@stll/anonymize-wasm 1.5.0 → 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.
- package/LICENSE +1 -2
- package/README.md +93 -4
- package/dist/native/index.wasi-browser.js +12002 -0
- package/dist/native/index.wasi.cjs +122 -0
- package/dist/native/index.wasm32-wasi.wasm +0 -0
- package/dist/native/native-pipeline.cs.stlanonpkg +0 -0
- package/dist/native/native-pipeline.de.stlanonpkg +0 -0
- package/dist/native/native-pipeline.en.stlanonpkg +0 -0
- package/dist/native/native-pipeline.stlanonpkg +0 -0
- package/dist/native/wasi-worker-browser.mjs +10766 -0
- package/dist/native/wasi-worker.mjs +63 -0
- package/dist/vite.d.mts +18 -2
- package/dist/vite.mjs +89 -12
- package/dist/vite.mjs.map +1 -1
- package/dist/wasm.d.mts +553 -1102
- package/dist/wasm.mjs +359 -16223
- package/dist/wasm.mjs.map +1 -1
- package/package.json +15 -9
- package/dist/address-boundaries.mjs +0 -195
- package/dist/address-boundaries.mjs.map +0 -1
- package/dist/address-prepositions.mjs +0 -182
- package/dist/address-prepositions.mjs.map +0 -1
- package/dist/address-stop-keywords.mjs +0 -137
- package/dist/address-stop-keywords.mjs.map +0 -1
- package/dist/address-stopwords.mjs +0 -84
- package/dist/address-stopwords.mjs.map +0 -1
- package/dist/allow-list.mjs +0 -196
- package/dist/allow-list.mjs.map +0 -1
- package/dist/clause-noun-heads.mjs +0 -75
- package/dist/clause-noun-heads.mjs.map +0 -1
- package/dist/common-words-en.mjs +0 -9887
- package/dist/common-words-en.mjs.map +0 -1
- package/dist/coreference.cs.mjs +0 -14
- package/dist/coreference.cs.mjs.map +0 -1
- package/dist/coreference.de.mjs +0 -14
- package/dist/coreference.de.mjs.map +0 -1
- package/dist/coreference.en.mjs +0 -14
- package/dist/coreference.en.mjs.map +0 -1
- package/dist/coreference.es.mjs +0 -22
- package/dist/coreference.es.mjs.map +0 -1
- package/dist/coreference.fr.mjs +0 -32
- package/dist/coreference.fr.mjs.map +0 -1
- package/dist/coreference.it.mjs +0 -27
- package/dist/coreference.it.mjs.map +0 -1
- package/dist/coreference.pl.mjs +0 -27
- package/dist/coreference.pl.mjs.map +0 -1
- package/dist/coreference.pt-br.mjs +0 -14
- package/dist/coreference.pt-br.mjs.map +0 -1
- package/dist/coreference.sk.mjs +0 -27
- package/dist/coreference.sk.mjs.map +0 -1
- package/dist/currencies.mjs +0 -231
- package/dist/currencies.mjs.map +0 -1
- package/dist/date-months.mjs +0 -618
- package/dist/date-months.mjs.map +0 -1
- package/dist/document-structure-headings.mjs +0 -90
- package/dist/document-structure-headings.mjs.map +0 -1
- package/dist/generic-roles.mjs +0 -244
- package/dist/generic-roles.mjs.map +0 -1
- package/dist/hotword-rules.mjs +0 -149
- package/dist/hotword-rules.mjs.map +0 -1
- package/dist/legal-form-leading-clauses.mjs +0 -23
- package/dist/legal-form-leading-clauses.mjs.map +0 -1
- package/dist/legal-forms.mjs +0 -2115
- package/dist/legal-forms.mjs.map +0 -1
- package/dist/legal-role-heads.cs.mjs +0 -42
- package/dist/legal-role-heads.cs.mjs.map +0 -1
- package/dist/legal-role-heads.de.mjs +0 -33
- package/dist/legal-role-heads.de.mjs.map +0 -1
- package/dist/legal-role-heads.en.mjs +0 -37
- package/dist/legal-role-heads.en.mjs.map +0 -1
- package/dist/legal-role-heads.es.mjs +0 -54
- package/dist/legal-role-heads.es.mjs.map +0 -1
- package/dist/legal-role-heads.fr.mjs +0 -72
- package/dist/legal-role-heads.fr.mjs.map +0 -1
- package/dist/legal-role-heads.it.mjs +0 -68
- package/dist/legal-role-heads.it.mjs.map +0 -1
- package/dist/legal-role-heads.pl.mjs +0 -84
- package/dist/legal-role-heads.pl.mjs.map +0 -1
- package/dist/legal-role-heads.pt-br.mjs +0 -63
- package/dist/legal-role-heads.pt-br.mjs.map +0 -1
- package/dist/legal-role-heads.sk.mjs +0 -80
- package/dist/legal-role-heads.sk.mjs.map +0 -1
- package/dist/manifest.mjs +0 -69
- package/dist/manifest.mjs.map +0 -1
- package/dist/names-exclusions.mjs +0 -223
- package/dist/names-exclusions.mjs.map +0 -1
- package/dist/names-first.mjs +0 -418
- package/dist/names-first.mjs.map +0 -1
- package/dist/names-nw-ar.mjs +0 -202
- package/dist/names-nw-ar.mjs.map +0 -1
- package/dist/names-nw-excluded-allcaps.mjs +0 -112
- package/dist/names-nw-excluded-allcaps.mjs.map +0 -1
- package/dist/names-nw-fil.mjs +0 -202
- package/dist/names-nw-fil.mjs.map +0 -1
- package/dist/names-nw-id.mjs +0 -210
- package/dist/names-nw-id.mjs.map +0 -1
- package/dist/names-nw-in.mjs +0 -526
- package/dist/names-nw-in.mjs.map +0 -1
- package/dist/names-nw-ja-latn.mjs +0 -260
- package/dist/names-nw-ja-latn.mjs.map +0 -1
- package/dist/names-nw-ko.mjs +0 -162
- package/dist/names-nw-ko.mjs.map +0 -1
- package/dist/names-nw-th.mjs +0 -188
- package/dist/names-nw-th.mjs.map +0 -1
- package/dist/names-nw-vi.mjs +0 -151
- package/dist/names-nw-vi.mjs.map +0 -1
- package/dist/names-nw-zh-latn.mjs +0 -197
- package/dist/names-nw-zh-latn.mjs.map +0 -1
- package/dist/names-surnames.mjs +0 -113
- package/dist/names-surnames.mjs.map +0 -1
- package/dist/names-title-tokens.mjs +0 -40
- package/dist/names-title-tokens.mjs.map +0 -1
- package/dist/person-stopwords.mjs +0 -205
- package/dist/person-stopwords.mjs.map +0 -1
- package/dist/section-headings.mjs +0 -64
- package/dist/section-headings.mjs.map +0 -1
- package/dist/sentence-verb-indicators.mjs +0 -232
- package/dist/sentence-verb-indicators.mjs.map +0 -1
- package/dist/signing-clauses.mjs +0 -78
- package/dist/signing-clauses.mjs.map +0 -1
- package/dist/stopwords.mjs +0 -9915
- package/dist/stopwords.mjs.map +0 -1
- package/dist/structural-single-cap-prefixes.mjs +0 -99
- package/dist/structural-single-cap-prefixes.mjs.map +0 -1
- package/dist/triggers.cs.mjs +0 -569
- package/dist/triggers.cs.mjs.map +0 -1
- package/dist/triggers.de.mjs +0 -139
- package/dist/triggers.de.mjs.map +0 -1
- package/dist/triggers.en.mjs +0 -119
- package/dist/triggers.en.mjs.map +0 -1
- package/dist/triggers.es.mjs +0 -96
- package/dist/triggers.es.mjs.map +0 -1
- package/dist/triggers.fr.mjs +0 -275
- package/dist/triggers.fr.mjs.map +0 -1
- package/dist/triggers.global.mjs +0 -79
- package/dist/triggers.global.mjs.map +0 -1
- package/dist/triggers.hu.mjs +0 -41
- package/dist/triggers.hu.mjs.map +0 -1
- package/dist/triggers.it.mjs +0 -74
- package/dist/triggers.it.mjs.map +0 -1
- package/dist/triggers.pl.mjs +0 -271
- package/dist/triggers.pl.mjs.map +0 -1
- package/dist/triggers.pt-br.mjs +0 -193
- package/dist/triggers.pt-br.mjs.map +0 -1
- package/dist/triggers.ro.mjs +0 -59
- package/dist/triggers.ro.mjs.map +0 -1
- package/dist/triggers.sk.mjs +0 -555
- package/dist/triggers.sk.mjs.map +0 -1
- package/dist/triggers.sv.mjs +0 -58
- package/dist/triggers.sv.mjs.map +0 -1
- package/dist/year-words.mjs +0 -62
- package/dist/year-words.mjs.map +0 -1
package/dist/allow-list.mjs
DELETED
|
@@ -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
|
package/dist/allow-list.mjs.map
DELETED
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"allow-list.mjs","names":[],"sources":["../../src/data/allow-list.json"],"sourcesContent":[""],"mappings":""}
|
|
@@ -1,75 +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
|
-
"oznámení"
|
|
37
|
-
];
|
|
38
|
-
var de = [
|
|
39
|
-
"vertrag",
|
|
40
|
-
"vertrages",
|
|
41
|
-
"vereinbarung",
|
|
42
|
-
"vereinbarungen",
|
|
43
|
-
"lizenz",
|
|
44
|
-
"anlage",
|
|
45
|
-
"anhang"
|
|
46
|
-
];
|
|
47
|
-
var es = [
|
|
48
|
-
"acuerdo",
|
|
49
|
-
"acuerdos",
|
|
50
|
-
"contrato",
|
|
51
|
-
"contratos",
|
|
52
|
-
"convenio",
|
|
53
|
-
"convenios",
|
|
54
|
-
"licencia",
|
|
55
|
-
"licencias",
|
|
56
|
-
"memorándum",
|
|
57
|
-
"notificación",
|
|
58
|
-
"anexo",
|
|
59
|
-
"anexos",
|
|
60
|
-
"adenda",
|
|
61
|
-
"apéndice",
|
|
62
|
-
"cláusula",
|
|
63
|
-
"cláusulas"
|
|
64
|
-
];
|
|
65
|
-
var clause_noun_heads_default = {
|
|
66
|
-
_comment,
|
|
67
|
-
en,
|
|
68
|
-
cs,
|
|
69
|
-
de,
|
|
70
|
-
es
|
|
71
|
-
};
|
|
72
|
-
//#endregion
|
|
73
|
-
export { _comment, cs, de, clause_noun_heads_default as default, en, es };
|
|
74
|
-
|
|
75
|
-
//# 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":""}
|