@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
@@ -1,151 +0,0 @@
1
- //#region src/data/names-nw-vi.json
2
- var _comment = "Non-Western name tokens for vi locale";
3
- var names = [
4
- "Anh",
5
- "Bao",
6
- "Bich",
7
- "Bind",
8
- "Binh",
9
- "Bui",
10
- "Cao",
11
- "Chanh",
12
- "Chau",
13
- "Chien",
14
- "Chon",
15
- "Chuan",
16
- "Cuc",
17
- "Cuong",
18
- "Dang",
19
- "Dao",
20
- "Dat",
21
- "Dau",
22
- "Diep",
23
- "Dinh",
24
- "Doanh",
25
- "Dok",
26
- "Duc",
27
- "Dung",
28
- "Duong",
29
- "Duyen",
30
- "Giang",
31
- "Giao",
32
- "Hai",
33
- "Han",
34
- "Hang",
35
- "Hanh",
36
- "Hao",
37
- "Hien",
38
- "Hieu",
39
- "Hoa",
40
- "Hoang",
41
- "Hong",
42
- "Huong",
43
- "Huu",
44
- "Huy",
45
- "Huynh",
46
- "Khang",
47
- "Khanh",
48
- "Khoa",
49
- "Khuyen",
50
- "Kie",
51
- "Kien",
52
- "Kieu",
53
- "Kim",
54
- "Kiu",
55
- "La",
56
- "Lam",
57
- "Lan",
58
- "Lao",
59
- "Le",
60
- "Lien",
61
- "Linh",
62
- "Loan",
63
- "Loc",
64
- "Long",
65
- "Luong",
66
- "Luu",
67
- "Ly",
68
- "Mai",
69
- "Manh",
70
- "Minh",
71
- "My",
72
- "Nam",
73
- "Nga",
74
- "Nghia",
75
- "Nghiem",
76
- "Ngo",
77
- "Ngoc",
78
- "Nguyen",
79
- "Nhan",
80
- "Nhat",
81
- "Nhi",
82
- "Nhu",
83
- "Nhung",
84
- "Ninh",
85
- "Nong",
86
- "Oanh",
87
- "Pham",
88
- "Phan",
89
- "Phat",
90
- "Phong",
91
- "Phu",
92
- "Phuc",
93
- "Phung",
94
- "Phuong",
95
- "Quach",
96
- "Quan",
97
- "Quang",
98
- "Quy",
99
- "Quyet",
100
- "Quynh",
101
- "Sang",
102
- "Sen",
103
- "Sy",
104
- "Ta",
105
- "Tan",
106
- "Thanh",
107
- "Thao",
108
- "Thi",
109
- "Thien",
110
- "Thiet",
111
- "Thip",
112
- "Thu",
113
- "Thuan",
114
- "Thuc",
115
- "Thuy",
116
- "Tien",
117
- "Tram",
118
- "Tran",
119
- "Trang",
120
- "Trieu",
121
- "Trinh",
122
- "Trong",
123
- "Truc",
124
- "Trung",
125
- "Truong",
126
- "Tuan",
127
- "Tuong",
128
- "Tuyet",
129
- "Ty",
130
- "Ung",
131
- "Uy",
132
- "Uyen",
133
- "Vien",
134
- "Viet",
135
- "Vinh",
136
- "Vo",
137
- "Vu",
138
- "Vuong",
139
- "Vy",
140
- "Xuan",
141
- "Yat",
142
- "Yen"
143
- ];
144
- var names_nw_vi_default = {
145
- _comment,
146
- names
147
- };
148
- //#endregion
149
- export { _comment, names_nw_vi_default as default, names };
150
-
151
- //# sourceMappingURL=names-nw-vi.mjs.map
@@ -1 +0,0 @@
1
- {"version":3,"file":"names-nw-vi.mjs","names":[],"sources":["../src/data/names-nw-vi.json"],"sourcesContent":[""],"mappings":""}
@@ -1,197 +0,0 @@
1
- //#region src/data/names-nw-zh-latn.json
2
- var _comment = "Non-Western name tokens for zh-latn locale";
3
- var names = [
4
- "Ang",
5
- "AuYeung",
6
- "Cao",
7
- "Chak",
8
- "Cham",
9
- "Chan",
10
- "Chang",
11
- "Chao",
12
- "Chau",
13
- "Cheak",
14
- "Chee",
15
- "Chek",
16
- "Chen",
17
- "Cheong",
18
- "Cheung",
19
- "Chia",
20
- "Chien",
21
- "Chik",
22
- "Chin",
23
- "Ching",
24
- "Chiu",
25
- "Chng",
26
- "Chon",
27
- "Chong",
28
- "Chow",
29
- "Choy",
30
- "Chu",
31
- "Chua",
32
- "Chuan",
33
- "Chui",
34
- "Chun",
35
- "Chung",
36
- "Cu",
37
- "Cua",
38
- "Dai",
39
- "Dong",
40
- "Dy",
41
- "Eah",
42
- "Fong",
43
- "Foo",
44
- "Fuk",
45
- "Fung",
46
- "Goh",
47
- "Guan",
48
- "Hak",
49
- "Han",
50
- "Hang",
51
- "Hao",
52
- "Heng",
53
- "Heung",
54
- "Ho",
55
- "Hock",
56
- "Hoi",
57
- "Hong",
58
- "Hsu",
59
- "Huang",
60
- "Huat",
61
- "Hui",
62
- "Hung",
63
- "Jia",
64
- "Jiang",
65
- "Jun",
66
- "Kam",
67
- "Kan",
68
- "Kee",
69
- "Kek",
70
- "Keung",
71
- "Khoo",
72
- "Kian",
73
- "Kiat",
74
- "Kie",
75
- "Kieu",
76
- "Kiok",
77
- "Kiu",
78
- "Ko",
79
- "Koh",
80
- "Kong",
81
- "Koo",
82
- "Koon",
83
- "Kowk",
84
- "Kuk",
85
- "Kwan",
86
- "Kwok",
87
- "Kwong",
88
- "La",
89
- "Lai",
90
- "Lam",
91
- "Lau",
92
- "Lee",
93
- "Leong",
94
- "Leung",
95
- "Li",
96
- "Lim",
97
- "Lin",
98
- "Ling",
99
- "Liu",
100
- "Lo",
101
- "Lok",
102
- "Lone",
103
- "Long",
104
- "Loo",
105
- "Luk",
106
- "Lun",
107
- "Ma",
108
- "Mak",
109
- "Mei",
110
- "Ming",
111
- "Mok",
112
- "Mong",
113
- "Nai",
114
- "Ng",
115
- "Ngai",
116
- "Ooi",
117
- "Oon",
118
- "Pai",
119
- "Pak",
120
- "Pang",
121
- "Peh",
122
- "Phua",
123
- "Ping",
124
- "Poe",
125
- "Poon",
126
- "Pua",
127
- "Puah",
128
- "Quek",
129
- "Ren",
130
- "Sen",
131
- "Seng",
132
- "Seow",
133
- "Sham",
134
- "Shek",
135
- "Shen",
136
- "Shum",
137
- "Situ",
138
- "Siu",
139
- "Soh",
140
- "Song",
141
- "Soon",
142
- "Suen",
143
- "Suk",
144
- "Sy",
145
- "Szeto",
146
- "Tai",
147
- "Tak",
148
- "Tam",
149
- "Tan",
150
- "Tang",
151
- "Tay",
152
- "Tee",
153
- "Teo",
154
- "Tham",
155
- "Thom",
156
- "Tiah",
157
- "Ting",
158
- "Tiu",
159
- "Tong",
160
- "Tsai",
161
- "Tse",
162
- "Tsen",
163
- "Tsui",
164
- "Tung",
165
- "Wai",
166
- "Wan",
167
- "Wang",
168
- "Wee",
169
- "Wing",
170
- "Wong",
171
- "Woo",
172
- "Wu",
173
- "Wun",
174
- "Xie",
175
- "Yam",
176
- "Yang",
177
- "Yap",
178
- "Yeoh",
179
- "Yeung",
180
- "Yi",
181
- "Yim",
182
- "Ying",
183
- "Yip",
184
- "Yiu",
185
- "Yuen",
186
- "Yuk",
187
- "Yum",
188
- "Zhang"
189
- ];
190
- var names_nw_zh_latn_default = {
191
- _comment,
192
- names
193
- };
194
- //#endregion
195
- export { _comment, names_nw_zh_latn_default as default, names };
196
-
197
- //# sourceMappingURL=names-nw-zh-latn.mjs.map
@@ -1 +0,0 @@
1
- {"version":3,"file":"names-nw-zh-latn.mjs","names":[],"sources":["../src/data/names-nw-zh-latn.json"],"sourcesContent":[""],"mappings":""}
@@ -1,113 +0,0 @@
1
- //#region src/data/names-surnames.json
2
- var _comment = "Common surnames for person detection. Czech, Slovak, German, Austrian sources.";
3
- var names = [
4
- "Bartoš",
5
- "Bauer",
6
- "Becker",
7
- "Beneš",
8
- "Berger",
9
- "Blaha",
10
- "Braun",
11
- "Brunner",
12
- "Čermák",
13
- "Černá",
14
- "Černý",
15
- "Doležal",
16
- "Duda",
17
- "Dvořák",
18
- "Dvořáková",
19
- "Fiala",
20
- "Fischer",
21
- "Fuchs",
22
- "Gruber",
23
- "Hájek",
24
- "Hartmann",
25
- "Hermann",
26
- "Hofer",
27
- "Hoffmann",
28
- "Holub",
29
- "Horák",
30
- "Horáková",
31
- "Hruška",
32
- "Huber",
33
- "Jelínek",
34
- "Kadlec",
35
- "Kaiser",
36
- "Koch",
37
- "Köhler",
38
- "Kolář",
39
- "Koller",
40
- "Konečný",
41
- "Kopecký",
42
- "Král",
43
- "Kratochvíl",
44
- "Krejčí",
45
- "Kříž",
46
- "Krüger",
47
- "Kubík",
48
- "Kučera",
49
- "Kučerová",
50
- "Lang",
51
- "Lehmann",
52
- "Maier",
53
- "Malý",
54
- "Mareš",
55
- "Mašek",
56
- "Mayer",
57
- "Meyer",
58
- "Moser",
59
- "Müller",
60
- "Musil",
61
- "Navrátil",
62
- "Němcová",
63
- "Němec",
64
- "Neumann",
65
- "Nguyen",
66
- "Novák",
67
- "Nováková",
68
- "Novotná",
69
- "Novotný",
70
- "Pavlík",
71
- "Pichler",
72
- "Pokorný",
73
- "Polák",
74
- "Pospíšil",
75
- "Procházka",
76
- "Procházková",
77
- "Richter",
78
- "Růžička",
79
- "Schmid",
80
- "Schmidt",
81
- "Schmitt",
82
- "Schneider",
83
- "Schröder",
84
- "Schulz",
85
- "Schwarz",
86
- "Sedláček",
87
- "Šimek",
88
- "Šťastný",
89
- "Steiner",
90
- "Svoboda",
91
- "Svobodová",
92
- "Tichý",
93
- "Urban",
94
- "Vaněk",
95
- "Veselá",
96
- "Veselý",
97
- "Vlček",
98
- "Wagner",
99
- "Weber",
100
- "Werner",
101
- "Winkler",
102
- "Wolf",
103
- "Zeman",
104
- "Zimmermann"
105
- ];
106
- var names_surnames_default = {
107
- _comment,
108
- names
109
- };
110
- //#endregion
111
- export { _comment, names_surnames_default as default, names };
112
-
113
- //# sourceMappingURL=names-surnames.mjs.map
@@ -1 +0,0 @@
1
- {"version":3,"file":"names-surnames.mjs","names":[],"sources":["../src/data/names-surnames.json"],"sourcesContent":[""],"mappings":""}
@@ -1,40 +0,0 @@
1
- //#region src/data/names-title-tokens.json
2
- var _comment = "Lowercase title tokens that signal a person name follows. Used for token classification in the name chain detector.";
3
- var tokens = [
4
- "bc",
5
- "dame",
6
- "doc",
7
- "dr",
8
- "frau",
9
- "herr",
10
- "ing",
11
- "judge",
12
- "judr",
13
- "justice",
14
- "lady",
15
- "lord",
16
- "maître",
17
- "mgr",
18
- "miss",
19
- "mlle",
20
- "mme",
21
- "mr",
22
- "mrs",
23
- "ms",
24
- "mudr",
25
- "mvdr",
26
- "paeddr",
27
- "phdr",
28
- "prof",
29
- "rndr",
30
- "sir",
31
- "thdr"
32
- ];
33
- var names_title_tokens_default = {
34
- _comment,
35
- tokens
36
- };
37
- //#endregion
38
- export { _comment, names_title_tokens_default as default, tokens };
39
-
40
- //# sourceMappingURL=names-title-tokens.mjs.map
@@ -1 +0,0 @@
1
- {"version":3,"file":"names-title-tokens.mjs","names":[],"sources":["../src/data/names-title-tokens.json"],"sourcesContent":[""],"mappings":""}
@@ -1,20 +0,0 @@
1
- //#region src/data/organization-unit-heads.json
2
- var _comment = "Administrative or organizational unit nouns that can appear in legal prose without denoting a person or a street/city suffix. Lowercased and organized per language.";
3
- var cs = [
4
- "agentura",
5
- "inspekce",
6
- "kancelář",
7
- "odbor",
8
- "oddělení",
9
- "sekretariát",
10
- "správa",
11
- "úřad"
12
- ];
13
- var organization_unit_heads_default = {
14
- _comment,
15
- cs
16
- };
17
- //#endregion
18
- export { _comment, cs, organization_unit_heads_default as default };
19
-
20
- //# sourceMappingURL=organization-unit-heads.mjs.map
@@ -1 +0,0 @@
1
- {"version":3,"file":"organization-unit-heads.mjs","names":[],"sources":["../src/data/organization-unit-heads.json"],"sourcesContent":[""],"mappings":""}