ["tax identification number"],
54
- "scoreAdjustment": .2,
55
- "proximityBefore": 60,
56
- "proximityAfter": 60
57
- },
58
- {
59
- "hotwords": [
60
- "tel.",
61
- "telefon",
62
- "phone",
63
- "mobil",
64
- "fax",
65
- "Telefon",
66
- "Handy"
67
- ],
68
- "targetLabels": ["phone number"],
69
- "scoreAdjustment": .15,
70
- "proximityBefore": 60,
71
- "proximityAfter": 60
72
- },
73
- {
74
- "hotwords": [
75
- "č.ú.",
76
- "číslo účtu",
77
- "bank account",
78
- "IBAN",
79
- "bankovní spojení"
80
- ],
81
- "targetLabels": ["bank account number", "iban"],
82
- "scoreAdjustment": .15,
83
- "proximityBefore": 60,
84
- "proximityAfter": 60
85
- },
86
- {
87
- "hotwords": [
88
- "e-mail",
89
- "email",
90
- "E-Mail",
91
- "pec"
92
- ],
93
- "targetLabels": ["email address"],
94
- "scoreAdjustment": .1,
95
- "proximityBefore": 60,
96
- "proximityAfter": 60
97
- },
98
- {
99
- "hotwords": [
100
- "adresa",
101
- "sídlo",
102
- "sídlem",
103
- "bytem",
104
- "Adresse",
105
- "address",
106
- "indirizzo",
107
- "sede legale",
108
- "sede"
109
- ],
110
- "targetLabels": ["address"],
111
- "scoreAdjustment": .15,
112
- "proximityBefore": 80,
113
- "proximityAfter": 150
114
- },
115
- {
116
- "hotwords": [
117
- "jednatel",
118
- "ředitel",
119
- "předseda",
120
- "prokurista",
121
- "Geschäftsführer",
122
- "director"
123
- ],
124
- "targetLabels": ["person"],
125
- "scoreAdjustment": .1,
126
- "proximityBefore": 200,
127
- "proximityAfter": 30
128
- },
129
- {
130
- "hotwords": [
131
- "článek",
132
- "článku",
133
- "paragraf",
134
- "section",
135
- "article",
136
- "bod",
137
- "odstavec"
138
- ],
139
- "targetLabels": ["date", "address"],
140
- "scoreAdjustment": -.3,
141
- "proximityBefore": 30,
142
- "proximityAfter": 10
143
- }
144
- ];
145
- var hotword_rules_default = { rules };
146
- //#endregion
147
- export { hotword_rules_default as default, rules };
148
-
149
- //# sourceMappingURL=hotword-rules.mjs.map
@@ -1 +0,0 @@
1
- {"version":3,"file":"hotword-rules.mjs","names":[],"sources":["../src/data/hotword-rules.json"],"sourcesContent":[""],"mappings":""}
@@ -1,23 +0,0 @@
1
- //#region src/data/legal-form-leading-clauses.json
2
- var _comment = "Per-language clause text that may precede an organization name before a legal-form suffix. `phrases` are trimmed unconditionally when found inside an over-captured legal-form span. `directPrefixes` are trimmed only when the text before the prefix already looks like prose, so company names such as `Stand By Me LLC` are preserved.";
3
- var en = {
4
- "phrases": [
5
- "by and among",
6
- "by and between",
7
- "is between"
8
- ],
9
- "directPrefixes": [
10
- "by",
11
- "among",
12
- "amongst",
13
- "between"
14
- ]
15
- };
16
- var legal_form_leading_clauses_default = {
17
- _comment,
18
- en
19
- };
20
- //#endregion
21
- export { _comment, legal_form_leading_clauses_default as default, en };
22
-
23
- //# sourceMappingURL=legal-form-leading-clauses.mjs.map
@@ -1 +0,0 @@
1
- {"version":3,"file":"legal-form-leading-clauses.mjs","names":[],"sources":["../src/data/legal-form-leading-clauses.json"],"sourcesContent":[""],"mappings":""}