visura-parser 0.1.0
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 +21 -0
- package/README.md +113 -0
- package/dist/errors.d.ts +6 -0
- package/dist/errors.d.ts.map +1 -0
- package/dist/errors.js +9 -0
- package/dist/errors.js.map +1 -0
- package/dist/index.d.ts +6 -0
- package/dist/index.d.ts.map +1 -0
- package/dist/index.js +4 -0
- package/dist/index.js.map +1 -0
- package/dist/internal/extract-pdf.d.ts +3 -0
- package/dist/internal/extract-pdf.d.ts.map +1 -0
- package/dist/internal/extract-pdf.js +114 -0
- package/dist/internal/extract-pdf.js.map +1 -0
- package/dist/internal/join-pages.d.ts +7 -0
- package/dist/internal/join-pages.d.ts.map +1 -0
- package/dist/internal/join-pages.js +37 -0
- package/dist/internal/join-pages.js.map +1 -0
- package/dist/internal/map-document.d.ts +4 -0
- package/dist/internal/map-document.d.ts.map +1 -0
- package/dist/internal/map-document.js +113 -0
- package/dist/internal/map-document.js.map +1 -0
- package/dist/internal/model.d.ts +20 -0
- package/dist/internal/model.d.ts.map +1 -0
- package/dist/internal/model.js +2 -0
- package/dist/internal/model.js.map +1 -0
- package/dist/internal/parse-capital.d.ts +5 -0
- package/dist/internal/parse-capital.d.ts.map +1 -0
- package/dist/internal/parse-capital.js +76 -0
- package/dist/internal/parse-capital.js.map +1 -0
- package/dist/internal/parse-classifications.d.ts +5 -0
- package/dist/internal/parse-classifications.d.ts.map +1 -0
- package/dist/internal/parse-classifications.js +40 -0
- package/dist/internal/parse-classifications.js.map +1 -0
- package/dist/internal/parse-document.d.ts +5 -0
- package/dist/internal/parse-document.d.ts.map +1 -0
- package/dist/internal/parse-document.js +385 -0
- package/dist/internal/parse-document.js.map +1 -0
- package/dist/internal/parse-people.d.ts +9 -0
- package/dist/internal/parse-people.d.ts.map +1 -0
- package/dist/internal/parse-people.js +231 -0
- package/dist/internal/parse-people.js.map +1 -0
- package/dist/internal/text.d.ts +12 -0
- package/dist/internal/text.d.ts.map +1 -0
- package/dist/internal/text.js +99 -0
- package/dist/internal/text.js.map +1 -0
- package/dist/internal/values.d.ts +5 -0
- package/dist/internal/values.d.ts.map +1 -0
- package/dist/internal/values.js +64 -0
- package/dist/internal/values.js.map +1 -0
- package/dist/parse-visura.d.ts +3 -0
- package/dist/parse-visura.d.ts.map +1 -0
- package/dist/parse-visura.js +17 -0
- package/dist/parse-visura.js.map +1 -0
- package/dist/types.d.ts +180 -0
- package/dist/types.d.ts.map +1 -0
- package/dist/types.js +2 -0
- package/dist/types.js.map +1 -0
- package/package.json +73 -0
|
@@ -0,0 +1,40 @@
|
|
|
1
|
+
import { normalizeText, semanticSpans, spansInReadingOrder } from './text.js';
|
|
2
|
+
/** Codes are read only inside an explicitly labelled ATECO/ATECORI block. */
|
|
3
|
+
export function parseClassifications(pdf) {
|
|
4
|
+
const result = [];
|
|
5
|
+
for (const page of pdf.pages) {
|
|
6
|
+
const spans = spansInReadingOrder(page);
|
|
7
|
+
const headings = semanticSpans(page).filter((span) => /^(?:classificazione(?: dell'attivit[aà]'?)?\s+)?ateco(?:ri)?\b/i.test(normalizeText(span.text)));
|
|
8
|
+
for (const heading of headings) {
|
|
9
|
+
let current;
|
|
10
|
+
let previousY = heading.y;
|
|
11
|
+
for (const span of spans.filter((span) => span.y < heading.y - 2)) {
|
|
12
|
+
const text = normalizeText(span.text);
|
|
13
|
+
if (/^(?:\d+\s+)?(?:classificazione|attivit[aà]'?|albi|licenze|soci|amministratori|sede|capitale|addetti|codice nace)\b/i.test(text) ||
|
|
14
|
+
previousY - span.y > 45)
|
|
15
|
+
break;
|
|
16
|
+
previousY = span.y;
|
|
17
|
+
const code = /^(?:codice\s*:?\s*)?(\d{2}(?:\.\d{1,2}){0,2})(?:\s*[-–:]\s*(.+))?$/i.exec(text);
|
|
18
|
+
if (code?.[1]) {
|
|
19
|
+
current = { code: code[1] };
|
|
20
|
+
if (code[2])
|
|
21
|
+
current.description = code[2];
|
|
22
|
+
result.push(current);
|
|
23
|
+
}
|
|
24
|
+
else if (current !== undefined) {
|
|
25
|
+
const importance = /^importanza\s*:\s*(.+)$/i.exec(text);
|
|
26
|
+
if (importance?.[1])
|
|
27
|
+
current.importance = importance[1];
|
|
28
|
+
else if (/^(?:prevalente|primaria|secondaria)$/i.test(text))
|
|
29
|
+
current.importance = text;
|
|
30
|
+
else if (/^fonte\b/i.test(text))
|
|
31
|
+
continue;
|
|
32
|
+
else if (/^[a-zà-ù]/i.test(text))
|
|
33
|
+
current.description = normalizeText(`${current.description ?? ''} ${text.replace(/^descrizione\s*:\s*/i, '')}`);
|
|
34
|
+
}
|
|
35
|
+
}
|
|
36
|
+
}
|
|
37
|
+
}
|
|
38
|
+
return result.filter((item, index) => result.findIndex((other) => JSON.stringify(other) === JSON.stringify(item)) === index);
|
|
39
|
+
}
|
|
40
|
+
//# sourceMappingURL=parse-classifications.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"parse-classifications.js","sourceRoot":"","sources":["../../src/internal/parse-classifications.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,aAAa,EAAE,aAAa,EAAE,mBAAmB,EAAE,MAAM,WAAW,CAAC;AAE9E,6EAA6E;AAC7E,MAAM,UAAU,oBAAoB,CAAC,GAAiB;IACpD,MAAM,MAAM,GAA0B,EAAE,CAAC;IACzC,KAAK,MAAM,IAAI,IAAI,GAAG,CAAC,KAAK,EAAE,CAAC;QAC7B,MAAM,KAAK,GAAG,mBAAmB,CAAC,IAAI,CAAC,CAAC;QACxC,MAAM,QAAQ,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CACnD,iEAAiE,CAAC,IAAI,CACpE,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CACzB,CACF,CAAC;QACF,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;YAC/B,IAAI,OAAwC,CAAC;YAC7C,IAAI,SAAS,GAAG,OAAO,CAAC,CAAC,CAAC;YAC1B,KAAK,MAAM,IAAI,IAAI,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,GAAG,OAAO,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;gBAClE,MAAM,IAAI,GAAG,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBACtC,IACE,qHAAqH,CAAC,IAAI,CACxH,IAAI,CACL;oBACD,SAAS,GAAG,IAAI,CAAC,CAAC,GAAG,EAAE;oBAEvB,MAAM;gBACR,SAAS,GAAG,IAAI,CAAC,CAAC,CAAC;gBACnB,MAAM,IAAI,GACR,qEAAqE,CAAC,IAAI,CACxE,IAAI,CACL,CAAC;gBACJ,IAAI,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;oBACd,OAAO,GAAG,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC;oBAC5B,IAAI,IAAI,CAAC,CAAC,CAAC;wBAAE,OAAO,CAAC,WAAW,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;oBAC3C,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;gBACvB,CAAC;qBAAM,IAAI,OAAO,KAAK,SAAS,EAAE,CAAC;oBACjC,MAAM,UAAU,GAAG,0BAA0B,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;oBACzD,IAAI,UAAU,EAAE,CAAC,CAAC,CAAC;wBAAE,OAAO,CAAC,UAAU,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC;yBACnD,IAAI,uCAAuC,CAAC,IAAI,CAAC,IAAI,CAAC;wBACzD,OAAO,CAAC,UAAU,GAAG,IAAI,CAAC;yBACvB,IAAI,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC;wBAAE,SAAS;yBACrC,IAAI,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC;wBAC9B,OAAO,CAAC,WAAW,GAAG,aAAa,CACjC,GAAG,OAAO,CAAC,WAAW,IAAI,EAAE,IAAI,IAAI,CAAC,OAAO,CAAC,sBAAsB,EAAE,EAAE,CAAC,EAAE,CAC3E,CAAC;gBACN,CAAC;YACH,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC,MAAM,CAClB,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CACd,MAAM,CAAC,SAAS,CACd,CAAC,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,KAAK,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,CAC1D,KAAK,KAAK,CACd,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,5 @@
|
|
|
1
|
+
import type { VisuraDocument } from '../types.js';
|
|
2
|
+
import type { ExtractedPdf } from './model.js';
|
|
3
|
+
export declare function isSupportedVisuraBlock(pdf: ExtractedPdf): boolean;
|
|
4
|
+
export declare function parseExtractedDocument(pdf: ExtractedPdf, filename?: string): VisuraDocument;
|
|
5
|
+
//# sourceMappingURL=parse-document.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"parse-document.d.ts","sourceRoot":"","sources":["../../src/internal/parse-document.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,aAAa,CAAC;AAClD,OAAO,KAAK,EAAiB,YAAY,EAAkB,MAAM,YAAY,CAAC;AA6W9E,wBAAgB,sBAAsB,CAAC,GAAG,EAAE,YAAY,GAAG,OAAO,CAQjE;AAwBD,wBAAgB,sBAAsB,CACpC,GAAG,EAAE,YAAY,EACjB,QAAQ,CAAC,EAAE,MAAM,GAChB,cAAc,CAkFhB"}
|
|
@@ -0,0 +1,385 @@
|
|
|
1
|
+
import { normalizeText, pageText, semanticSpans } from './text.js';
|
|
2
|
+
import { parseItalianDate } from './values.js';
|
|
3
|
+
import { mapDocument } from './map-document.js';
|
|
4
|
+
import { isPersonRole, parsePeople } from './parse-people.js';
|
|
5
|
+
import { parseClassifications } from './parse-classifications.js';
|
|
6
|
+
import { parseCapital } from './parse-capital.js';
|
|
7
|
+
import { joinPages } from './join-pages.js';
|
|
8
|
+
const FIELD_SPECS = [
|
|
9
|
+
{
|
|
10
|
+
key: 'registeredOfficeAddress',
|
|
11
|
+
labels: [
|
|
12
|
+
/^indirizzo (?:della )?sede\s*legale$/i,
|
|
13
|
+
/^indirizzo sede$/i,
|
|
14
|
+
/^indirizzo$/i,
|
|
15
|
+
/^sede legale$/i,
|
|
16
|
+
],
|
|
17
|
+
pages: 'cover',
|
|
18
|
+
multiline: true,
|
|
19
|
+
},
|
|
20
|
+
{
|
|
21
|
+
key: 'certifiedEmail',
|
|
22
|
+
labels: [/^domicilio digitale\/?pec$/i, /^pec$/i],
|
|
23
|
+
pages: 'cover',
|
|
24
|
+
},
|
|
25
|
+
{ key: 'reaNumber', labels: [/^numero rea$/i], pages: 'cover' },
|
|
26
|
+
{
|
|
27
|
+
key: 'taxCode',
|
|
28
|
+
labels: [/^codice fiscale(?: e n\.iscr\. al)?$/i, /^codice fiscale$/i],
|
|
29
|
+
pages: 'cover',
|
|
30
|
+
},
|
|
31
|
+
{ key: 'vatNumber', labels: [/^partita iva$/i], pages: 'all' },
|
|
32
|
+
{ key: 'leiCode', labels: [/^codice lei$/i], pages: 'cover' },
|
|
33
|
+
{
|
|
34
|
+
key: 'legalForm',
|
|
35
|
+
labels: [/^forma giuridica$/i],
|
|
36
|
+
pages: 'cover',
|
|
37
|
+
multiline: true,
|
|
38
|
+
},
|
|
39
|
+
{
|
|
40
|
+
key: 'incorporationDate',
|
|
41
|
+
labels: [/^data atto di costituzione$/i],
|
|
42
|
+
pages: 'identity',
|
|
43
|
+
},
|
|
44
|
+
{
|
|
45
|
+
key: 'registrationDate',
|
|
46
|
+
labels: [/^data (?:di )?iscrizione$/i],
|
|
47
|
+
pages: 'identity',
|
|
48
|
+
},
|
|
49
|
+
{
|
|
50
|
+
key: 'lastProtocolDate',
|
|
51
|
+
labels: [/^data ultimo protocollo$/i],
|
|
52
|
+
pages: 'cover',
|
|
53
|
+
},
|
|
54
|
+
{ key: 'status', labels: [/^stato attivit[aà]'?$/i], pages: 'all' },
|
|
55
|
+
{
|
|
56
|
+
key: 'startDate',
|
|
57
|
+
labels: [/^data inizio (?:dell')?attivit[aà]'?(?: dell'impresa)?$/i],
|
|
58
|
+
pages: 'activity',
|
|
59
|
+
},
|
|
60
|
+
{
|
|
61
|
+
key: 'primaryActivity',
|
|
62
|
+
labels: [/^attivit[aà]'? prevalente$/i],
|
|
63
|
+
pages: 'activity',
|
|
64
|
+
multiline: true,
|
|
65
|
+
},
|
|
66
|
+
{ key: 'atecoCode', labels: [/^codice ateco$/i], pages: 'cover' },
|
|
67
|
+
{ key: 'naceCode', labels: [/^codice nace(?: 2\.1)?$/i], pages: 'cover' },
|
|
68
|
+
{
|
|
69
|
+
key: 'importExportActivity',
|
|
70
|
+
labels: [/^attivit[aà]'? import export$/i],
|
|
71
|
+
pages: 'cover',
|
|
72
|
+
},
|
|
73
|
+
{ key: 'networkContract', labels: [/^contratto di rete$/i], pages: 'cover' },
|
|
74
|
+
{
|
|
75
|
+
key: 'licensesAndRegistrations',
|
|
76
|
+
labels: [/^albi ruoli e licenze$/i],
|
|
77
|
+
pages: 'cover',
|
|
78
|
+
},
|
|
79
|
+
{
|
|
80
|
+
key: 'environmentalRegistrations',
|
|
81
|
+
labels: [/^albi e registri ambientali$/i],
|
|
82
|
+
pages: 'cover',
|
|
83
|
+
},
|
|
84
|
+
{ key: 'coverCapital', labels: [/^capitale sociale$/i], pages: 'cover' },
|
|
85
|
+
{
|
|
86
|
+
key: 'employeeCount',
|
|
87
|
+
labels: [/^addetti(?: al .+|\s*\(.+\))?$/i],
|
|
88
|
+
pages: 'cover',
|
|
89
|
+
},
|
|
90
|
+
{
|
|
91
|
+
key: 'filingCount',
|
|
92
|
+
labels: [/^documenti ri(?: dal .+|\s*\(.+\))?$/i],
|
|
93
|
+
pages: 'cover',
|
|
94
|
+
},
|
|
95
|
+
{
|
|
96
|
+
key: 'shareholdersCount',
|
|
97
|
+
labels: [/^soci e titolari di diritti su azioni e quote$/i, /^soci$/i],
|
|
98
|
+
pages: 'cover',
|
|
99
|
+
},
|
|
100
|
+
{ key: 'directorsCount', labels: [/^amministratori$/i], pages: 'cover' },
|
|
101
|
+
{
|
|
102
|
+
key: 'officeHoldersCount',
|
|
103
|
+
labels: [/^titolari di cariche$/i],
|
|
104
|
+
pages: 'cover',
|
|
105
|
+
},
|
|
106
|
+
{ key: 'localUnitsCount', labels: [/^unit[aà]'? locali$/i], pages: 'cover' },
|
|
107
|
+
{
|
|
108
|
+
key: 'filingsLast12Months',
|
|
109
|
+
labels: [/^pratiche inviate negli ultimi 12 mesi$/i],
|
|
110
|
+
pages: 'cover',
|
|
111
|
+
},
|
|
112
|
+
{ key: 'soa', labels: [/^attestazioni soa$/i], pages: 'cover' },
|
|
113
|
+
{
|
|
114
|
+
key: 'shareTransfersCount',
|
|
115
|
+
labels: [/^trasferimenti di quote$/i],
|
|
116
|
+
pages: 'cover',
|
|
117
|
+
},
|
|
118
|
+
{
|
|
119
|
+
key: 'registeredOfficeTransfersCount',
|
|
120
|
+
labels: [/^trasferimenti di sede$/i],
|
|
121
|
+
pages: 'cover',
|
|
122
|
+
},
|
|
123
|
+
{ key: 'hasEquityInterests', labels: [/^partecipazioni$/i], pages: 'cover' },
|
|
124
|
+
{ key: 'companyFileAvailable', labels: [/^fascicolo$/i], pages: 'cover' },
|
|
125
|
+
{
|
|
126
|
+
key: 'articlesOfAssociationAvailable',
|
|
127
|
+
labels: [/^statuto$/i],
|
|
128
|
+
pages: 'cover',
|
|
129
|
+
},
|
|
130
|
+
{ key: 'otherActsCount', labels: [/^altri atti$/i], pages: 'cover' },
|
|
131
|
+
{
|
|
132
|
+
key: 'quality',
|
|
133
|
+
labels: [/^certificazioni di qualit[aà]'?$/i],
|
|
134
|
+
pages: 'cover',
|
|
135
|
+
},
|
|
136
|
+
{ key: 'financialStatementYears', labels: [/^bilanci$/i], pages: 'cover' },
|
|
137
|
+
];
|
|
138
|
+
const ALL_LABELS = FIELD_SPECS.flatMap((field) => field.labels);
|
|
139
|
+
const DATE_FIELDS = new Set([
|
|
140
|
+
'incorporationDate',
|
|
141
|
+
'registrationDate',
|
|
142
|
+
'lastProtocolDate',
|
|
143
|
+
'startDate',
|
|
144
|
+
]);
|
|
145
|
+
function matchesAny(text, patterns) {
|
|
146
|
+
const normalized = normalizeText(text);
|
|
147
|
+
return patterns.some((pattern) => pattern.test(normalized));
|
|
148
|
+
}
|
|
149
|
+
function isKnownLabel(text) {
|
|
150
|
+
return (isPersonRole(text) ||
|
|
151
|
+
ALL_LABELS.some((pattern) => pattern.test(normalizeText(text))));
|
|
152
|
+
}
|
|
153
|
+
function samePanel(page, left, right) {
|
|
154
|
+
const midpoint = page.width / 2;
|
|
155
|
+
return left.x < midpoint === right.x < midpoint;
|
|
156
|
+
}
|
|
157
|
+
function inlineValue(label, patterns) {
|
|
158
|
+
const text = normalizeText(label.text);
|
|
159
|
+
const separator = text.indexOf(':');
|
|
160
|
+
if (separator < 0 || !matchesAny(text.slice(0, separator), patterns))
|
|
161
|
+
return undefined;
|
|
162
|
+
const value = normalizeText(text.slice(separator + 1));
|
|
163
|
+
if (value.length > 0)
|
|
164
|
+
return value;
|
|
165
|
+
return undefined;
|
|
166
|
+
}
|
|
167
|
+
function findValueOnPage(page, patterns, multiline = false, fullWidth = false) {
|
|
168
|
+
const labels = semanticSpans(page).filter((span) => matchesAny(span.text, patterns) ||
|
|
169
|
+
inlineValue(span, patterns) !== undefined);
|
|
170
|
+
for (const label of labels) {
|
|
171
|
+
const inline = inlineValue(label, patterns);
|
|
172
|
+
if (inline !== undefined)
|
|
173
|
+
return inline;
|
|
174
|
+
const candidates = page.spans
|
|
175
|
+
.filter((span) => span !== label &&
|
|
176
|
+
(fullWidth || samePanel(page, label, span)) &&
|
|
177
|
+
span.x >= label.x + Math.max(label.width, 10) - 2 &&
|
|
178
|
+
Math.abs(span.y - label.y) <= Math.max(5, label.height * 0.75) &&
|
|
179
|
+
!isKnownLabel(span.text))
|
|
180
|
+
.sort((left, right) => left.x - right.x);
|
|
181
|
+
let value = candidates[0];
|
|
182
|
+
if (value === undefined) {
|
|
183
|
+
value = page.spans
|
|
184
|
+
.filter((span) => span !== label &&
|
|
185
|
+
(fullWidth || samePanel(page, label, span)) &&
|
|
186
|
+
span.x > label.x + Math.max(label.width, 10) - 2 &&
|
|
187
|
+
label.y - span.y >= -3 &&
|
|
188
|
+
label.y - span.y <= 18 &&
|
|
189
|
+
!isKnownLabel(span.text))
|
|
190
|
+
.sort((left, right) => {
|
|
191
|
+
const vertical = Math.abs(left.y - label.y) - Math.abs(right.y - label.y);
|
|
192
|
+
return vertical === 0 ? left.x - right.x : vertical;
|
|
193
|
+
})[0];
|
|
194
|
+
}
|
|
195
|
+
if (value === undefined)
|
|
196
|
+
continue;
|
|
197
|
+
if (!multiline)
|
|
198
|
+
return normalizeText(value.text);
|
|
199
|
+
const nextLabel = semanticSpans(page)
|
|
200
|
+
.filter((span) => (fullWidth || samePanel(page, value, span)) &&
|
|
201
|
+
span.x < value.x - 10 &&
|
|
202
|
+
span.y < value.y - 2 &&
|
|
203
|
+
(fullWidth || isKnownLabel(span.text)))
|
|
204
|
+
.sort((left, right) => right.y - left.y)[0];
|
|
205
|
+
const lowerBoundary = nextLabel?.y ?? (fullWidth ? -Infinity : value.y - 120);
|
|
206
|
+
const continuations = page.spans
|
|
207
|
+
.filter((span) => (fullWidth || samePanel(page, value, span)) &&
|
|
208
|
+
(fullWidth
|
|
209
|
+
? span.x >= value.x - 8
|
|
210
|
+
: Math.abs(span.x - value.x) <= 8) &&
|
|
211
|
+
value.y - span.y > 2 &&
|
|
212
|
+
span.y > lowerBoundary + 2 &&
|
|
213
|
+
!isKnownLabel(span.text))
|
|
214
|
+
.sort((left, right) => right.y - left.y || left.x - right.x);
|
|
215
|
+
const parts = fullWidth
|
|
216
|
+
? page.spans
|
|
217
|
+
.filter((span) => span.x >= value.x - 2 && Math.abs(span.y - value.y) <= 2)
|
|
218
|
+
.sort((a, b) => a.x - b.x)
|
|
219
|
+
.map((span) => span.text)
|
|
220
|
+
: [value.text];
|
|
221
|
+
let previousY = value.y;
|
|
222
|
+
for (const continuation of continuations) {
|
|
223
|
+
if (previousY - continuation.y > 18)
|
|
224
|
+
break;
|
|
225
|
+
parts.push(continuation.text);
|
|
226
|
+
previousY = continuation.y;
|
|
227
|
+
}
|
|
228
|
+
return normalizeText(parts.join(' '));
|
|
229
|
+
}
|
|
230
|
+
return undefined;
|
|
231
|
+
}
|
|
232
|
+
function findCoverPage(pdf) {
|
|
233
|
+
return (pdf.pages.find((page) => /\bvisura\b/i.test(pageText(page))) ??
|
|
234
|
+
pdf.pages.find((page) => /registro imprese/i.test(pageText(page))) ??
|
|
235
|
+
pdf.pages.find((page) => page.spans.length > 0));
|
|
236
|
+
}
|
|
237
|
+
function findBusinessName(page) {
|
|
238
|
+
const spans = semanticSpans(page);
|
|
239
|
+
const title = spans
|
|
240
|
+
.filter((span) => span.y > page.height * 0.35 &&
|
|
241
|
+
/^(?:visura\s+(?:ordinaria|storica|(?:di\s+)?evasione)\b|soci e titolari di diritti|esito evasione protocollo)/i.test(normalizeText(span.text)))
|
|
242
|
+
.sort((left, right) => right.y - left.y || right.width - left.width)[0];
|
|
243
|
+
if (title === undefined)
|
|
244
|
+
return undefined;
|
|
245
|
+
const first = page.spans
|
|
246
|
+
.filter((span) => span.y < title.y - 8 &&
|
|
247
|
+
span.y > title.y - 95 &&
|
|
248
|
+
span.x < page.width / 2 &&
|
|
249
|
+
/[a-zà-ù]/i.test(span.text) &&
|
|
250
|
+
!/^(camera di commercio|registro imprese|informazioni|sede|codice|rea\b)/i.test(normalizeText(span.text)))
|
|
251
|
+
.sort((left, right) => {
|
|
252
|
+
const height = right.height - left.height;
|
|
253
|
+
return Math.abs(height) > 0.5 ? height : right.y - left.y;
|
|
254
|
+
})[0];
|
|
255
|
+
if (first === undefined)
|
|
256
|
+
return undefined;
|
|
257
|
+
const parts = [first.text];
|
|
258
|
+
let previous = first;
|
|
259
|
+
for (const candidate of [...page.spans].sort((a, b) => b.y - a.y)) {
|
|
260
|
+
if (candidate.y >= previous.y - 3 || Math.abs(candidate.x - first.x) > 3)
|
|
261
|
+
continue;
|
|
262
|
+
if (previous.y - candidate.y > 20 ||
|
|
263
|
+
Math.abs(candidate.height - first.height) > 0.5 ||
|
|
264
|
+
isKnownLabel(candidate.text))
|
|
265
|
+
break;
|
|
266
|
+
parts.push(candidate.text);
|
|
267
|
+
previous = candidate;
|
|
268
|
+
}
|
|
269
|
+
return parts.join(' ');
|
|
270
|
+
}
|
|
271
|
+
function findBusinessNameFromLabels(pdf) {
|
|
272
|
+
const labels = [/^denominazione$/i, /^ditta$/i, /^ragione sociale$/i];
|
|
273
|
+
for (const page of pdf.pages) {
|
|
274
|
+
const value = findValueOnPage(page, labels, true);
|
|
275
|
+
if (value !== undefined)
|
|
276
|
+
return value;
|
|
277
|
+
}
|
|
278
|
+
return undefined;
|
|
279
|
+
}
|
|
280
|
+
function detectVisuraType(cover) {
|
|
281
|
+
const titles = semanticSpans(cover).filter((span) => span.y > cover.height * 0.35);
|
|
282
|
+
const title = titles.find((span) => /^visura\s+(?:ordinaria|storica|(?:di\s+)?evasione)\b/i.test(normalizeText(span.text)));
|
|
283
|
+
if (title !== undefined)
|
|
284
|
+
return normalizeText(title.text);
|
|
285
|
+
const evasion = titles.find((span) => /^esito evasione protocollo\b/i.test(normalizeText(span.text)));
|
|
286
|
+
if (evasion !== undefined)
|
|
287
|
+
return 'Visura di Evasione';
|
|
288
|
+
return undefined;
|
|
289
|
+
}
|
|
290
|
+
export function isSupportedVisuraBlock(pdf) {
|
|
291
|
+
const cover = findCoverPage(pdf);
|
|
292
|
+
return (/dati anagrafici/i.test(pageText(cover)) &&
|
|
293
|
+
cover.spans.some((span) => /^soci e titolari di diritti/i.test(normalizeText(span.text))));
|
|
294
|
+
}
|
|
295
|
+
function beforeSection(pdf, heading) {
|
|
296
|
+
const pages = [];
|
|
297
|
+
for (const page of pdf.pages) {
|
|
298
|
+
const boundary = semanticSpans(page)
|
|
299
|
+
.filter((span) => span.x < page.width * 0.15 && heading.test(normalizeText(span.text)))
|
|
300
|
+
.sort((left, right) => right.y - left.y)[0];
|
|
301
|
+
if (boundary === undefined) {
|
|
302
|
+
pages.push(page);
|
|
303
|
+
continue;
|
|
304
|
+
}
|
|
305
|
+
pages.push({
|
|
306
|
+
...page,
|
|
307
|
+
spans: page.spans.filter((span) => span.y > boundary.y),
|
|
308
|
+
});
|
|
309
|
+
break;
|
|
310
|
+
}
|
|
311
|
+
return { pages };
|
|
312
|
+
}
|
|
313
|
+
export function parseExtractedDocument(pdf, filename) {
|
|
314
|
+
const cover = findCoverPage(pdf);
|
|
315
|
+
const currentPdf = beforeSection(pdf, /^(?:\d+\s+)?(?:storia (?:delle|dei|di)|informazioni storiche|protocollo evaso\b)/i);
|
|
316
|
+
// Company dates precede the domain sections; later registration dates can
|
|
317
|
+
// belong to a shareholder or an appointment rather than the company.
|
|
318
|
+
const identityPdf = beforeSection(currentPdf, /^\d+\s+(?:capitale|soci|amministratori|sindaci|titolari|attivit[aà]'?|trasferimenti|scioglimento|altre cariche)\b/i);
|
|
319
|
+
const fields = {};
|
|
320
|
+
const activityBody = joinPages(currentPdf.pages.filter((page) => page.number !== cover.number));
|
|
321
|
+
for (const field of FIELD_SPECS) {
|
|
322
|
+
const key = field.key;
|
|
323
|
+
const pages = field.pages === 'cover'
|
|
324
|
+
? [cover]
|
|
325
|
+
: field.pages === 'identity'
|
|
326
|
+
? identityPdf.pages
|
|
327
|
+
: field.pages === 'activity'
|
|
328
|
+
? [cover, activityBody]
|
|
329
|
+
: currentPdf.pages;
|
|
330
|
+
for (const page of pages) {
|
|
331
|
+
const value = findValueOnPage(page, field.labels, field.multiline, page === activityBody);
|
|
332
|
+
if (value === undefined)
|
|
333
|
+
continue;
|
|
334
|
+
const normalized = DATE_FIELDS.has(key) ? parseItalianDate(value) : value;
|
|
335
|
+
if (normalized !== undefined) {
|
|
336
|
+
fields[key] = normalized;
|
|
337
|
+
break;
|
|
338
|
+
}
|
|
339
|
+
}
|
|
340
|
+
}
|
|
341
|
+
const result = mapDocument(fields);
|
|
342
|
+
if (filename !== undefined)
|
|
343
|
+
result.filename = filename;
|
|
344
|
+
const reportType = detectVisuraType(cover);
|
|
345
|
+
if (reportType !== undefined)
|
|
346
|
+
result.reportType = reportType;
|
|
347
|
+
const name = findBusinessName(cover) ?? findBusinessNameFromLabels(currentPdf);
|
|
348
|
+
if (name !== undefined)
|
|
349
|
+
result.companyName = normalizeText(name);
|
|
350
|
+
// Dates belong to their printed summary label, not the report extraction date.
|
|
351
|
+
for (const label of semanticSpans(cover)) {
|
|
352
|
+
const text = normalizeText(label.text);
|
|
353
|
+
const employees = /^addetti(?: al)?\s*\(?(\d{2}\/\d{2}\/\d{4})\)?$/i.exec(text);
|
|
354
|
+
const filings = /^documenti ri(?: dal)?\s*\(?(\d{2}\/\d{2}\/\d{4})\)?$/i.exec(text);
|
|
355
|
+
const rawDate = employees?.[1] ?? filings?.[1];
|
|
356
|
+
const date = rawDate === undefined ? undefined : parseItalianDate(rawDate);
|
|
357
|
+
if (date === undefined)
|
|
358
|
+
continue;
|
|
359
|
+
if (employees)
|
|
360
|
+
result.employees = { ...result.employees, referenceDate: date };
|
|
361
|
+
if (filings)
|
|
362
|
+
result.businessRegisterFilings = {
|
|
363
|
+
...result.businessRegisterFilings,
|
|
364
|
+
sinceDate: date,
|
|
365
|
+
};
|
|
366
|
+
}
|
|
367
|
+
const people = parsePeople(currentPdf, cover.number);
|
|
368
|
+
if (people.primaryRepresentative)
|
|
369
|
+
result.primaryRepresentative = people.primaryRepresentative;
|
|
370
|
+
if (people.officers.length)
|
|
371
|
+
result.officers = people.officers;
|
|
372
|
+
if (people.shareholders.length)
|
|
373
|
+
result.shareholders = people.shareholders;
|
|
374
|
+
const classifications = parseClassifications(currentPdf);
|
|
375
|
+
if (classifications.length)
|
|
376
|
+
result.activity = {
|
|
377
|
+
...result.activity,
|
|
378
|
+
atecoClassifications: classifications,
|
|
379
|
+
};
|
|
380
|
+
const capital = parseCapital(currentPdf);
|
|
381
|
+
if (capital !== undefined)
|
|
382
|
+
result.shareCapital = capital;
|
|
383
|
+
return result;
|
|
384
|
+
}
|
|
385
|
+
//# sourceMappingURL=parse-document.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"parse-document.js","sourceRoot":"","sources":["../../src/internal/parse-document.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,aAAa,EAAE,QAAQ,EAAE,aAAa,EAAE,MAAM,WAAW,CAAC;AACnE,OAAO,EAAE,gBAAgB,EAAE,MAAM,aAAa,CAAC;AAC/C,OAAO,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAChD,OAAO,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAC9D,OAAO,EAAE,oBAAoB,EAAE,MAAM,4BAA4B,CAAC;AAClE,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAClD,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAS5C,MAAM,WAAW,GAAG;IAClB;QACE,GAAG,EAAE,yBAAyB;QAC9B,MAAM,EAAE;YACN,uCAAuC;YACvC,mBAAmB;YACnB,cAAc;YACd,gBAAgB;SACjB;QACD,KAAK,EAAE,OAAO;QACd,SAAS,EAAE,IAAI;KAChB;IACD;QACE,GAAG,EAAE,gBAAgB;QACrB,MAAM,EAAE,CAAC,6BAA6B,EAAE,QAAQ,CAAC;QACjD,KAAK,EAAE,OAAO;KACf;IACD,EAAE,GAAG,EAAE,WAAW,EAAE,MAAM,EAAE,CAAC,eAAe,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;IAC/D;QACE,GAAG,EAAE,SAAS;QACd,MAAM,EAAE,CAAC,uCAAuC,EAAE,mBAAmB,CAAC;QACtE,KAAK,EAAE,OAAO;KACf;IACD,EAAE,GAAG,EAAE,WAAW,EAAE,MAAM,EAAE,CAAC,gBAAgB,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE;IAC9D,EAAE,GAAG,EAAE,SAAS,EAAE,MAAM,EAAE,CAAC,eAAe,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;IAC7D;QACE,GAAG,EAAE,WAAW;QAChB,MAAM,EAAE,CAAC,oBAAoB,CAAC;QAC9B,KAAK,EAAE,OAAO;QACd,SAAS,EAAE,IAAI;KAChB;IACD;QACE,GAAG,EAAE,mBAAmB;QACxB,MAAM,EAAE,CAAC,8BAA8B,CAAC;QACxC,KAAK,EAAE,UAAU;KAClB;IACD;QACE,GAAG,EAAE,kBAAkB;QACvB,MAAM,EAAE,CAAC,4BAA4B,CAAC;QACtC,KAAK,EAAE,UAAU;KAClB;IACD;QACE,GAAG,EAAE,kBAAkB;QACvB,MAAM,EAAE,CAAC,2BAA2B,CAAC;QACrC,KAAK,EAAE,OAAO;KACf;IACD,EAAE,GAAG,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC,wBAAwB,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE;IACnE;QACE,GAAG,EAAE,WAAW;QAChB,MAAM,EAAE,CAAC,0DAA0D,CAAC;QACpE,KAAK,EAAE,UAAU;KAClB;IACD;QACE,GAAG,EAAE,iBAAiB;QACtB,MAAM,EAAE,CAAC,6BAA6B,CAAC;QACvC,KAAK,EAAE,UAAU;QACjB,SAAS,EAAE,IAAI;KAChB;IACD,EAAE,GAAG,EAAE,WAAW,EAAE,MAAM,EAAE,CAAC,iBAAiB,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;IACjE,EAAE,GAAG,EAAE,UAAU,EAAE,MAAM,EAAE,CAAC,0BAA0B,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;IACzE;QACE,GAAG,EAAE,sBAAsB;QAC3B,MAAM,EAAE,CAAC,gCAAgC,CAAC;QAC1C,KAAK,EAAE,OAAO;KACf;IACD,EAAE,GAAG,EAAE,iBAAiB,EAAE,MAAM,EAAE,CAAC,sBAAsB,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;IAC5E;QACE,GAAG,EAAE,0BAA0B;QAC/B,MAAM,EAAE,CAAC,yBAAyB,CAAC;QACnC,KAAK,EAAE,OAAO;KACf;IACD;QACE,GAAG,EAAE,4BAA4B;QACjC,MAAM,EAAE,CAAC,+BAA+B,CAAC;QACzC,KAAK,EAAE,OAAO;KACf;IACD,EAAE,GAAG,EAAE,cAAc,EAAE,MAAM,EAAE,CAAC,qBAAqB,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;IACxE;QACE,GAAG,EAAE,eAAe;QACpB,MAAM,EAAE,CAAC,iCAAiC,CAAC;QAC3C,KAAK,EAAE,OAAO;KACf;IACD;QACE,GAAG,EAAE,aAAa;QAClB,MAAM,EAAE,CAAC,uCAAuC,CAAC;QACjD,KAAK,EAAE,OAAO;KACf;IACD;QACE,GAAG,EAAE,mBAAmB;QACxB,MAAM,EAAE,CAAC,iDAAiD,EAAE,SAAS,CAAC;QACtE,KAAK,EAAE,OAAO;KACf;IACD,EAAE,GAAG,EAAE,gBAAgB,EAAE,MAAM,EAAE,CAAC,mBAAmB,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;IACxE;QACE,GAAG,EAAE,oBAAoB;QACzB,MAAM,EAAE,CAAC,wBAAwB,CAAC;QAClC,KAAK,EAAE,OAAO;KACf;IACD,EAAE,GAAG,EAAE,iBAAiB,EAAE,MAAM,EAAE,CAAC,sBAAsB,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;IAC5E;QACE,GAAG,EAAE,qBAAqB;QAC1B,MAAM,EAAE,CAAC,0CAA0C,CAAC;QACpD,KAAK,EAAE,OAAO;KACf;IACD,EAAE,GAAG,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC,qBAAqB,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;IAC/D;QACE,GAAG,EAAE,qBAAqB;QAC1B,MAAM,EAAE,CAAC,2BAA2B,CAAC;QACrC,KAAK,EAAE,OAAO;KACf;IACD;QACE,GAAG,EAAE,gCAAgC;QACrC,MAAM,EAAE,CAAC,0BAA0B,CAAC;QACpC,KAAK,EAAE,OAAO;KACf;IACD,EAAE,GAAG,EAAE,oBAAoB,EAAE,MAAM,EAAE,CAAC,mBAAmB,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;IAC5E,EAAE,GAAG,EAAE,sBAAsB,EAAE,MAAM,EAAE,CAAC,cAAc,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;IACzE;QACE,GAAG,EAAE,gCAAgC;QACrC,MAAM,EAAE,CAAC,YAAY,CAAC;QACtB,KAAK,EAAE,OAAO;KACf;IACD,EAAE,GAAG,EAAE,gBAAgB,EAAE,MAAM,EAAE,CAAC,eAAe,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;IACpE;QACE,GAAG,EAAE,SAAS;QACd,MAAM,EAAE,CAAC,mCAAmC,CAAC;QAC7C,KAAK,EAAE,OAAO;KACf;IACD,EAAE,GAAG,EAAE,yBAAyB,EAAE,MAAM,EAAE,CAAC,YAAY,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE;CACnC,CAAC;AAI1C,MAAM,UAAU,GAAG,WAAW,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;AAChE,MAAM,WAAW,GAA2B,IAAI,GAAG,CAAC;IAClD,mBAAmB;IACnB,kBAAkB;IAClB,kBAAkB;IAClB,WAAW;CACZ,CAAC,CAAC;AAEH,SAAS,UAAU,CAAC,IAAY,EAAE,QAA2B;IAC3D,MAAM,UAAU,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IACvC,OAAO,QAAQ,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC;AAC9D,CAAC;AAED,SAAS,YAAY,CAAC,IAAY;IAChC,OAAO,CACL,YAAY,CAAC,IAAI,CAAC;QAClB,UAAU,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC,CAChE,CAAC;AACJ,CAAC;AAED,SAAS,SAAS,CAChB,IAAmB,EACnB,IAAoB,EACpB,KAAqB;IAErB,MAAM,QAAQ,GAAG,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC;IAChC,OAAO,IAAI,CAAC,CAAC,GAAG,QAAQ,KAAK,KAAK,CAAC,CAAC,GAAG,QAAQ,CAAC;AAClD,CAAC;AAED,SAAS,WAAW,CAClB,KAAqB,EACrB,QAA2B;IAE3B,MAAM,IAAI,GAAG,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACvC,MAAM,SAAS,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;IACpC,IAAI,SAAS,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,SAAS,CAAC,EAAE,QAAQ,CAAC;QAClE,OAAO,SAAS,CAAC;IACnB,MAAM,KAAK,GAAG,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC;IACvD,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,KAAK,CAAC;IACnC,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,SAAS,eAAe,CACtB,IAAmB,EACnB,QAA2B,EAC3B,SAAS,GAAG,KAAK,EACjB,SAAS,GAAG,KAAK;IAEjB,MAAM,MAAM,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC,MAAM,CACvC,CAAC,IAAI,EAAE,EAAE,CACP,UAAU,CAAC,IAAI,CAAC,IAAI,EAAE,QAAQ,CAAC;QAC/B,WAAW,CAAC,IAAI,EAAE,QAAQ,CAAC,KAAK,SAAS,CAC5C,CAAC;IAEF,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,MAAM,MAAM,GAAG,WAAW,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAC;QAC5C,IAAI,MAAM,KAAK,SAAS;YAAE,OAAO,MAAM,CAAC;QAExC,MAAM,UAAU,GAAG,IAAI,CAAC,KAAK;aAC1B,MAAM,CACL,CAAC,IAAI,EAAE,EAAE,CACP,IAAI,KAAK,KAAK;YACd,CAAC,SAAS,IAAI,SAAS,CAAC,IAAI,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC;YAC3C,IAAI,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,EAAE,EAAE,CAAC,GAAG,CAAC;YACjD,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,MAAM,GAAG,IAAI,CAAC;YAC9D,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAC3B;aACA,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QAE3C,IAAI,KAAK,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC;QAC1B,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,KAAK,GAAG,IAAI,CAAC,KAAK;iBACf,MAAM,CACL,CAAC,IAAI,EAAE,EAAE,CACP,IAAI,KAAK,KAAK;gBACd,CAAC,SAAS,IAAI,SAAS,CAAC,IAAI,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC;gBAC3C,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,EAAE,EAAE,CAAC,GAAG,CAAC;gBAChD,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC;gBACtB,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,EAAE;gBACtB,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAC3B;iBACA,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;gBACpB,MAAM,QAAQ,GACZ,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;gBAC3D,OAAO,QAAQ,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC;YACtD,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QACV,CAAC;QACD,IAAI,KAAK,KAAK,SAAS;YAAE,SAAS;QAElC,IAAI,CAAC,SAAS;YAAE,OAAO,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAEjD,MAAM,SAAS,GAAG,aAAa,CAAC,IAAI,CAAC;aAClC,MAAM,CACL,CAAC,IAAI,EAAE,EAAE,CACP,CAAC,SAAS,IAAI,SAAS,CAAC,IAAI,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC;YAC3C,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,GAAG,EAAE;YACrB,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC;YACpB,CAAC,SAAS,IAAI,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CACzC;aACA,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAC9C,MAAM,aAAa,GACjB,SAAS,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC;QAC1D,MAAM,aAAa,GAAG,IAAI,CAAC,KAAK;aAC7B,MAAM,CACL,CAAC,IAAI,EAAE,EAAE,CACP,CAAC,SAAS,IAAI,SAAS,CAAC,IAAI,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC;YAC3C,CAAC,SAAS;gBACR,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,GAAG,CAAC;gBACvB,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;YACpC,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC;YACpB,IAAI,CAAC,CAAC,GAAG,aAAa,GAAG,CAAC;YAC1B,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAC3B;aACA,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QAC/D,MAAM,KAAK,GAAG,SAAS;YACrB,CAAC,CAAC,IAAI,CAAC,KAAK;iBACP,MAAM,CACL,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CACnE;iBACA,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;iBACzB,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC;YAC7B,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACjB,IAAI,SAAS,GAAG,KAAK,CAAC,CAAC,CAAC;QACxB,KAAK,MAAM,YAAY,IAAI,aAAa,EAAE,CAAC;YACzC,IAAI,SAAS,GAAG,YAAY,CAAC,CAAC,GAAG,EAAE;gBAAE,MAAM;YAC3C,KAAK,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;YAC9B,SAAS,GAAG,YAAY,CAAC,CAAC,CAAC;QAC7B,CAAC;QAED,OAAO,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;IACxC,CAAC;IAED,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,SAAS,aAAa,CAAC,GAAiB;IACtC,OAAO,CACL,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,aAAa,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;QAC5D,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,mBAAmB,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;QAClE,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAE,CACjD,CAAC;AACJ,CAAC;AAED,SAAS,gBAAgB,CAAC,IAAmB;IAC3C,MAAM,KAAK,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IAClC,MAAM,KAAK,GAAG,KAAK;SAChB,MAAM,CACL,CAAC,IAAI,EAAE,EAAE,CACP,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,MAAM,GAAG,IAAI;QAC3B,gHAAgH,CAAC,IAAI,CACnH,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CACzB,CACJ;SACA,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,KAAK,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1E,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAE1C,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK;SACrB,MAAM,CACL,CAAC,IAAI,EAAE,EAAE,CACP,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,GAAG,CAAC;QACpB,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,GAAG,EAAE;QACrB,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,GAAG,CAAC;QACvB,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC;QAC3B,CAAC,yEAAyE,CAAC,IAAI,CAC7E,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CACzB,CACJ;SACA,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;QACpB,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC;QAC1C,OAAO,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC;IAC5D,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACR,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAE1C,MAAM,KAAK,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC3B,IAAI,QAAQ,GAAG,KAAK,CAAC;IACrB,KAAK,MAAM,SAAS,IAAI,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;QAClE,IAAI,SAAS,CAAC,CAAC,IAAI,QAAQ,CAAC,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC;YACtE,SAAS;QACX,IACE,QAAQ,CAAC,CAAC,GAAG,SAAS,CAAC,CAAC,GAAG,EAAE;YAC7B,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,GAAG,GAAG;YAC/C,YAAY,CAAC,SAAS,CAAC,IAAI,CAAC;YAE5B,MAAM;QACR,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;QAC3B,QAAQ,GAAG,SAAS,CAAC;IACvB,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AACzB,CAAC;AAED,SAAS,0BAA0B,CAAC,GAAiB;IACnD,MAAM,MAAM,GAAG,CAAC,kBAAkB,EAAE,UAAU,EAAE,oBAAoB,CAAC,CAAC;IACtE,KAAK,MAAM,IAAI,IAAI,GAAG,CAAC,KAAK,EAAE,CAAC;QAC7B,MAAM,KAAK,GAAG,eAAe,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,CAAC,CAAC;QAClD,IAAI,KAAK,KAAK,SAAS;YAAE,OAAO,KAAK,CAAC;IACxC,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,SAAS,gBAAgB,CAAC,KAAoB;IAC5C,MAAM,MAAM,GAAG,aAAa,CAAC,KAAK,CAAC,CAAC,MAAM,CACxC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,MAAM,GAAG,IAAI,CACvC,CAAC;IACF,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CACjC,uDAAuD,CAAC,IAAI,CAC1D,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CACzB,CACF,CAAC;IACF,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC1D,MAAM,OAAO,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CACnC,+BAA+B,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAC/D,CAAC;IACF,IAAI,OAAO,KAAK,SAAS;QAAE,OAAO,oBAAoB,CAAC;IACvD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,MAAM,UAAU,sBAAsB,CAAC,GAAiB;IACtD,MAAM,KAAK,GAAG,aAAa,CAAC,GAAG,CAAC,CAAC;IACjC,OAAO,CACL,kBAAkB,CAAC,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC;QACxC,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CACxB,8BAA8B,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAC9D,CACF,CAAC;AACJ,CAAC;AAED,SAAS,aAAa,CAAC,GAAiB,EAAE,OAAe;IACvD,MAAM,KAAK,GAAoB,EAAE,CAAC;IAClC,KAAK,MAAM,IAAI,IAAI,GAAG,CAAC,KAAK,EAAE,CAAC;QAC7B,MAAM,QAAQ,GAAG,aAAa,CAAC,IAAI,CAAC;aACjC,MAAM,CACL,CAAC,IAAI,EAAE,EAAE,CACP,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,GAAG,IAAI,IAAI,OAAO,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CACvE;aACA,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAC9C,IAAI,QAAQ,KAAK,SAAS,EAAE,CAAC;YAC3B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACjB,SAAS;QACX,CAAC;QACD,KAAK,CAAC,IAAI,CAAC;YACT,GAAG,IAAI;YACP,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC;SACxD,CAAC,CAAC;QACH,MAAM;IACR,CAAC;IACD,OAAO,EAAE,KAAK,EAAE,CAAC;AACnB,CAAC;AAED,MAAM,UAAU,sBAAsB,CACpC,GAAiB,EACjB,QAAiB;IAEjB,MAAM,KAAK,GAAG,aAAa,CAAC,GAAG,CAAC,CAAC;IACjC,MAAM,UAAU,GAAG,aAAa,CAC9B,GAAG,EACH,mFAAmF,CACpF,CAAC;IACF,0EAA0E;IAC1E,qEAAqE;IACrE,MAAM,WAAW,GAAG,aAAa,CAC/B,UAAU,EACV,oHAAoH,CACrH,CAAC;IACF,MAAM,MAAM,GAAuC,EAAE,CAAC;IACtD,MAAM,YAAY,GAAG,SAAS,CAC5B,UAAU,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,KAAK,KAAK,CAAC,MAAM,CAAC,CAChE,CAAC;IACF,KAAK,MAAM,KAAK,IAAI,WAAmC,EAAE,CAAC;QACxD,MAAM,GAAG,GAAG,KAAK,CAAC,GAAgB,CAAC;QACnC,MAAM,KAAK,GACT,KAAK,CAAC,KAAK,KAAK,OAAO;YACrB,CAAC,CAAC,CAAC,KAAK,CAAC;YACT,CAAC,CAAC,KAAK,CAAC,KAAK,KAAK,UAAU;gBAC1B,CAAC,CAAC,WAAW,CAAC,KAAK;gBACnB,CAAC,CAAC,KAAK,CAAC,KAAK,KAAK,UAAU;oBAC1B,CAAC,CAAC,CAAC,KAAK,EAAE,YAAY,CAAC;oBACvB,CAAC,CAAC,UAAU,CAAC,KAAK,CAAC;QAC3B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,KAAK,GAAG,eAAe,CAC3B,IAAI,EACJ,KAAK,CAAC,MAAM,EACZ,KAAK,CAAC,SAAS,EACf,IAAI,KAAK,YAAY,CACtB,CAAC;YACF,IAAI,KAAK,KAAK,SAAS;gBAAE,SAAS;YAClC,MAAM,UAAU,GAAG,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,gBAAgB,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;YAC1E,IAAI,UAAU,KAAK,SAAS,EAAE,CAAC;gBAC7B,MAAM,CAAC,GAAG,CAAC,GAAG,UAAU,CAAC;gBACzB,MAAM;YACR,CAAC;QACH,CAAC;IACH,CAAC;IACD,MAAM,MAAM,GAAG,WAAW,CAAC,MAAM,CAAC,CAAC;IACnC,IAAI,QAAQ,KAAK,SAAS;QAAE,MAAM,CAAC,QAAQ,GAAG,QAAQ,CAAC;IACvD,MAAM,UAAU,GAAG,gBAAgB,CAAC,KAAK,CAAC,CAAC;IAC3C,IAAI,UAAU,KAAK,SAAS;QAAE,MAAM,CAAC,UAAU,GAAG,UAAU,CAAC;IAC7D,MAAM,IAAI,GACR,gBAAgB,CAAC,KAAK,CAAC,IAAI,0BAA0B,CAAC,UAAU,CAAC,CAAC;IACpE,IAAI,IAAI,KAAK,SAAS;QAAE,MAAM,CAAC,WAAW,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IAEjE,+EAA+E;IAC/E,KAAK,MAAM,KAAK,IAAI,aAAa,CAAC,KAAK,CAAC,EAAE,CAAC;QACzC,MAAM,IAAI,GAAG,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QACvC,MAAM,SAAS,GAAG,kDAAkD,CAAC,IAAI,CACvE,IAAI,CACL,CAAC;QACF,MAAM,OAAO,GACX,wDAAwD,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACtE,MAAM,OAAO,GAAG,SAAS,EAAE,CAAC,CAAC,CAAC,IAAI,OAAO,EAAE,CAAC,CAAC,CAAC,CAAC;QAC/C,MAAM,IAAI,GAAG,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,gBAAgB,CAAC,OAAO,CAAC,CAAC;QAC3E,IAAI,IAAI,KAAK,SAAS;YAAE,SAAS;QACjC,IAAI,SAAS;YACX,MAAM,CAAC,SAAS,GAAG,EAAE,GAAG,MAAM,CAAC,SAAS,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC;QAClE,IAAI,OAAO;YACT,MAAM,CAAC,uBAAuB,GAAG;gBAC/B,GAAG,MAAM,CAAC,uBAAuB;gBACjC,SAAS,EAAE,IAAI;aAChB,CAAC;IACN,CAAC;IACD,MAAM,MAAM,GAAG,WAAW,CAAC,UAAU,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC;IACrD,IAAI,MAAM,CAAC,qBAAqB;QAC9B,MAAM,CAAC,qBAAqB,GAAG,MAAM,CAAC,qBAAqB,CAAC;IAC9D,IAAI,MAAM,CAAC,QAAQ,CAAC,MAAM;QAAE,MAAM,CAAC,QAAQ,GAAG,MAAM,CAAC,QAAQ,CAAC;IAC9D,IAAI,MAAM,CAAC,YAAY,CAAC,MAAM;QAAE,MAAM,CAAC,YAAY,GAAG,MAAM,CAAC,YAAY,CAAC;IAC1E,MAAM,eAAe,GAAG,oBAAoB,CAAC,UAAU,CAAC,CAAC;IACzD,IAAI,eAAe,CAAC,MAAM;QACxB,MAAM,CAAC,QAAQ,GAAG;YAChB,GAAG,MAAM,CAAC,QAAQ;YAClB,oBAAoB,EAAE,eAAe;SACtC,CAAC;IACJ,MAAM,OAAO,GAAG,YAAY,CAAC,UAAU,CAAC,CAAC;IACzC,IAAI,OAAO,KAAK,SAAS;QAAE,MAAM,CAAC,YAAY,GAAG,OAAO,CAAC;IACzD,OAAO,MAAM,CAAC;AAChB,CAAC"}
|
|
@@ -0,0 +1,9 @@
|
|
|
1
|
+
import type { Officer, Representative, Shareholder } from '../types.js';
|
|
2
|
+
import type { ExtractedPdf } from './model.js';
|
|
3
|
+
export declare function isPersonRole(text: string): boolean;
|
|
4
|
+
export declare function parsePeople(pdf: ExtractedPdf, coverPage: number): {
|
|
5
|
+
primaryRepresentative?: Representative;
|
|
6
|
+
officers: Officer[];
|
|
7
|
+
shareholders: Shareholder[];
|
|
8
|
+
};
|
|
9
|
+
//# sourceMappingURL=parse-people.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"parse-people.d.ts","sourceRoot":"","sources":["../../src/internal/parse-people.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,OAAO,EAAE,cAAc,EAAE,WAAW,EAAE,MAAM,aAAa,CAAC;AACxE,OAAO,KAAK,EAAiB,YAAY,EAAkB,MAAM,YAAY,CAAC;AAiB9E,wBAAgB,YAAY,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAGlD;AA+HD,wBAAgB,WAAW,CACzB,GAAG,EAAE,YAAY,EACjB,SAAS,EAAE,MAAM,GAChB;IACD,qBAAqB,CAAC,EAAE,cAAc,CAAC;IACvC,QAAQ,EAAE,OAAO,EAAE,CAAC;IACpB,YAAY,EAAE,WAAW,EAAE,CAAC;CAC7B,CAiMA"}
|