visura-parser 0.1.0

package/dist/internal/parse-people.js

@@ -0,0 +1,231 @@
+import { foldText, normalizeText, semanticSpans } from './text.js';
+import { parseItalianAmount } from './values.js';
+import { joinPages } from './join-pages.js';
+const OFFICER_ROLE = /^(?:presidente(?: (?:del )?consiglio (?:di |d')?amministrazione| del collegio sindacale)?|vice ?presidente(?: (?:del )?consiglio (?:di |d')?amministrazione)?|amministratore(?: unico| delegato)?|consigliere(?: delegato)?|sindaco(?: effettivo| supplente)?|titolare(?: dell'impresa individuale| di impresa individuale| firmatari[oa])?|liquidatore|procuratore(?: speciale| generale)?)$/i;
+const SHAREHOLDER_ROLE = /^(?:soci[oa](?: unic[oa]| accomandante| accomandatari[oa])?|(?:nuda )?propriet[aà]'?|usufrutto)$/i;
+const RIGHT = /^(?:(?:nuda )?propriet[aà]'?|usufrutto)$/i;
+const TAX_CODE = /^(?:codice fiscale\s*:?\s*)?([A-Z0-9]{16}|\d{11})$/i;
+const SECTION = /^(?:\d+\s+)(?:soci|amministratori|sindaci|titolari|altre cariche|attivit[aà]'?|sede|capitale|storia|informazioni|trasferimenti|scioglimento|protocollo)\b/i;
+const PEOPLE_SECTION = /^\d+\s+(?:soci|amministratori|sindaci|titolari|altre cariche)\b/i;
+const OWNERSHIP_ROW_TOLERANCE = 5;
+export function isPersonRole(text) {
+    const value = normalizeText(text);
+    return OFFICER_ROLE.test(value) || SHAREHOLDER_ROLE.test(value);
+}
+function isName(text) {
+    return (text.length >= 2 &&
+        text.length <= 160 &&
+        /[a-zà-ù]/i.test(text) &&
+        !/[:;]|^\d|^[a-z]{0,3}\d{5,}/i.test(text) &&
+        !TAX_CODE.test(text) &&
+        !/^(?:elenco|numero|codice|nato|nata|residente|domicilio|residenza|indirizzo|\(?in carica|dal\s+\d|durata|quota|valore|tipo diritto|rappresentante|informazioni|amministratori|soci\b|sindaci|titolari|data |carica\b|poteri\b|registro|camera|visura|via\b|viale\b|piazza\b|cittadinanza|proprieta|versato|euro\b|capitale)/i.test(text) &&
+        !OFFICER_ROLE.test(text) &&
+        !SHAREHOLDER_ROLE.test(text));
+}
+function identities(page) {
+    const result = [];
+    const spans = semanticSpans(page);
+    for (const span of spans) {
+        const match = TAX_CODE.exec(normalizeText(span.text));
+        if (!match?.[1])
+            continue;
+        const name = page.spans
+            .filter((candidate) => candidate.y >= span.y - 2 &&
+            candidate.y - span.y < 85 &&
+            Math.abs(candidate.x - span.x) < 220 &&
+            isName(normalizeText(candidate.text)) &&
+            !page.spans.some((label) => Math.abs(label.y - candidate.y) <= 2 &&
+                label.x < candidate.x &&
+                /^(?:nato|nata|domicilio|residenza|indirizzo|cittadinanza)\b/i.test(normalizeText(label.text))))
+            .sort((a, b) => a.y - b.y)[0];
+        if (name &&
+            !result.some((item) => item.name === name && item.taxCode === match[1]))
+            result.push({ name, taxCode: match[1], y: span.y });
+    }
+    return result;
+}
+function adjacentName(page, role, roleFragments) {
+    const samePanel = (span) => span.x < page.width / 2 === role.x < page.width / 2;
+    const candidates = page.spans.filter((span) => span !== role &&
+        !roleFragments.has(span) &&
+        isName(normalizeText(span.text)));
+    return (candidates
+        .filter((span) => samePanel(span) &&
+        span.x > role.x + role.width - 2 &&
+        Math.abs(span.y - role.y) <= 4)
+        .sort((a, b) => a.x - b.x)[0] ??
+        candidates
+            .filter((span) => samePanel(span) &&
+            Math.abs(span.x - role.x) <= 35 &&
+            role.y - span.y >= 4 &&
+            role.y - span.y <= 34)
+            .sort((a, b) => b.y - a.y)[0]);
+}
+function fullName(page, first) {
+    const parts = [first.text];
+    let previous = first;
+    for (const span of [...page.spans].sort((a, b) => b.y - a.y)) {
+        if (span.y >= previous.y - 2 || Math.abs(span.x - first.x) > 3)
+            continue;
+        if (previous.y - span.y > 15 ||
+            Math.abs(span.height - first.height) > 0.5 ||
+            !isName(normalizeText(span.text)) ||
+            page.spans.some((label) => label.x < first.x - 10 && Math.abs(label.y - span.y) <= 2))
+            break;
+        parts.push(span.text);
+        previous = span;
+    }
+    return normalizeText(parts.join(' '));
+}
+function mergeOfficer(officers, entry) {
+    const existing = officers.find((other) => entry.taxCode !== undefined && other.taxCode !== undefined
+        ? entry.taxCode === other.taxCode
+        : foldText(other.name ?? '') === foldText(entry.name ?? ''));
+    if (!existing) {
+        officers.push(entry);
+        return;
+    }
+    if (entry.taxCode !== undefined)
+        existing.taxCode = entry.taxCode;
+    for (const role of entry.roles ?? []) {
+        existing.roles ??= [];
+        if (!existing.roles.some((other) => foldText(other) === foldText(role)))
+            existing.roles.push(role);
+    }
+}
+export function parsePeople(pdf, coverPage) {
+    const result = { officers: [], shareholders: [] };
+    const page = joinPages(pdf.pages);
+    const spans = semanticSpans(page);
+    const records = identities(page);
+    const sections = spans
+        .filter((span) => span.x < page.width * 0.15 && SECTION.test(normalizeText(span.text)))
+        .sort((a, b) => a.y - b.y);
+    // Longer reconstructed roles take precedence over a contained 'Presidente'.
+    const roles = spans
+        .filter((span) => OFFICER_ROLE.test(normalizeText(span.text)) ||
+        SHAREHOLDER_ROLE.test(normalizeText(span.text)))
+        .filter((span) => {
+        const section = sections.find((heading) => heading.y >= span.y);
+        return (section === undefined ||
+            PEOPLE_SECTION.test(normalizeText(section.text)));
+    })
+        .filter((span) => !spans.some((other) => other !== span &&
+        other.x <= span.x &&
+        Math.abs(other.y - span.y) <= 2 &&
+        other.width > span.width &&
+        other.x + other.width >= span.x + span.width &&
+        (OFFICER_ROLE.test(normalizeText(other.text)) ||
+            SHAREHOLDER_ROLE.test(normalizeText(other.text)))))
+        .sort((a, b) => b.y - a.y || a.x - b.x);
+    const roleFragments = new Set(page.spans.filter((span) => roles.some((role) => role !== span &&
+        Math.abs(role.x - span.x) <= 3 &&
+        role.y - span.y >= 0 &&
+        role.y - span.y <= 36 &&
+        normalizeText(role.text).endsWith(` ${normalizeText(span.text)}`))));
+    for (const role of roles) {
+        const text = normalizeText(role.text);
+        // A value beside 'carica' belongs to the preceding identity. The next
+        // row can be another qualification, not a new person's name.
+        const isDetailRole = page.spans.some((span) => span.x < role.x &&
+            Math.abs(span.y - role.y) <= 2 &&
+            /^carica$/i.test(normalizeText(span.text)));
+        const adjacent = isDetailRole
+            ? undefined
+            : adjacentName(page, role, roleFragments);
+        let identity = adjacent === undefined
+            ? undefined
+            : records.find((record) => record.name === adjacent);
+        if (adjacent === undefined) {
+            identity = records
+                .filter((record) => record.y > role.y &&
+                record.y - role.y < 220 &&
+                !spans.some((span) => SECTION.test(normalizeText(span.text)) &&
+                    span.y < record.y &&
+                    span.y > role.y) &&
+                !records.some((other) => other !== record &&
+                    other.name.y < record.y &&
+                    other.name.y > role.y))
+                .sort((a, b) => a.y - b.y)[0];
+        }
+        const name = adjacent ?? identity?.name;
+        if (name === undefined)
+            continue;
+        const person = {
+            name: fullName(page, name),
+            ...(identity ? { taxCode: identity.taxCode } : {}),
+        };
+        if (OFFICER_ROLE.test(text)) {
+            mergeOfficer(result.officers, { ...person, roles: [text] });
+            if (role.page === coverPage &&
+                adjacent !== undefined &&
+                !spans.some((span) => SECTION.test(normalizeText(span.text)) && span.y > role.y) &&
+                result.primaryRepresentative === undefined &&
+                !/^(?:sindaco|presidente del collegio)/i.test(text)) {
+                result.primaryRepresentative = { name: person.name, role: text };
+            }
+        }
+        else {
+            const shareholder = { ...person };
+            if (RIGHT.test(text))
+                shareholder.rightType = text;
+            if (/^soci[oa] unic[oa]$/i.test(text))
+                shareholder.isSoleShareholder = true;
+            if (role.page === coverPage &&
+                adjacent !== undefined &&
+                result.primaryRepresentative === undefined &&
+                spans.some((span) => span.page === coverPage &&
+                    Math.abs(span.x - name.x) <= 3 &&
+                    name.y - span.y > 0 &&
+                    name.y - span.y <= 20 &&
+                    /^rappresentante dell'impresa$/i.test(normalizeText(span.text))))
+                result.primaryRepresentative = { name: person.name, role: text };
+            // Quota facts must be inside this ownership record. No capital/paid-up fallback.
+            const nextRole = roles.find((other) => other.y < name.y - 2);
+            const nextSection = spans
+                .filter((span) => SECTION.test(normalizeText(span.text)) && span.y < name.y)
+                .sort((a, b) => b.y - a.y)[0];
+            const lower = Math.max(name.y - 180, nextRole === undefined
+                ? -Infinity
+                : nextRole.y + OWNERSHIP_ROW_TOLERANCE, nextSection?.y ?? -Infinity);
+            const detail = spans.filter((span) => span.y <= role.y + OWNERSHIP_ROW_TOLERANCE &&
+                span.y > lower &&
+                !SECTION.test(normalizeText(span.text)));
+            for (const span of detail) {
+                const value = normalizeText(span.text);
+                const nominal = /^(?:valore nominale|quota)\s*(?:di nominali)?\s*:?\s*([\d.,]+)\s*(euro|EUR)?$/i.exec(value);
+                if (nominal?.[1]) {
+                    const amount = parseItalianAmount(nominal[1]);
+                    if (amount !== undefined)
+                        shareholder.nominalValue = amount;
+                    if (nominal[2])
+                        shareholder.currency = 'EUR';
+                }
+                const percentage = /^(?:percentuale(?: di partecipazione)?\s*:?\s*)?([\d.,]+)\s*%$/i.exec(value);
+                const amount = percentage?.[1] === undefined
+                    ? undefined
+                    : parseItalianAmount(percentage[1]);
+                if (amount !== undefined && amount <= 100)
+                    shareholder.ownershipPercentage = amount;
+                const right = /^tipo (?:di )?diritto\s*:\s*(.+)$/i.exec(value);
+                if (right?.[1] && RIGHT.test(right[1]))
+                    shareholder.rightType = right[1];
+            }
+            const existing = result.shareholders.find((other) => (person.taxCode !== undefined && other.taxCode !== undefined
+                ? other.taxCode === person.taxCode
+                : foldText(other.name ?? '') === foldText(person.name)) &&
+                (other.rightType === undefined ||
+                    shareholder.rightType === undefined ||
+                    foldText(other.rightType) === foldText(shareholder.rightType)) &&
+                (other.nominalValue === undefined ||
+                    shareholder.nominalValue === undefined ||
+                    other.nominalValue === shareholder.nominalValue));
+            if (existing)
+                Object.assign(existing, shareholder);
+            else
+                result.shareholders.push(shareholder);
+        }
+    }
+    return result;
+}
+//# sourceMappingURL=parse-people.js.map

package/dist/internal/parse-people.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"parse-people.js","sourceRoot":"","sources":["../../src/internal/parse-people.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,QAAQ,EAAE,aAAa,EAAE,aAAa,EAAE,MAAM,WAAW,CAAC;AACnE,OAAO,EAAE,kBAAkB,EAAE,MAAM,aAAa,CAAC;AACjD,OAAO,EAAE,SAAS,EAAE,MAAM,iBAAiB,CAAC;AAE5C,MAAM,YAAY,GAChB,gYAAgY,CAAC;AACnY,MAAM,gBAAgB,GACpB,mGAAmG,CAAC;AACtG,MAAM,KAAK,GAAG,2CAA2C,CAAC;AAC1D,MAAM,QAAQ,GAAG,qDAAqD,CAAC;AACvE,MAAM,OAAO,GACX,4JAA4J,CAAC;AAC/J,MAAM,cAAc,GAClB,kEAAkE,CAAC;AACrE,MAAM,uBAAuB,GAAG,CAAC,CAAC;AAElC,MAAM,UAAU,YAAY,CAAC,IAAY;IACvC,MAAM,KAAK,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IAClC,OAAO,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,gBAAgB,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;AAClE,CAAC;AAED,SAAS,MAAM,CAAC,IAAY;IAC1B,OAAO,CACL,IAAI,CAAC,MAAM,IAAI,CAAC;QAChB,IAAI,CAAC,MAAM,IAAI,GAAG;QAClB,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC;QACtB,CAAC,6BAA6B,CAAC,IAAI,CAAC,IAAI,CAAC;QACzC,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC;QACpB,CAAC,6TAA6T,CAAC,IAAI,CACjU,IAAI,CACL;QACD,CAAC,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC;QACxB,CAAC,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,CAC7B,CAAC;AACJ,CAAC;AAQD,SAAS,UAAU,CAAC,IAAmB;IACrC,MAAM,MAAM,GAAe,EAAE,CAAC;IAC9B,MAAM,KAAK,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IAClC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,KAAK,GAAG,QAAQ,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;QACtD,IAAI,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;YAAE,SAAS;QAC1B,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK;aACpB,MAAM,CACL,CAAC,SAAS,EAAE,EAAE,CACZ,SAAS,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,GAAG,CAAC;YACzB,SAAS,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,EAAE;YACzB,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,GAAG,GAAG;YACpC,MAAM,CAAC,aAAa,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;YACrC,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CACd,CAAC,KAAK,EAAE,EAAE,CACR,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,GAAG,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;gBACpC,KAAK,CAAC,CAAC,GAAG,SAAS,CAAC,CAAC;gBACrB,8DAA8D,CAAC,IAAI,CACjE,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAC1B,CACJ,CACJ;aACA,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAChC,IACE,IAAI;YACJ,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,KAAK,IAAI,IAAI,IAAI,CAAC,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC;YAEvE,MAAM,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,IAAI,CAAC,CAAC,EAAE,CAAC,CAAC;IACxD,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,YAAY,CACnB,IAAmB,EACnB,IAAoB,EACpB,aAA0C;IAE1C,MAAM,SAAS,GAAG,CAAC,IAAoB,EAAE,EAAE,CACzC,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,GAAG,CAAC,KAAK,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC;IACtD,MAAM,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,CAClC,CAAC,IAAI,EAAE,EAAE,CACP,IAAI,KAAK,IAAI;QACb,CAAC,aAAa,CAAC,GAAG,CAAC,IAAI,CAAC;QACxB,MAAM,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CACnC,CAAC;IACF,OAAO,CACL,UAAU;SACP,MAAM,CACL,CAAC,IAAI,EAAE,EAAE,CACP,SAAS,CAAC,IAAI,CAAC;QACf,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,GAAG,CAAC;QAChC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CACjC;SACA,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAC/B,UAAU;aACP,MAAM,CACL,CAAC,IAAI,EAAE,EAAE,CACP,SAAS,CAAC,IAAI,CAAC;YACf,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE;YAC/B,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,CAAC;YACpB,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,EAAE,CACxB;aACA,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAChC,CAAC;AACJ,CAAC;AAED,SAAS,QAAQ,CAAC,IAAmB,EAAE,KAAqB;IAC1D,MAAM,KAAK,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC3B,IAAI,QAAQ,GAAG,KAAK,CAAC;IACrB,KAAK,MAAM,IAAI,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;QAC7D,IAAI,IAAI,CAAC,CAAC,IAAI,QAAQ,CAAC,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC;YAAE,SAAS;QACzE,IACE,QAAQ,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,EAAE;YACxB,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM,CAAC,GAAG,GAAG;YAC1C,CAAC,MAAM,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACjC,IAAI,CAAC,KAAK,CAAC,IAAI,CACb,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,GAAG,EAAE,IAAI,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CACrE;YAED,MAAM;QACR,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACtB,QAAQ,GAAG,IAAI,CAAC;IAClB,CAAC;IACD,OAAO,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;AACxC,CAAC;AAED,SAAS,YAAY,CAAC,QAAmB,EAAE,KAAc;IACvD,MAAM,QAAQ,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CACvC,KAAK,CAAC,OAAO,KAAK,SAAS,IAAI,KAAK,CAAC,OAAO,KAAK,SAAS;QACxD,CAAC,CAAC,KAAK,CAAC,OAAO,KAAK,KAAK,CAAC,OAAO;QACjC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,IAAI,EAAE,CAAC,KAAK,QAAQ,CAAC,KAAK,CAAC,IAAI,IAAI,EAAE,CAAC,CAC9D,CAAC;IACF,IAAI,CAAC,QAAQ,EAAE,CAAC;QACd,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACrB,OAAO;IACT,CAAC;IACD,IAAI,KAAK,CAAC,OAAO,KAAK,SAAS;QAAE,QAAQ,CAAC,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC;IAClE,KAAK,MAAM,IAAI,IAAI,KAAK,CAAC,KAAK,IAAI,EAAE,EAAE,CAAC;QACrC,QAAQ,CAAC,KAAK,KAAK,EAAE,CAAC;QACtB,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,KAAK,QAAQ,CAAC,IAAI,CAAC,CAAC;YACrE,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC9B,CAAC;AACH,CAAC;AAED,MAAM,UAAU,WAAW,CACzB,GAAiB,EACjB,SAAiB;IAMjB,MAAM,MAAM,GAIR,EAAE,QAAQ,EAAE,EAAE,EAAE,YAAY,EAAE,EAAE,EAAE,CAAC;IACvC,MAAM,IAAI,GAAG,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;IAClC,MAAM,KAAK,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;IAClC,MAAM,OAAO,GAAG,UAAU,CAAC,IAAI,CAAC,CAAC;IACjC,MAAM,QAAQ,GAAG,KAAK;SACnB,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;SACA,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;IAC7B,4EAA4E;IAC5E,MAAM,KAAK,GAAG,KAAK;SAChB,MAAM,CACL,CAAC,IAAI,EAAE,EAAE,CACP,YAAY,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC3C,gBAAgB,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAClD;SACA,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE;QACf,MAAM,OAAO,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,CAAC;QAChE,OAAO,CACL,OAAO,KAAK,SAAS;YACrB,cAAc,CAAC,IAAI,CAAC,aAAa,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CACjD,CAAC;IACJ,CAAC,CAAC;SACD,MAAM,CACL,CAAC,IAAI,EAAE,EAAE,CACP,CAAC,KAAK,CAAC,IAAI,CACT,CAAC,KAAK,EAAE,EAAE,CACR,KAAK,KAAK,IAAI;QACd,KAAK,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC;QACjB,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC;QAC/B,KAAK,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK;QACxB,KAAK,CAAC,CAAC,GAAG,KAAK,CAAC,KAAK,IAAI,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK;QAC5C,CAAC,YAAY,CAAC,IAAI,CAAC,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;YAC3C,gBAAgB,CAAC,IAAI,CAAC,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,CACtD,CACJ;SACA,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1C,MAAM,aAAa,GAAG,IAAI,GAAG,CAC3B,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CACzB,KAAK,CAAC,IAAI,CACR,CAAC,IAAI,EAAE,EAAE,CACP,IAAI,KAAK,IAAI;QACb,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC;QAC9B,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,CAAC;QACpB,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,EAAE;QACrB,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,IAAI,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CACpE,CACF,CACF,CAAC;IACF,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,IAAI,GAAG,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACtC,sEAAsE;QACtE,6DAA6D;QAC7D,MAAM,YAAY,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAClC,CAAC,IAAI,EAAE,EAAE,CACP,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;YACf,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC;YAC9B,WAAW,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAC7C,CAAC;QACF,MAAM,QAAQ,GAAG,YAAY;YAC3B,CAAC,CAAC,SAAS;YACX,CAAC,CAAC,YAAY,CAAC,IAAI,EAAE,IAAI,EAAE,aAAa,CAAC,CAAC;QAC5C,IAAI,QAAQ,GACV,QAAQ,KAAK,SAAS;YACpB,CAAC,CAAC,SAAS;YACX,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,KAAK,QAAQ,CAAC,CAAC;QACzD,IAAI,QAAQ,KAAK,SAAS,EAAE,CAAC;YAC3B,QAAQ,GAAG,OAAO;iBACf,MAAM,CACL,CAAC,MAAM,EAAE,EAAE,CACT,MAAM,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC;gBACjB,MAAM,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,GAAG;gBACvB,CAAC,KAAK,CAAC,IAAI,CACT,CAAC,IAAI,EAAE,EAAE,CACP,OAAO,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;oBACtC,IAAI,CAAC,CAAC,GAAG,MAAM,CAAC,CAAC;oBACjB,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAClB;gBACD,CAAC,OAAO,CAAC,IAAI,CACX,CAAC,KAAK,EAAE,EAAE,CACR,KAAK,KAAK,MAAM;oBAChB,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,MAAM,CAAC,CAAC;oBACvB,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CACxB,CACJ;iBACA,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAClC,CAAC;QACD,MAAM,IAAI,GAAG,QAAQ,IAAI,QAAQ,EAAE,IAAI,CAAC;QACxC,IAAI,IAAI,KAAK,SAAS;YAAE,SAAS;QACjC,MAAM,MAAM,GAAG;YACb,IAAI,EAAE,QAAQ,CAAC,IAAI,EAAE,IAAI,CAAC;YAC1B,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,OAAO,EAAE,QAAQ,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;SACnD,CAAC;QACF,IAAI,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAC5B,YAAY,CAAC,MAAM,CAAC,QAAQ,EAAE,EAAE,GAAG,MAAM,EAAE,KAAK,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YAC5D,IACE,IAAI,CAAC,IAAI,KAAK,SAAS;gBACvB,QAAQ,KAAK,SAAS;gBACtB,CAAC,KAAK,CAAC,IAAI,CACT,CAAC,IAAI,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CACpE;gBACD,MAAM,CAAC,qBAAqB,KAAK,SAAS;gBAC1C,CAAC,uCAAuC,CAAC,IAAI,CAAC,IAAI,CAAC,EACnD,CAAC;gBACD,MAAM,CAAC,qBAAqB,GAAG,EAAE,IAAI,EAAE,MAAM,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;YACnE,CAAC;QACH,CAAC;aAAM,CAAC;YACN,MAAM,WAAW,GAAgB,EAAE,GAAG,MAAM,EAAE,CAAC;YAC/C,IAAI,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC;gBAAE,WAAW,CAAC,SAAS,GAAG,IAAI,CAAC;YACnD,IAAI,sBAAsB,CAAC,IAAI,CAAC,IAAI,CAAC;gBACnC,WAAW,CAAC,iBAAiB,GAAG,IAAI,CAAC;YACvC,IACE,IAAI,CAAC,IAAI,KAAK,SAAS;gBACvB,QAAQ,KAAK,SAAS;gBACtB,MAAM,CAAC,qBAAqB,KAAK,SAAS;gBAC1C,KAAK,CAAC,IAAI,CACR,CAAC,IAAI,EAAE,EAAE,CACP,IAAI,CAAC,IAAI,KAAK,SAAS;oBACvB,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC;oBAC9B,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC;oBACnB,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,EAAE;oBACrB,gCAAgC,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAClE;gBAED,MAAM,CAAC,qBAAqB,GAAG,EAAE,IAAI,EAAE,MAAM,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;YACnE,iFAAiF;YACjF,MAAM,QAAQ,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YAC7D,MAAM,WAAW,GAAG,KAAK;iBACtB,MAAM,CACL,CAAC,IAAI,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CACpE;iBACA,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;YAChC,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,CACpB,IAAI,CAAC,CAAC,GAAG,GAAG,EACZ,QAAQ,KAAK,SAAS;gBACpB,CAAC,CAAC,CAAC,QAAQ;gBACX,CAAC,CAAC,QAAQ,CAAC,CAAC,GAAG,uBAAuB,EACxC,WAAW,EAAE,CAAC,IAAI,CAAC,QAAQ,CAC5B,CAAC;YACF,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,CACzB,CAAC,IAAI,EAAE,EAAE,CACP,IAAI,CAAC,CAAC,IAAI,IAAI,CAAC,CAAC,GAAG,uBAAuB;gBAC1C,IAAI,CAAC,CAAC,GAAG,KAAK;gBACd,CAAC,OAAO,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAC1C,CAAC;YACF,KAAK,MAAM,IAAI,IAAI,MAAM,EAAE,CAAC;gBAC1B,MAAM,KAAK,GAAG,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBACvC,MAAM,OAAO,GACX,gFAAgF,CAAC,IAAI,CACnF,KAAK,CACN,CAAC;gBACJ,IAAI,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;oBACjB,MAAM,MAAM,GAAG,kBAAkB,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC;oBAC9C,IAAI,MAAM,KAAK,SAAS;wBAAE,WAAW,CAAC,YAAY,GAAG,MAAM,CAAC;oBAC5D,IAAI,OAAO,CAAC,CAAC,CAAC;wBAAE,WAAW,CAAC,QAAQ,GAAG,KAAK,CAAC;gBAC/C,CAAC;gBACD,MAAM,UAAU,GACd,iEAAiE,CAAC,IAAI,CACpE,KAAK,CACN,CAAC;gBACJ,MAAM,MAAM,GACV,UAAU,EAAE,CAAC,CAAC,CAAC,KAAK,SAAS;oBAC3B,CAAC,CAAC,SAAS;oBACX,CAAC,CAAC,kBAAkB,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;gBACxC,IAAI,MAAM,KAAK,SAAS,IAAI,MAAM,IAAI,GAAG;oBACvC,WAAW,CAAC,mBAAmB,GAAG,MAAM,CAAC;gBAC3C,MAAM,KAAK,GAAG,oCAAoC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;gBAC/D,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;oBACpC,WAAW,CAAC,SAAS,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;YACrC,CAAC;YACD,MAAM,QAAQ,GAAG,MAAM,CAAC,YAAY,CAAC,IAAI,CACvC,CAAC,KAAK,EAAE,EAAE,CACR,CAAC,MAAM,CAAC,OAAO,KAAK,SAAS,IAAI,KAAK,CAAC,OAAO,KAAK,SAAS;gBAC1D,CAAC,CAAC,KAAK,CAAC,OAAO,KAAK,MAAM,CAAC,OAAO;gBAClC,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,IAAI,EAAE,CAAC,KAAK,QAAQ,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;gBACzD,CAAC,KAAK,CAAC,SAAS,KAAK,SAAS;oBAC5B,WAAW,CAAC,SAAS,KAAK,SAAS;oBACnC,QAAQ,CAAC,KAAK,CAAC,SAAS,CAAC,KAAK,QAAQ,CAAC,WAAW,CAAC,SAAS,CAAC,CAAC;gBAChE,CAAC,KAAK,CAAC,YAAY,KAAK,SAAS;oBAC/B,WAAW,CAAC,YAAY,KAAK,SAAS;oBACtC,KAAK,CAAC,YAAY,KAAK,WAAW,CAAC,YAAY,CAAC,CACrD,CAAC;YACF,IAAI,QAAQ;gBAAE,MAAM,CAAC,MAAM,CAAC,QAAQ,EAAE,WAAW,CAAC,CAAC;;gBAC9C,MAAM,CAAC,YAAY,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC;QAC7C,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC"}

package/dist/internal/text.d.ts

@@ -0,0 +1,12 @@
+import type { ExtractedPage, PositionedSpan } from './model.js';
+export declare function normalizeText(value: string): string;
+export declare function foldText(value: string): string;
+export declare function unique(values: readonly string[]): string[];
+export declare function spansInReadingOrder(page: ExtractedPage): PositionedSpan[];
+/**
+ * Adds deterministic same-line phrases for PDFs that split one printed label
+ * into several text items. Original spans are retained for value lookup.
+ */
+export declare function semanticSpans(page: ExtractedPage): PositionedSpan[];
+export declare function pageText(page: ExtractedPage): string;
+//# sourceMappingURL=text.d.ts.map

package/dist/internal/text.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"text.d.ts","sourceRoot":"","sources":["../../src/internal/text.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,aAAa,EAAE,cAAc,EAAE,MAAM,YAAY,CAAC;AAEhE,wBAAgB,aAAa,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAMnD;AAED,wBAAgB,QAAQ,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAE9C;AAED,wBAAgB,MAAM,CAAC,MAAM,EAAE,SAAS,MAAM,EAAE,GAAG,MAAM,EAAE,CAa1D;AAED,wBAAgB,mBAAmB,CAAC,IAAI,EAAE,aAAa,GAAG,cAAc,EAAE,CAMzE;AAED;;;GAGG;AACH,wBAAgB,aAAa,CAAC,IAAI,EAAE,aAAa,GAAG,cAAc,EAAE,CAkEnE;AAED,wBAAgB,QAAQ,CAAC,IAAI,EAAE,aAAa,GAAG,MAAM,CAIpD"}

package/dist/internal/text.js

@@ -0,0 +1,99 @@
+export function normalizeText(value) {
+    return value
+        .normalize('NFKC')
+        .replace(/[’‘`´]/g, "'")
+        .replace(/\s+/g, ' ')
+        .trim();
+}
+export function foldText(value) {
+    return normalizeText(value).toLowerCase();
+}
+export function unique(values) {
+    const seen = new Set();
+    const result = [];
+    for (const value of values) {
+        const normalized = normalizeText(value);
+        const key = foldText(normalized);
+        if (normalized.length === 0 || seen.has(key))
+            continue;
+        seen.add(key);
+        result.push(normalized);
+    }
+    return result;
+}
+export function spansInReadingOrder(page) {
+    return [...page.spans].sort((left, right) => {
+        const rowDifference = right.y - left.y;
+        if (Math.abs(rowDifference) > 2)
+            return rowDifference;
+        return left.x - right.x;
+    });
+}
+/**
+ * Adds deterministic same-line phrases for PDFs that split one printed label
+ * into several text items. Original spans are retained for value lookup.
+ */
+export function semanticSpans(page) {
+    const original = spansInReadingOrder(page);
+    const rows = [];
+    for (const span of original) {
+        const row = rows.find((candidate) => Math.abs((candidate[0]?.y ?? span.y) - span.y) <= 2);
+        if (row === undefined)
+            rows.push([span]);
+        else
+            row.push(span);
+    }
+    const phrases = [...original];
+    for (const row of rows) {
+        row.sort((left, right) => left.x - right.x);
+        for (let start = 0; start < row.length; start += 1) {
+            const first = row[start];
+            if (first === undefined)
+                continue;
+            let text = first.text;
+            let right = first.x + first.width;
+            for (let end = start + 1; end < Math.min(row.length, start + 7); end += 1) {
+                const next = row[end];
+                if (next === undefined || next.x - right > 18)
+                    break;
+                text = `${text} ${next.text}`;
+                right = Math.max(right, next.x + next.width);
+                phrases.push({
+                    ...first,
+                    text: normalizeText(text),
+                    width: right - first.x,
+                    height: Math.max(first.height, next.height),
+                    hasEol: next.hasEol,
+                });
+            }
+        }
+    }
+    for (const first of original) {
+        let previous = first;
+        let text = first.text;
+        let width = first.width;
+        for (let depth = 0; depth < 2; depth += 1) {
+            const next = original.find((span) => Math.abs(span.x - first.x) <= 3 &&
+                previous.y - span.y > 3 &&
+                previous.y - span.y <= 18);
+            if (next === undefined)
+                break;
+            text = `${text} ${next.text}`;
+            width = Math.max(width, next.width);
+            phrases.push({
+                ...first,
+                text: normalizeText(text),
+                width,
+                hasEol: next.hasEol,
+            });
+            previous = next;
+        }
+    }
+    return phrases;
+}
+export function pageText(page) {
+    return spansInReadingOrder(page)
+        .map((span) => span.text)
+        .join('\n');
+}
+//# sourceMappingURL=text.js.map

package/dist/internal/text.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"text.js","sourceRoot":"","sources":["../../src/internal/text.ts"],"names":[],"mappings":"AAEA,MAAM,UAAU,aAAa,CAAC,KAAa;IACzC,OAAO,KAAK;SACT,SAAS,CAAC,MAAM,CAAC;SACjB,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;SACvB,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC;SACpB,IAAI,EAAE,CAAC;AACZ,CAAC;AAED,MAAM,UAAU,QAAQ,CAAC,KAAa;IACpC,OAAO,aAAa,CAAC,KAAK,CAAC,CAAC,WAAW,EAAE,CAAC;AAC5C,CAAC;AAED,MAAM,UAAU,MAAM,CAAC,MAAyB;IAC9C,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,MAAM,MAAM,GAAa,EAAE,CAAC;IAE5B,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,MAAM,UAAU,GAAG,aAAa,CAAC,KAAK,CAAC,CAAC;QACxC,MAAM,GAAG,GAAG,QAAQ,CAAC,UAAU,CAAC,CAAC;QACjC,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC;YAAE,SAAS;QACvD,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACd,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;IAC1B,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,MAAM,UAAU,mBAAmB,CAAC,IAAmB;IACrD,OAAO,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;QAC1C,MAAM,aAAa,GAAG,KAAK,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC;QACvC,IAAI,IAAI,CAAC,GAAG,CAAC,aAAa,CAAC,GAAG,CAAC;YAAE,OAAO,aAAa,CAAC;QACtD,OAAO,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC;IAC1B,CAAC,CAAC,CAAC;AACL,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,aAAa,CAAC,IAAmB;IAC/C,MAAM,QAAQ,GAAG,mBAAmB,CAAC,IAAI,CAAC,CAAC;IAC3C,MAAM,IAAI,GAAuB,EAAE,CAAC;IAEpC,KAAK,MAAM,IAAI,IAAI,QAAQ,EAAE,CAAC;QAC5B,MAAM,GAAG,GAAG,IAAI,CAAC,IAAI,CACnB,CAAC,SAAS,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CACnE,CAAC;QACF,IAAI,GAAG,KAAK,SAAS;YAAE,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;;YACpC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACtB,CAAC;IAED,MAAM,OAAO,GAAqB,CAAC,GAAG,QAAQ,CAAC,CAAC;IAChD,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,GAAG,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QAC5C,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,GAAG,CAAC,MAAM,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;YACnD,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC;YACzB,IAAI,KAAK,KAAK,SAAS;gBAAE,SAAS;YAClC,IAAI,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC;YACtB,IAAI,KAAK,GAAG,KAAK,CAAC,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC;YAElC,KACE,IAAI,GAAG,GAAG,KAAK,GAAG,CAAC,EACnB,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,MAAM,EAAE,KAAK,GAAG,CAAC,CAAC,EACrC,GAAG,IAAI,CAAC,EACR,CAAC;gBACD,MAAM,IAAI,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC;gBACtB,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,CAAC,GAAG,KAAK,GAAG,EAAE;oBAAE,MAAM;gBACrD,IAAI,GAAG,GAAG,IAAI,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC;gBAC9B,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC;gBAC7C,OAAO,CAAC,IAAI,CAAC;oBACX,GAAG,KAAK;oBACR,IAAI,EAAE,aAAa,CAAC,IAAI,CAAC;oBACzB,KAAK,EAAE,KAAK,GAAG,KAAK,CAAC,CAAC;oBACtB,MAAM,EAAE,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,MAAM,EAAE,IAAI,CAAC,MAAM,CAAC;oBAC3C,MAAM,EAAE,IAAI,CAAC,MAAM;iBACpB,CAAC,CAAC;YACL,CAAC;QACH,CAAC;IACH,CAAC;IAED,KAAK,MAAM,KAAK,IAAI,QAAQ,EAAE,CAAC;QAC7B,IAAI,QAAQ,GAAG,KAAK,CAAC;QACrB,IAAI,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC;QACtB,IAAI,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC;QACxB,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,CAAC,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;YAC1C,MAAM,IAAI,GAAG,QAAQ,CAAC,IAAI,CACxB,CAAC,IAAI,EAAE,EAAE,CACP,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC;gBAC/B,QAAQ,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC;gBACvB,QAAQ,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,EAAE,CAC5B,CAAC;YACF,IAAI,IAAI,KAAK,SAAS;gBAAE,MAAM;YAC9B,IAAI,GAAG,GAAG,IAAI,IAAI,IAAI,CAAC,IAAI,EAAE,CAAC;YAC9B,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC;YACpC,OAAO,CAAC,IAAI,CAAC;gBACX,GAAG,KAAK;gBACR,IAAI,EAAE,aAAa,CAAC,IAAI,CAAC;gBACzB,KAAK;gBACL,MAAM,EAAE,IAAI,CAAC,MAAM;aACpB,CAAC,CAAC;YACH,QAAQ,GAAG,IAAI,CAAC;QAClB,CAAC;IACH,CAAC;IAED,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,MAAM,UAAU,QAAQ,CAAC,IAAmB;IAC1C,OAAO,mBAAmB,CAAC,IAAI,CAAC;SAC7B,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC;SACxB,IAAI,CAAC,IAAI,CAAC,CAAC;AAChB,CAAC"}

package/dist/internal/values.d.ts

@@ -0,0 +1,5 @@
+export declare function parseItalianDate(value: string): string | undefined;
+export declare function parseItalianAmount(value: string): number | undefined;
+export declare function parseItalianCount(value: string): number | undefined;
+export declare function parseItalianBoolean(value: string): boolean | undefined;
+//# sourceMappingURL=values.d.ts.map

package/dist/internal/values.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"values.d.ts","sourceRoot":"","sources":["../../src/internal/values.ts"],"names":[],"mappings":"AAQA,wBAAgB,gBAAgB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAyBlE;AAED,wBAAgB,kBAAkB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAepE;AAED,wBAAgB,iBAAiB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAMnE;AAED,wBAAgB,mBAAmB,CAAC,KAAK,EAAE,MAAM,GAAG,OAAO,GAAG,SAAS,CAKtE"}

package/dist/internal/values.js

@@ -0,0 +1,64 @@
+const ITALIAN_DATE = /^(\d{2})\/(\d{2})\/(\d{4})$/;
+const ITALIAN_AMOUNT = /^(?:\d+|\d{1,3}(?:\.\d{3})+)(?:,\d{1,2})?$/;
+const ITALIAN_INTEGER = /^(?:\d+|\d{1,3}(?:\.\d{3})+)$/;
+function isLeapYear(year) {
+    return year % 4 === 0 && (year % 100 !== 0 || year % 400 === 0);
+}
+export function parseItalianDate(value) {
+    const match = ITALIAN_DATE.exec(value.trim());
+    if (match === null)
+        return undefined;
+    const day = Number(match[1]);
+    const month = Number(match[2]);
+    const year = Number(match[3]);
+    const monthLengths = [
+        31,
+        isLeapYear(year) ? 29 : 28,
+        31,
+        30,
+        31,
+        30,
+        31,
+        31,
+        30,
+        31,
+        30,
+        31,
+    ];
+    const maximumDay = monthLengths[month - 1];
+    if (maximumDay === undefined || day < 1 || day > maximumDay)
+        return undefined;
+    return `${String(year).padStart(4, '0')}-${String(month).padStart(2, '0')}-${String(day).padStart(2, '0')}`;
+}
+export function parseItalianAmount(value) {
+    const normalized = value.trim();
+    if (!ITALIAN_AMOUNT.test(normalized))
+        return undefined;
+    const [integerPart = '', decimalPart = ''] = normalized.split(',');
+    const centsText = `${integerPart.replaceAll('.', '')}${decimalPart.padEnd(2, '0')}`;
+    const cents = Number(centsText);
+    if (!Number.isSafeInteger(cents))
+        return undefined;
+    const amount = cents / 100;
+    // JSON uses the shortest decimal representation. Reject a value when that
+    // representation would change a source cent at the upper end of the range.
+    const [whole = '', fraction = ''] = String(amount).split('.');
+    const serializedCents = BigInt(`${whole}${fraction.padEnd(2, '0')}`);
+    return serializedCents === BigInt(centsText) ? amount : undefined;
+}
+export function parseItalianCount(value) {
+    const normalized = value.trim();
+    if (!ITALIAN_INTEGER.test(normalized))
+        return undefined;
+    const count = Number(normalized.replaceAll('.', ''));
+    return Number.isSafeInteger(count) ? count : undefined;
+}
+export function parseItalianBoolean(value) {
+    const normalized = value.trim().toLowerCase();
+    if (normalized === 'si' || normalized === 'sì')
+        return true;
+    if (normalized === 'no')
+        return false;
+    return undefined;
+}
+//# sourceMappingURL=values.js.map

package/dist/internal/values.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"values.js","sourceRoot":"","sources":["../../src/internal/values.ts"],"names":[],"mappings":"AAAA,MAAM,YAAY,GAAG,6BAA6B,CAAC;AACnD,MAAM,cAAc,GAAG,4CAA4C,CAAC;AACpE,MAAM,eAAe,GAAG,+BAA+B,CAAC;AAExD,SAAS,UAAU,CAAC,IAAY;IAC9B,OAAO,IAAI,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,GAAG,GAAG,KAAK,CAAC,IAAI,IAAI,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC;AAClE,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,KAAa;IAC5C,MAAM,KAAK,GAAG,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC;IAC9C,IAAI,KAAK,KAAK,IAAI;QAAE,OAAO,SAAS,CAAC;IAErC,MAAM,GAAG,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC7B,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC/B,MAAM,IAAI,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAC9B,MAAM,YAAY,GAAG;QACnB,EAAE;QACF,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE;QAC1B,EAAE;QACF,EAAE;QACF,EAAE;QACF,EAAE;QACF,EAAE;QACF,EAAE;QACF,EAAE;QACF,EAAE;QACF,EAAE;QACF,EAAE;KACH,CAAC;IACF,MAAM,UAAU,GAAG,YAAY,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;IAC3C,IAAI,UAAU,KAAK,SAAS,IAAI,GAAG,GAAG,CAAC,IAAI,GAAG,GAAG,UAAU;QAAE,OAAO,SAAS,CAAC;IAE9E,OAAO,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,CAAC,EAAE,GAAG,CAAC,IAAI,MAAM,CAAC,KAAK,CAAC,CAAC,QAAQ,CAAC,CAAC,EAAE,GAAG,CAAC,IAAI,MAAM,CAAC,GAAG,CAAC,CAAC,QAAQ,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC;AAC9G,CAAC;AAED,MAAM,UAAU,kBAAkB,CAAC,KAAa;IAC9C,MAAM,UAAU,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;IAChC,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,UAAU,CAAC;QAAE,OAAO,SAAS,CAAC;IAEvD,MAAM,CAAC,WAAW,GAAG,EAAE,EAAE,WAAW,GAAG,EAAE,CAAC,GAAG,UAAU,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IACnE,MAAM,SAAS,GAAG,GAAG,WAAW,CAAC,UAAU,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,WAAW,CAAC,MAAM,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC;IACpF,MAAM,KAAK,GAAG,MAAM,CAAC,SAAS,CAAC,CAAC;IAChC,IAAI,CAAC,MAAM,CAAC,aAAa,CAAC,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IAEnD,MAAM,MAAM,GAAG,KAAK,GAAG,GAAG,CAAC;IAC3B,0EAA0E;IAC1E,2EAA2E;IAC3E,MAAM,CAAC,KAAK,GAAG,EAAE,EAAE,QAAQ,GAAG,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAC9D,MAAM,eAAe,GAAG,MAAM,CAAC,GAAG,KAAK,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;IACrE,OAAO,eAAe,KAAK,MAAM,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,SAAS,CAAC;AACpE,CAAC;AAED,MAAM,UAAU,iBAAiB,CAAC,KAAa;IAC7C,MAAM,UAAU,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;IAChC,IAAI,CAAC,eAAe,CAAC,IAAI,CAAC,UAAU,CAAC;QAAE,OAAO,SAAS,CAAC;IAExD,MAAM,KAAK,GAAG,MAAM,CAAC,UAAU,CAAC,UAAU,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC;IACrD,OAAO,MAAM,CAAC,aAAa,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAAC;AACzD,CAAC;AAED,MAAM,UAAU,mBAAmB,CAAC,KAAa;IAC/C,MAAM,UAAU,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;IAC9C,IAAI,UAAU,KAAK,IAAI,IAAI,UAAU,KAAK,IAAI;QAAE,OAAO,IAAI,CAAC;IAC5D,IAAI,UAAU,KAAK,IAAI;QAAE,OAAO,KAAK,CAAC;IACtC,OAAO,SAAS,CAAC;AACnB,CAAC"}

package/dist/parse-visura.d.ts

@@ -0,0 +1,3 @@
+import type { ParseVisuraOptions, VisuraDocument } from './types.js';
+export declare function parseVisura(input: Uint8Array, options?: ParseVisuraOptions): Promise<VisuraDocument>;
+//# sourceMappingURL=parse-visura.d.ts.map

package/dist/parse-visura.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"parse-visura.d.ts","sourceRoot":"","sources":["../src/parse-visura.ts"],"names":[],"mappings":"AAMA,OAAO,KAAK,EAAE,kBAAkB,EAAE,cAAc,EAAE,MAAM,YAAY,CAAC;AAErE,wBAAsB,WAAW,CAC/B,KAAK,EAAE,UAAU,EACjB,OAAO,GAAE,kBAAuB,GAC/B,OAAO,CAAC,cAAc,CAAC,CAsBzB"}

package/dist/parse-visura.js

@@ -0,0 +1,17 @@
+import { extractPdf } from './internal/extract-pdf.js';
+import { isSupportedVisuraBlock, parseExtractedDocument, } from './internal/parse-document.js';
+import { VisuraParseError } from './errors.js';
+export async function parseVisura(input, options = {}) {
+    const pdf = await extractPdf(input);
+    const result = parseExtractedDocument(pdf, options.filename);
+    if (result.reportType === undefined && !isSupportedVisuraBlock(pdf)) {
+        throw new VisuraParseError('unsupported-pdf', 'The PDF is text-based but is not a recognized Visura camerale.');
+    }
+    if (result.companyName === undefined &&
+        result.taxCode === undefined &&
+        result.reaNumber === undefined) {
+        throw new VisuraParseError('malformed-content', 'The Visura type was recognized, but no subject identity could be parsed.');
+    }
+    return result;
+}
+//# sourceMappingURL=parse-visura.js.map

package/dist/parse-visura.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"parse-visura.js","sourceRoot":"","sources":["../src/parse-visura.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,2BAA2B,CAAC;AACvD,OAAO,EACL,sBAAsB,EACtB,sBAAsB,GACvB,MAAM,8BAA8B,CAAC;AACtC,OAAO,EAAE,gBAAgB,EAAE,MAAM,aAAa,CAAC;AAG/C,MAAM,CAAC,KAAK,UAAU,WAAW,CAC/B,KAAiB,EACjB,UAA8B,EAAE;IAEhC,MAAM,GAAG,GAAG,MAAM,UAAU,CAAC,KAAK,CAAC,CAAC;IACpC,MAAM,MAAM,GAAG,sBAAsB,CAAC,GAAG,EAAE,OAAO,CAAC,QAAQ,CAAC,CAAC;IAE7D,IAAI,MAAM,CAAC,UAAU,KAAK,SAAS,IAAI,CAAC,sBAAsB,CAAC,GAAG,CAAC,EAAE,CAAC;QACpE,MAAM,IAAI,gBAAgB,CACxB,iBAAiB,EACjB,gEAAgE,CACjE,CAAC;IACJ,CAAC;IACD,IACE,MAAM,CAAC,WAAW,KAAK,SAAS;QAChC,MAAM,CAAC,OAAO,KAAK,SAAS;QAC5B,MAAM,CAAC,SAAS,KAAK,SAAS,EAC9B,CAAC;QACD,MAAM,IAAI,gBAAgB,CACxB,mBAAmB,EACnB,0EAA0E,CAC3E,CAAC;IACJ,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC"}