@stll/anonymize-wasm 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.
- 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 +181 -1356
- package/dist/wasm.mjs +185 -18644
- package/dist/wasm.mjs.map +1 -1
- package/package.json +15 -9
- package/dist/address-boundaries.mjs +0 -197
- package/dist/address-boundaries.mjs.map +0 -1
- package/dist/address-jurisdiction-prefixes.mjs +0 -16
- package/dist/address-jurisdiction-prefixes.mjs.map +0 -1
- package/dist/address-stop-keywords.mjs +0 -148
- 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/address-unit-abbreviations.mjs +0 -15
- package/dist/address-unit-abbreviations.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 -79
- 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-org-determiners.mjs +0 -19
- package/dist/coreference-org-determiners.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/defined-term-heads.mjs +0 -15
- package/dist/defined-term-heads.mjs.map +0 -1
- package/dist/document-structure-headings.mjs +0 -90
- package/dist/document-structure-headings.mjs.map +0 -1
- package/dist/false-positive-shapes.mjs +0 -36
- package/dist/false-positive-shapes.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 -48
- 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/organization-unit-heads.mjs +0 -20
- package/dist/organization-unit-heads.mjs.map +0 -1
- package/dist/person-stopwords.mjs +0 -211
- 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 -102
- 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/package.json
CHANGED
|
@@ -1,7 +1,18 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@stll/anonymize-wasm",
|
|
3
|
-
"version": "2.0.0
|
|
4
|
-
"description": "PII detection and anonymization for browsers via WebAssembly. Same API as @stll/anonymize.",
|
|
3
|
+
"version": "2.0.0",
|
|
4
|
+
"description": "PII detection and anonymization for browsers via WebAssembly. Same native SDK API as @stll/anonymize.",
|
|
5
|
+
"keywords": [
|
|
6
|
+
"anonymization",
|
|
7
|
+
"pii",
|
|
8
|
+
"redaction",
|
|
9
|
+
"ner",
|
|
10
|
+
"pseudonymization",
|
|
11
|
+
"gdpr",
|
|
12
|
+
"privacy",
|
|
13
|
+
"wasm",
|
|
14
|
+
"browser"
|
|
15
|
+
],
|
|
5
16
|
"type": "module",
|
|
6
17
|
"module": "dist/wasm.mjs",
|
|
7
18
|
"types": "dist/wasm.d.mts",
|
|
@@ -30,20 +41,15 @@
|
|
|
30
41
|
"type": "git",
|
|
31
42
|
"url": "git+https://github.com/stella/anonymize.git"
|
|
32
43
|
},
|
|
44
|
+
"homepage": "https://github.com/stella/anonymize",
|
|
33
45
|
"license": "MIT",
|
|
34
46
|
"dependencies": {
|
|
35
|
-
"@
|
|
36
|
-
"@stll/stdnum": "^2.1.1",
|
|
37
|
-
"@stll/text-search-wasm": "^1.0.7"
|
|
47
|
+
"@napi-rs/wasm-runtime": "^1.1.6"
|
|
38
48
|
},
|
|
39
49
|
"peerDependencies": {
|
|
40
|
-
"@stll/anonymize-data": "^0.0.6",
|
|
41
50
|
"vite": ">=5 <10"
|
|
42
51
|
},
|
|
43
52
|
"peerDependenciesMeta": {
|
|
44
|
-
"@stll/anonymize-data": {
|
|
45
|
-
"optional": true
|
|
46
|
-
},
|
|
47
53
|
"vite": {
|
|
48
54
|
"optional": true
|
|
49
55
|
}
|
|
@@ -1,197 +0,0 @@
|
|
|
1
|
-
//#region src/data/address-boundaries.json
|
|
2
|
-
var address_boundaries_default = {
|
|
3
|
-
_comment: "Boundary words that signal the end of an address span. Multi-language. Used by address seed detector to stop expansion.",
|
|
4
|
-
cs: [
|
|
5
|
-
"jednající",
|
|
6
|
-
"jednajícím",
|
|
7
|
-
"jejímž jménem",
|
|
8
|
-
"kontaktní osoba",
|
|
9
|
-
"pokud",
|
|
10
|
-
"zapsán",
|
|
11
|
-
"zapsaná",
|
|
12
|
-
"zapsané",
|
|
13
|
-
"zapsaný",
|
|
14
|
-
"zastoupená",
|
|
15
|
-
"zastoupené",
|
|
16
|
-
"zastoupený"
|
|
17
|
-
],
|
|
18
|
-
de: [
|
|
19
|
-
"amtsgericht",
|
|
20
|
-
"eingetragen",
|
|
21
|
-
"geschäftsführer",
|
|
22
|
-
"handelsregister",
|
|
23
|
-
"vertreten"
|
|
24
|
-
],
|
|
25
|
-
en: [
|
|
26
|
-
"attention",
|
|
27
|
-
"attn",
|
|
28
|
-
"c/o",
|
|
29
|
-
"care of",
|
|
30
|
-
"concerning",
|
|
31
|
-
"contact person",
|
|
32
|
-
"for any",
|
|
33
|
-
"incorporated",
|
|
34
|
-
"postmarked",
|
|
35
|
-
"registered",
|
|
36
|
-
"represented",
|
|
37
|
-
"shall govern",
|
|
38
|
-
"shall be governed",
|
|
39
|
-
"to be enforced",
|
|
40
|
-
"with a copy",
|
|
41
|
-
"with the intention",
|
|
42
|
-
"without reference",
|
|
43
|
-
"without regard"
|
|
44
|
-
],
|
|
45
|
-
es: [
|
|
46
|
-
"con CIF",
|
|
47
|
-
"con C.I.F.",
|
|
48
|
-
"con NIF",
|
|
49
|
-
"con N.I.F.",
|
|
50
|
-
"con DNI",
|
|
51
|
-
"con D.N.I.",
|
|
52
|
-
"con NIE",
|
|
53
|
-
"con N.I.E.",
|
|
54
|
-
"inscrita",
|
|
55
|
-
"inscrito",
|
|
56
|
-
"registro mercantil",
|
|
57
|
-
"registro",
|
|
58
|
-
"representada por",
|
|
59
|
-
"representado por",
|
|
60
|
-
"actuando en nombre",
|
|
61
|
-
"número de identificación fiscal",
|
|
62
|
-
"código postal"
|
|
63
|
-
],
|
|
64
|
-
fr: [
|
|
65
|
-
"agissant",
|
|
66
|
-
"ayant pour",
|
|
67
|
-
"ayant son",
|
|
68
|
-
"ci-après",
|
|
69
|
-
"ci-dessous",
|
|
70
|
-
"ci-dessus",
|
|
71
|
-
"désigné",
|
|
72
|
-
"désignée",
|
|
73
|
-
"dénommé",
|
|
74
|
-
"dénommée",
|
|
75
|
-
"immatriculé",
|
|
76
|
-
"immatriculée",
|
|
77
|
-
"inscrit",
|
|
78
|
-
"inscrite",
|
|
79
|
-
"RCS",
|
|
80
|
-
"représenté",
|
|
81
|
-
"représentée",
|
|
82
|
-
"SIREN",
|
|
83
|
-
"SIRET",
|
|
84
|
-
"TVA"
|
|
85
|
-
],
|
|
86
|
-
hu: [
|
|
87
|
-
"bejegyzett",
|
|
88
|
-
"cégjegyzékszám",
|
|
89
|
-
"képviseli"
|
|
90
|
-
],
|
|
91
|
-
it: [
|
|
92
|
-
"C.F.",
|
|
93
|
-
"P. IVA",
|
|
94
|
-
"P.IVA",
|
|
95
|
-
"codice fiscale",
|
|
96
|
-
"iscritta",
|
|
97
|
-
"iscritto",
|
|
98
|
-
"partita IVA",
|
|
99
|
-
"punti di contatto",
|
|
100
|
-
"rappresentata",
|
|
101
|
-
"rappresentato",
|
|
102
|
-
"registro imprese",
|
|
103
|
-
"responsabile del procedimento"
|
|
104
|
-
],
|
|
105
|
-
pl: [
|
|
106
|
-
"kapitał zakładowy",
|
|
107
|
-
"kapitale zakładowym",
|
|
108
|
-
"KRS",
|
|
109
|
-
"NIP",
|
|
110
|
-
"REGON",
|
|
111
|
-
"PESEL",
|
|
112
|
-
"reprezentowana",
|
|
113
|
-
"reprezentowany",
|
|
114
|
-
"reprezentowaną",
|
|
115
|
-
"reprezentowanego",
|
|
116
|
-
"reprezentowanym",
|
|
117
|
-
"reprezentowanej",
|
|
118
|
-
"wpisana",
|
|
119
|
-
"wpisane",
|
|
120
|
-
"wpisano",
|
|
121
|
-
"wpisany",
|
|
122
|
-
"wpisaną",
|
|
123
|
-
"wpisanego",
|
|
124
|
-
"wpisanej",
|
|
125
|
-
"zarejestrowana",
|
|
126
|
-
"zarejestrowany",
|
|
127
|
-
"zarejestrowaną",
|
|
128
|
-
"zarejestrowanego",
|
|
129
|
-
"zwana dalej",
|
|
130
|
-
"zwany dalej",
|
|
131
|
-
"zwaną dalej",
|
|
132
|
-
"zwanego dalej"
|
|
133
|
-
],
|
|
134
|
-
"pt-br": [
|
|
135
|
-
"CNPJ",
|
|
136
|
-
"CPF",
|
|
137
|
-
"RG",
|
|
138
|
-
"OAB",
|
|
139
|
-
"inscrita",
|
|
140
|
-
"inscrito",
|
|
141
|
-
"Inscrição Estadual",
|
|
142
|
-
"Inscrição Municipal",
|
|
143
|
-
"neste ato representada",
|
|
144
|
-
"neste ato representado",
|
|
145
|
-
"doravante denominada",
|
|
146
|
-
"doravante denominado",
|
|
147
|
-
"registrada",
|
|
148
|
-
"registrado",
|
|
149
|
-
"portador",
|
|
150
|
-
"portadora"
|
|
151
|
-
],
|
|
152
|
-
ro: [
|
|
153
|
-
"înregistrată",
|
|
154
|
-
"reprezentat",
|
|
155
|
-
"reprezentată"
|
|
156
|
-
],
|
|
157
|
-
sk: [
|
|
158
|
-
"IČO",
|
|
159
|
-
"DIČ",
|
|
160
|
-
"IČ DPH",
|
|
161
|
-
"bankové spojenie",
|
|
162
|
-
"číslo účtu",
|
|
163
|
-
"konajúci",
|
|
164
|
-
"konajúca",
|
|
165
|
-
"kontaktná osoba",
|
|
166
|
-
"registrovaný",
|
|
167
|
-
"registrovaná",
|
|
168
|
-
"splnomocnený zástupca",
|
|
169
|
-
"splnomocnená zástupkyňa",
|
|
170
|
-
"štatutárny orgán",
|
|
171
|
-
"štatutárny zástupca",
|
|
172
|
-
"v zastúpení",
|
|
173
|
-
"zapísaná",
|
|
174
|
-
"zapísaný",
|
|
175
|
-
"zapísané",
|
|
176
|
-
"zaregistrovaná",
|
|
177
|
-
"zaregistrovaný",
|
|
178
|
-
"zastúpená",
|
|
179
|
-
"zastúpený",
|
|
180
|
-
"zastúpené"
|
|
181
|
-
],
|
|
182
|
-
universal: [
|
|
183
|
-
"e-mail",
|
|
184
|
-
"email",
|
|
185
|
-
"fax",
|
|
186
|
-
"http",
|
|
187
|
-
"https",
|
|
188
|
-
"phone",
|
|
189
|
-
"tel",
|
|
190
|
-
"telephone",
|
|
191
|
-
"www"
|
|
192
|
-
]
|
|
193
|
-
};
|
|
194
|
-
//#endregion
|
|
195
|
-
export { address_boundaries_default as default };
|
|
196
|
-
|
|
197
|
-
//# sourceMappingURL=address-boundaries.mjs.map
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"address-boundaries.mjs","names":[],"sources":["../../src/data/address-boundaries.json"],"sourcesContent":[""],"mappings":""}
|
|
@@ -1,16 +0,0 @@
|
|
|
1
|
-
//#region src/data/address-jurisdiction-prefixes.json
|
|
2
|
-
var _comment = "Address-like jurisdiction prefixes that are valid location/address spans without digits or street-type words. Lowercased and organized per language.";
|
|
3
|
-
var en = [
|
|
4
|
-
"commonwealth of",
|
|
5
|
-
"district of",
|
|
6
|
-
"state of",
|
|
7
|
-
"territory of"
|
|
8
|
-
];
|
|
9
|
-
var address_jurisdiction_prefixes_default = {
|
|
10
|
-
_comment,
|
|
11
|
-
en
|
|
12
|
-
};
|
|
13
|
-
//#endregion
|
|
14
|
-
export { _comment, address_jurisdiction_prefixes_default as default, en };
|
|
15
|
-
|
|
16
|
-
//# sourceMappingURL=address-jurisdiction-prefixes.mjs.map
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"address-jurisdiction-prefixes.mjs","names":[],"sources":["../../src/data/address-jurisdiction-prefixes.json"],"sourcesContent":[""],"mappings":""}
|
|
@@ -1,148 +0,0 @@
|
|
|
1
|
-
//#region src/data/address-stop-keywords.json
|
|
2
|
-
var _comment = "Field-label keywords that terminate address-strategy scanning. When the address strategy hits a comma followed by one of these labels, the address span stops before the keyword so 'IČO:', 'Bankové spojenie:', 'IBAN:' don't get absorbed into the address. Lowercased; matched case-insensitively. Organised per-language because the labels are language-specific; at runtime the trigger detector unions every language's list into a single stop set because address spans don't carry a language tag.";
|
|
3
|
-
var cs = [
|
|
4
|
-
"číslo účtu",
|
|
5
|
-
"registrační",
|
|
6
|
-
"zastoupen",
|
|
7
|
-
"bankovní",
|
|
8
|
-
"datová",
|
|
9
|
-
"vložka",
|
|
10
|
-
"sp.zn.",
|
|
11
|
-
"jednatel",
|
|
12
|
-
"ředitel",
|
|
13
|
-
"oddíl",
|
|
14
|
-
"telefon",
|
|
15
|
-
"č.ú.",
|
|
16
|
-
"dič",
|
|
17
|
-
"ičo",
|
|
18
|
-
"ič"
|
|
19
|
-
];
|
|
20
|
-
var de = [
|
|
21
|
-
"bank",
|
|
22
|
-
"bic",
|
|
23
|
-
"iban",
|
|
24
|
-
"steuer-id",
|
|
25
|
-
"steueridentifikationsnummer",
|
|
26
|
-
"steuernummer",
|
|
27
|
-
"ust-idnr",
|
|
28
|
-
"ust-idnr."
|
|
29
|
-
];
|
|
30
|
-
var en = [
|
|
31
|
-
"e-mail",
|
|
32
|
-
"email",
|
|
33
|
-
"tel",
|
|
34
|
-
"swift",
|
|
35
|
-
"iban",
|
|
36
|
-
"bic"
|
|
37
|
-
];
|
|
38
|
-
var pl = [
|
|
39
|
-
"nip",
|
|
40
|
-
"krs",
|
|
41
|
-
"regon",
|
|
42
|
-
"pesel",
|
|
43
|
-
"numer konta",
|
|
44
|
-
"konto bankowe",
|
|
45
|
-
"rachunek bankowy",
|
|
46
|
-
"tel.",
|
|
47
|
-
"telefon",
|
|
48
|
-
"e-mail",
|
|
49
|
-
"email"
|
|
50
|
-
];
|
|
51
|
-
var fr = [
|
|
52
|
-
"e-mail",
|
|
53
|
-
"email",
|
|
54
|
-
"courriel",
|
|
55
|
-
"téléphone",
|
|
56
|
-
"tél.",
|
|
57
|
-
"tél",
|
|
58
|
-
"portable",
|
|
59
|
-
"fax",
|
|
60
|
-
"siren",
|
|
61
|
-
"siret",
|
|
62
|
-
"tva",
|
|
63
|
-
"rcs",
|
|
64
|
-
"naf",
|
|
65
|
-
"ape",
|
|
66
|
-
"iban",
|
|
67
|
-
"bic",
|
|
68
|
-
"swift",
|
|
69
|
-
"code postal",
|
|
70
|
-
"cp",
|
|
71
|
-
"nir"
|
|
72
|
-
];
|
|
73
|
-
var sk = [
|
|
74
|
-
"číslo účtu",
|
|
75
|
-
"bankové spojenie",
|
|
76
|
-
"ičo",
|
|
77
|
-
"ič dph",
|
|
78
|
-
"ičdph",
|
|
79
|
-
"dič",
|
|
80
|
-
"konateľ",
|
|
81
|
-
"konateľka",
|
|
82
|
-
"konateľom",
|
|
83
|
-
"konateľkou",
|
|
84
|
-
"riaditeľ",
|
|
85
|
-
"riaditeľka",
|
|
86
|
-
"vložka",
|
|
87
|
-
"oddiel",
|
|
88
|
-
"sp.zn.",
|
|
89
|
-
"sp. zn.",
|
|
90
|
-
"spisovej značky",
|
|
91
|
-
"č.ú.",
|
|
92
|
-
"č. účtu"
|
|
93
|
-
];
|
|
94
|
-
var pt_br = [
|
|
95
|
-
"bic",
|
|
96
|
-
"cnpj",
|
|
97
|
-
"cnpj/mf",
|
|
98
|
-
"conta",
|
|
99
|
-
"cpf",
|
|
100
|
-
"cpf/mf",
|
|
101
|
-
"doravante",
|
|
102
|
-
"e-mail",
|
|
103
|
-
"email",
|
|
104
|
-
"fax",
|
|
105
|
-
"iban",
|
|
106
|
-
"inscrição estadual",
|
|
107
|
-
"inscrição municipal",
|
|
108
|
-
"neste ato",
|
|
109
|
-
"nire",
|
|
110
|
-
"oab",
|
|
111
|
-
"portador",
|
|
112
|
-
"portadora",
|
|
113
|
-
"representada",
|
|
114
|
-
"representado",
|
|
115
|
-
"rg",
|
|
116
|
-
"swift",
|
|
117
|
-
"tel",
|
|
118
|
-
"telefone"
|
|
119
|
-
];
|
|
120
|
-
var es = [];
|
|
121
|
-
var it = [
|
|
122
|
-
"c.f.",
|
|
123
|
-
"c.f. e p.iva",
|
|
124
|
-
"codice fiscale",
|
|
125
|
-
"p. iva",
|
|
126
|
-
"p.iva",
|
|
127
|
-
"partita iva",
|
|
128
|
-
"pec",
|
|
129
|
-
"registro imprese",
|
|
130
|
-
"rappresentata",
|
|
131
|
-
"rappresentato"
|
|
132
|
-
];
|
|
133
|
-
var address_stop_keywords_default = {
|
|
134
|
-
_comment,
|
|
135
|
-
cs,
|
|
136
|
-
de,
|
|
137
|
-
en,
|
|
138
|
-
pl,
|
|
139
|
-
fr,
|
|
140
|
-
sk,
|
|
141
|
-
"pt-br": pt_br,
|
|
142
|
-
es,
|
|
143
|
-
it
|
|
144
|
-
};
|
|
145
|
-
//#endregion
|
|
146
|
-
export { _comment, cs, de, address_stop_keywords_default as default, en, es, fr, it, pl, pt_br as "pt-br", sk };
|
|
147
|
-
|
|
148
|
-
//# sourceMappingURL=address-stop-keywords.mjs.map
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"address-stop-keywords.mjs","names":[],"sources":["../../src/data/address-stop-keywords.json"],"sourcesContent":[""],"mappings":""}
|
|
@@ -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":""}
|