cry-bizisi-parser 0.2.0

package/src/bizisi.ts

@@ -0,0 +1,720 @@
+/**
+ * bizi.si scraper — čisti Bun fetch, brez brskalnika.
+ *
+ * bizi.si uporablja ASP.NET WebForms, zato je treba:
+ * 1. Shranjevati piškotke (ASP.NET_SessionId, .ASPXAUTH)
+ * 2. Prenašati __VIEWSTATE in __EVENTVALIDATION med GET in POST
+ *
+ * Odvisnosti: 0 — samo Bun built-in API-ji.
+ */
+
+import { createWorker, type Worker } from "tesseract.js";
+
+import type {
+  CompanySearchResult,
+  CompanyDetail,
+  FinancialData,
+  BankAccount,
+  KeyPeople,
+  ScrapeResult,
+} from "./types.js";
+
+// ---------------------------------------------------------------------------
+// Config
+// ---------------------------------------------------------------------------
+
+const BASE_URL = "https://www.bizi.si";
+
+// ---------------------------------------------------------------------------
+// Cookie jar — ročno upravljanje prek Set-Cookie / Cookie
+// ---------------------------------------------------------------------------
+
+class CookieJar {
+  private cookies = new Map<string, { value: string; domain: string; path: string }>();
+
+  /** Prebere Set-Cookie headerje in shrani (Node.js + Bun). */
+  setFromHeaders(headers: Headers): void {
+    // Node.js 19+ ima getSetCookie(), Bun ima getAll()
+    let setCookieValues: string[];
+    if (typeof (headers as any).getSetCookie === "function") {
+      setCookieValues = (headers as any).getSetCookie();
+    } else if (typeof (headers as any).getAll === "function") {
+      const raw = (headers as any).getAll("Set-Cookie");
+      setCookieValues = Array.isArray(raw) ? raw : [raw];
+    } else {
+      setCookieValues = [];
+      for (const [k, v] of headers as any) {
+        if (k.toLowerCase() === "set-cookie") setCookieValues.push(v);
+      }
+    }
+
+    for (const entry of setCookieValues) {
+      if (!entry) continue;
+      const [nv, ...rest] = entry.split(";");
+      if (!nv) continue;
+      const eq = nv.indexOf("=");
+      if (eq < 0) continue;
+      const name = nv.substring(0, eq).trim();
+      const value = nv.substring(eq + 1).trim();
+
+      let domain = ".bizi.si";
+      let path = "/";
+      for (const attr of rest) {
+        const a = attr.trim();
+        if (a.toLowerCase().startsWith("domain=")) domain = a.split("=")[1] ?? domain;
+        if (a.toLowerCase().startsWith("path=")) path = a.split("=")[1] ?? path;
+        if (a.toLowerCase() === "max-age=0" || a.toLowerCase().startsWith("expires=")) {
+          // Could handle expiry but not critical
+        }
+      }
+      this.cookies.set(name, { value, domain, path });
+    }
+  }
+
+  /** Vrne Cookie header za dani URL. */
+  getRequestHeader(): string {
+    const parts: string[] = [];
+    for (const [name, c] of this.cookies) {
+      if (c.domain && ".bizi.si".endsWith(c.domain)) {
+        parts.push(`${name}=${c.value}`);
+      }
+    }
+    return parts.join("; ");
+  }
+
+  /** Izbriši vse piškotke. */
+  clear(): void {
+    this.cookies.clear();
+  }
+}
+
+// ---------------------------------------------------------------------------
+// HTTP helper — fetch s piškotki
+// ---------------------------------------------------------------------------
+
+async function fetchWithCookies(
+  url: string,
+  jar: CookieJar,
+  init: RequestInit = {},
+): Promise<Response> {
+  const headers = new Headers(init.headers);
+  // Default headers
+  if (!headers.has("User-Agent")) {
+    headers.set(
+      "User-Agent",
+      "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/125.0.0.0 Safari/537.36",
+    );
+  }
+  if (!headers.has("Accept")) {
+    headers.set("Accept", "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8");
+  }
+  if (!headers.has("Accept-Language")) {
+    headers.set("Accept-Language", "sl-SI,sl;q=0.9,en;q=0.8");
+  }
+
+  const cookie = jar.getRequestHeader();
+  if (cookie) headers.set("Cookie", cookie);
+
+  if (init.body && !headers.has("Content-Type")) {
+    headers.set("Content-Type", "application/x-www-form-urlencoded");
+  }
+
+  const res = await fetch(url, { ...init, headers, redirect: "manual" });
+
+  // Shrani nove piškotke
+  jar.setFromHeaders(res.headers);
+
+  // Sledi redirectom ročno (da ohranimo piškotke)
+  if (res.status >= 300 && res.status < 400) {
+    const location = res.headers.get("location");
+    if (location) {
+      const redirectUrl = location.startsWith("http") ? location : `${new URL(url).origin}${location}`;
+      return fetchWithCookies(redirectUrl, jar, { method: "GET" });
+    }
+  }
+
+  return res;
+}
+
+// ---------------------------------------------------------------------------
+// ASP.NET WebForms helperji
+// ---------------------------------------------------------------------------
+
+/** Izlušči __VIEWSTATE, __EVENTVALIDATION itd. iz HTML. */
+function extractAspFormFields(html: string): Record<string, string> {
+  const fields: Record<string, string> = {};
+  const regex = /<input[^>]*(?:name|id)="(?:__VIEWSTATE|__EVENTVALIDATION|__VIEWSTATEGENERATOR)"[^>]*\/?>/gi;
+  let m: RegExpExecArray | null;
+  while ((m = regex.exec(html)) !== null) {
+    const nameMatch = m[0].match(/name="([^"]+)"/);
+    const valueMatch = m[0].match(/value="([^"]*)"/);
+    if (nameMatch) {
+      fields[nameMatch[1]!] = valueMatch ? valueMatch[1]! : "";
+    }
+  }
+  return fields;
+}
+
+/** Build URL-encoded form body. */
+function encodeForm(data: Record<string, string>): string {
+  return Object.entries(data)
+    .map(([k, v]) => `${encodeURIComponent(k)}=${encodeURIComponent(v)}`)
+    .join("&");
+}
+
+// ---------------------------------------------------------------------------
+// Parser helperji
+// ---------------------------------------------------------------------------
+
+function normalize(str: string): string {
+  return str.replace(/\s+/g, " ").trim();
+}
+
+/** Pretvori slovenski datum "25. 3. 2015" v ISO "2015-03-25". */
+function formatDate(raw: string): string {
+  const m = raw.match(/(\d{1,2})\.\s*(\d{1,2})\.\s*(\d{4})/);
+  if (!m) return raw;
+  const d = m[1]!.padStart(2, "0");
+  const mo = m[2]!.padStart(2, "0");
+  const y = m[3]!;
+  return `${y}-${mo}-${d}`;
+}
+
+function parseEmployees(raw: string): number {
+  const s = raw.trim();
+  if (!s || /^(ni|brez|ne|\/)/i.test(s)) return 0;
+  const range = s.match(/(\d+)\s*(?:ali|do|-|–|,)\s*(\d+)/i);
+  if (range) return (parseInt(range[1] ?? "0", 10) + parseInt(range[2] ?? "0", 10)) / 2;
+  const single = s.match(/(\d+)/);
+  return single ? parseInt(single[1] ?? "0", 10) : 0;
+}
+
+function parseSlovenianNumber(raw: string | null): number | null {
+  if (!raw) return null;
+  let s = raw.trim();
+  if (s === "" || s === "-") return null;
+  if (s === "0") return 0;
+  s = s.replace(/\./g, "").replace(/,/, ".");
+  const n = parseFloat(s);
+  return Number.isFinite(n) ? n : null;
+}
+
+/** Poišči besedilo med dvema oznakama v HTML (brez HTML tagov). */
+function extractBetween(html: string, startMarker: string, endMarker: string): string {
+  const s = html.indexOf(startMarker);
+  if (s < 0) return "";
+  const start = s + startMarker.length;
+  const e = endMarker ? html.indexOf(endMarker, start) : html.length;
+  if (e < 0) return html.substring(start).trim();
+  return html.substring(start, e).trim();
+}
+
+/** Odstrani HTML tage. */
+function stripTags(html: string): string {
+  return html.replace(/<[^>]+>/g, " ").replace(/&nbsp;/g, " ").replace(/&amp;/g, "&");
+}
+
+/** Poišči tabelo v HTML in vrni vrstice kot array celic. */
+function parseTable(html: string, tableIdx = 0): string[][] {
+  const tables: string[][] = [];
+  const tableRegex = /<table[^>]*>([\s\S]*?)<\/table>/gi;
+  let tMatch: RegExpExecArray | null;
+  while ((tMatch = tableRegex.exec(html)) !== null) {
+    const rows: string[][] = [];
+    const rowRegex = /<tr[^>]*>([\s\S]*?)<\/tr>/gi;
+    let rMatch: RegExpExecArray | null;
+    while ((rMatch = rowRegex.exec(tMatch[1]!)) !== null) {
+      const cells: string[] = [];
+      const cellRegex = /<t[dh][^>]*>([\s\S]*?)<\/t[dh]>/gi;
+      let cMatch: RegExpExecArray | null;
+      while ((cMatch = cellRegex.exec(rMatch[1]!)) !== null) {
+        cells.push(normalize(stripTags(cMatch[1]!)));
+      }
+      if (cells.length > 0) rows.push(cells);
+    }
+    if (rows.length > 0) tables.push(rows.flat());
+  }
+  return tables;
+}
+
+// ---------------------------------------------------------------------------
+// FetchBizisiScraper
+// ---------------------------------------------------------------------------
+
+export class BizisiScraper {
+  private jar = new CookieJar();
+  private _loggedIn = false;
+  private ocrWorker: Worker | null = null;
+
+  /** Vrne (ali ustvari) OCR worker. */
+  async #getOcrWorker(): Promise<Worker> {
+    if (!this.ocrWorker) {
+      this.ocrWorker = await createWorker("slv", 1, {
+        logger: () => {}, // utišaj log
+      });
+    }
+    return this.ocrWorker;
+  }
+
+  /** Počisti OCR worker (pokliči ko končaš). */
+  async close(): Promise<void> {
+    if (this.ocrWorker) {
+      try { await this.ocrWorker.terminate(); } catch {}
+      this.ocrWorker = null;
+    }
+  }
+
+  /** OCR ene slike — vrne prepoznano besedilo. */
+  async #ocrImage(url: string): Promise<string> {
+    try {
+      const resp = await fetch(url);
+      const buf = Buffer.from(await resp.arrayBuffer());
+      const worker = await this.#getOcrWorker();
+      const { data } = await worker.recognize(buf);
+      return data.text.trim();
+    } catch {
+      return url; // fallback na URL
+    }
+  }
+
+  get loggedIn(): boolean {
+    return this._loggedIn;
+  }
+
+  // -----------------------------------------------------------------------
+  // Login
+  // -----------------------------------------------------------------------
+
+  async login(username?: string, password?: string): Promise<void> {
+    const user = username ?? process.env.BIZISI_USERNAME;
+    const pass = password ?? process.env.BIZISI_PASSWORD;
+    if (!user || !pass) throw new Error("Missing BIZISI_USERNAME or BIZISI_PASSWORD");
+
+    // 1. GET login page — poberi __VIEWSTATE itd.
+    const loginPage = await fetchWithCookies(`${BASE_URL}/prijava/`, this.jar);
+    const html = await loginPage.text();
+
+    const aspFields = extractAspFormFields(html);
+    if (!aspFields["__VIEWSTATE"]) {
+      throw new Error("Could not find __VIEWSTATE on login page");
+    }
+
+    // 2. POST login form
+    const formData: Record<string, string> = {
+      ...aspFields,
+      "ctl00$cphMain$loginBox1$UserName": user,
+      "ctl00$cphMain$loginBox1$Password": pass,
+      "ctl00$cphMain$loginBox1$ButtonLogin": "Prijavi se",
+      "ctl00$cphMain$loginBox1$cvUN": "",
+      "ctl00$cphMain$loginBox1$cvPW": "",
+      __LASTFOCUS: "",
+      __SCROLLPOSITION: "",
+    };
+
+    const loginRes = await fetchWithCookies(`${BASE_URL}/prijava/?rw=1`, this.jar, {
+      method: "POST",
+      body: encodeForm(formData),
+      headers: {
+        Referer: `${BASE_URL}/prijava/`,
+        Origin: BASE_URL,
+      },
+    });
+
+    const loginHtml = await loginRes.text();
+
+    // Preveri, če je prijava uspela
+    if (loginHtml.includes("Prijavi se") && !loginHtml.includes("Moj Bizi")) {
+      if (/napačno|neveljaven|Presegli/i.test(loginHtml)) {
+        throw new Error("Login failed: invalid credentials or rate-limited");
+      }
+      throw new Error("Login failed: could not verify");
+    }
+
+    this._loggedIn = true;
+  }
+
+  // -----------------------------------------------------------------------
+  // Search
+  // -----------------------------------------------------------------------
+
+  async search(query: string): Promise<CompanySearchResult[]> {
+    const res = await fetchWithCookies(
+      `${BASE_URL}/iskanje?q=${encodeURIComponent(query)}`,
+      this.jar,
+    );
+    const html = await res.text();
+
+    return this.#parseSearchResults(html);
+  }
+
+  #parseSearchResults(html: string): CompanySearchResult[] {
+    const results: CompanySearchResult[] = [];
+
+    // Najdi vrstice rezultatov: div.row.b-table-row z id-jem, ki vsebuje RepeaterResults
+    const rowRegex =
+      /<div[^>]*id="ctl00_cphMain_DisplayRecords1_RepeaterResults_[^"]*"[^>]*class="[^"]*row b-table-row[^"]*"[^>]*>([\s\S]*?)<\/div>\s*<\/div>/gi;
+    let rowMatch: RegExpExecArray | null;
+
+    while ((rowMatch = rowRegex.exec(html)) !== null) {
+      const rowHtml = rowMatch[1]!;
+
+      // Company link
+      const linkMatch = rowHtml.match(
+        /<a[^>]*class="[^"]*b-link-company[^"]*"[^>]*href="([^"]*)"[^>]*>[\s\S]*?<span[^>]*class="b-company-title"[^>]*>([\s\S]*?)<\/span>/i,
+      );
+      if (!linkMatch) continue;
+
+      const href = linkMatch[1]!;
+      const name = normalize(stripTags(linkMatch[2]!));
+
+      // Celice: div.col.b-table-cell
+      const cells = Array.from(rowHtml.matchAll(/<div[^>]*class="[^"]*col b-table-cell[^"]*"[^>]*>([\s\S]*?)<\/div>/gi));
+
+      const getCell = (idx: number): string => {
+        if (idx >= cells.length) return "";
+        return normalize(stripTags(cells[idx]![1]!));
+      };
+
+      // Cells: 0=checkbox, 1=name, 2=address, 3=city, 4=reg#, 5=tax#, 6=activity, 7=employees
+      const url = href.startsWith("http") ? "/" + href.split("/").slice(3).join("/") : href;
+
+      results.push({
+        name,
+        url: url.replace(/\/+$/, "") + "/",
+        address: getCell(2),
+        city: getCell(3),
+        registrationNumber: getCell(4),
+        taxNumber: getCell(5),
+        activity: getCell(6),
+        employees: parseEmployees(getCell(7)),
+      });
+    }
+
+    return results;
+  }
+
+  // -----------------------------------------------------------------------
+  // Company detail
+  // -----------------------------------------------------------------------
+
+  async getCompanyDetail(relativeUrl: string): Promise<CompanyDetail> {
+    const url = `${BASE_URL}${relativeUrl.startsWith("/") ? "" : "/"}${relativeUrl}`;
+    const res = await fetchWithCookies(url, this.jar);
+    const html = await res.text();
+
+    // Odstrani skripte in stile za čistejše regex ujemanje
+    const cleanHtml = html
+      .replace(/<script[\s\S]*?<\/script>/gi, "")
+      .replace(/<style[\s\S]*?<\/style>/gi, "")
+      .replace(/&nbsp;/g, " ")
+      .replace(/&amp;/g, "&");
+
+    const fullText = normalize(stripTags(cleanHtml));
+
+    // Poišči položaj h1 — za njim je glavna vsebina, pred njim sidebar
+    const h1Start = cleanHtml.indexOf('<h1');
+    const afterH1Html = h1Start >= 0 ? cleanHtml.substring(h1Start) : cleanHtml;
+
+    // Helper: extractBetween na HTML-ju za h1
+    const extractAfterH1 = (start: string, end: string): string => {
+      const s = afterH1Html.indexOf(start);
+      if (s < 0) return '';
+      const begin = s + start.length;
+      const e = end ? afterH1Html.indexOf(end, begin) : afterH1Html.length;
+      if (e < 0) return afterH1Html.substring(begin).trim();
+      return afterH1Html.substring(begin, e).trim();
+    };
+
+    // --- Header info ---
+    const h1Match = cleanHtml.match(/<h1[^>]*>([\s\S]*?)<\/h1>/i);
+    const h1Name = h1Match ? normalize(stripTags(h1Match[1]!)) : "";
+    const titleMatch = cleanHtml.match(/<title[^>]*>([\s\S]*?)<\/title>/i);
+    const title = titleMatch ? titleMatch[1]!.replace(/ - .*$/, "").trim() : "";
+    const name = h1Name || title;
+
+    // Phone: poišči tel: povezavo
+    const phoneMatch = cleanHtml.match(/<a[^>]*href="tel:([^"]*)"[^>]*>([\s\S]*?)<\/a>/i);
+    const phone = phoneMatch ? normalize(phoneMatch[2] ?? phoneMatch[1]!) : "";
+
+    // Email: poišči mailto: povezavo
+    const emailMatch = cleanHtml.match(/<a[^>]*href="mailto:([^"]*)"[^>]*>([\s\S]*?)<\/a>/i);
+    const email = emailMatch ? normalize(emailMatch[2] ?? emailMatch[1]!) : "";
+
+    // Website: poišči v header delu (za h1, pred h5) link z URL-jem kot besedilom
+    let website = "";
+    const headerHtml = extractAfterH1("</h1>", "Matično podjetje");
+    // Poisci <a> brez href (aspNetDisabled) kjer je text URL, ali <a> s href
+    const headerLinks = Array.from(headerHtml.matchAll(/<a[^>]*>([\s\S]*?)<\/a>/gi));
+    for (const link of headerLinks) {
+      const text = normalize(link[1]!);
+      if (text.startsWith("http") && !text.includes("bizi.si")) {
+        website = text;
+        break;
+      }
+    }
+
+    // Address: vzemi samo vrstico z naslovom (vsebuje številko in poštno številko)
+    let address = "";
+    const addrRaw = extractAfterH1("</h1>", '<h5 class="b-box-title');
+    if (addrRaw) {
+      const addrText = normalize(stripTags(addrRaw));
+      // Vzemi besedilo do prvega emaila, telefona (3 cifre + presledek + 3-6 cifre), "Več" ali "Dodaj"
+      const upToMarker = addrText.split(
+        /\s+\S+@\S+|\s+\d{3}\s\d{3,6}|\s+Več\s+|\s+Dodaj\s+|\s+http/i,
+      )[0]?.trim() ?? '';
+      // Izlušči del s poštno številko (4-mestna)
+      const m = upToMarker.match(/([A-ZČŠŽa-zčšž][^,]+\d+[^,]*,\s*[^,]+,\s*\d{4}\s+[A-ZČŠŽa-zčšž][^,]+)/);
+      if (m) address = m[1]!.trim();
+    }
+
+    // --- Matično podjetje section (only after h1) ---
+    const maticnoHtml = extractAfterH1("Matično podjetje", "Ključne osebe");
+    const maticnoText = normalize(stripTags(maticnoHtml));
+
+    const davcnaM = maticnoText.match(/Davčna številka.*?(\d{7,9})/i);
+    const maticnaM = maticnoText.match(/Matična.*?(\d{7,10})/i);
+    const ddvM = maticnoText.match(/Zavezanec za DDV.*?(Da|Ne)/i);
+    const skisM = maticnoText.match(/SKIS\s*:?\s*([^\n]+?)(?:\s+(?:Datum|Vir))/i);
+    const vpisM = maticnoText.match(/Datum vpisa\s*:?\s*([^\n]+?)(?:\s+(?:Dejavnost|Vir))/i);
+    const dejavnostM = maticnoText.match(/Dejavnost TSmedia\s*:?\s*([^\n]+?)(?:\s+Vir)/i);
+
+    const taxNumber = davcnaM ? davcnaM[1]! : "";
+    const registrationNumber = maticnaM ? maticnaM[1]! : "";
+    const vatPayer = ddvM ? ddvM[1]!.toLowerCase() === "da" : false;
+    const vatNumM = fullText.match(/SI(\d{7,9})/);
+    const vatNumber = vatNumM ? "SI" + vatNumM[1]! : "";
+    const skis = skisM ? skisM[1]!.trim() : "";
+    const dateOfEntry = vpisM ? formatDate(vpisM[1]!.trim()) : "";
+    const activity = dejavnostM ? dejavnostM[1]!.trim() : "";
+
+    // --- TRR (only after h1) ---
+    const trrHtml = extractAfterH1("TRR in blokade", "Bizi obveščevalec");
+    const trrText = normalize(stripTags(trrHtml));
+    const bankAccounts: BankAccount[] = [];
+    const ibanRegex = /(SI\d{2}\s*\d{4}\s*\d{4}\s*\d{4}\s*\d{3})/g;
+    let ibanM: RegExpExecArray | null;
+    while ((ibanM = ibanRegex.exec(trrText)) !== null) {
+      bankAccounts.push({
+        iban: ibanM[1]!.replace(/\s+/g, " ").trim(),
+        status: "odprt",
+        openedDate: null,
+        bank: null,
+      });
+    }
+
+    // --- Key people (only after h1, before Bonitetna ocena) ---
+    const osebeHtml = extractAfterH1("Ključne osebe", "Bonitetna ocena");
+    const keyPeople = await this.#ocrKeyPeople(this.#parseKeyPeople(osebeHtml));
+
+    // --- Financiranje iz proračuna (only after h1) ---
+    const finHtml = extractAfterH1("Financiranje iz proračuna", "Ne spreglejte");
+    const finText = normalize(stripTags(finHtml));
+    const financingFromBudget = this.#parseFinancing(finText);
+
+    return {
+      name,
+      fullAddress: address || "",
+      phone,
+      website,
+      email,
+      taxNumber,
+      registrationNumber,
+      vatPayer,
+      vatNumber,
+      skis,
+      dateOfEntry,
+      activity,
+      keyPeople,
+      bankAccounts,
+      financingFromBudget,
+    };
+  }
+
+  // -----------------------------------------------------------------------
+  // Financial data
+  // -----------------------------------------------------------------------
+
+  async getFinancialData(relativeUrl: string): Promise<FinancialData[]> {
+    const basePath = relativeUrl.replace(/\/+$/, "");
+    const finUrl = `${BASE_URL}${basePath}/financni-podatki/`;
+
+    const res = await fetchWithCookies(finUrl, this.jar);
+    const html = await res.text();
+
+    // Preveri paywall
+    if (/registriraj.*se.*brezplačno|naroči.*polni.*dostop|preizkusi/i.test(html)) {
+      return [];
+    }
+
+    const cleanHtml = html
+      .replace(/<script[\s\S]*?<\/script>/gi, "")
+      .replace(/<style[\s\S]*?<\/style>/gi, "");
+
+    // Poišči tabelo s finančnimi podatki
+    const tables = parseTable(cleanHtml);
+    const data: FinancialData[] = [];
+
+    for (const row of tables) {
+      const joined = row.join(" ");
+      const yearMatch = joined.match(/\b(20[2-9]\d)\b/);
+      if (!yearMatch) continue;
+
+      const year = parseInt(yearMatch[1]!, 10);
+      // Prvi večji znesek = čisti prihodki, drugi = poslovni izzid
+      const amounts = row
+        .map((c) => parseSlovenianNumber(c))
+        .filter((n): n is number => n !== null && n > 0);
+
+      data.push({
+        year,
+        netSalesRevenue: amounts[0] ?? null,
+        operatingResult: amounts[1] ?? null,
+      });
+    }
+
+    return data;
+  }
+
+  // -----------------------------------------------------------------------
+  // Scrape
+  // -----------------------------------------------------------------------
+
+  async scrapeCompany(relativeUrl: string): Promise<ScrapeResult> {
+    const company = await this.getCompanyDetail(relativeUrl);
+    const financialData = await this.getFinancialData(relativeUrl);
+    return { company, financialData };
+  }
+
+  async scrape(
+    query: string,
+    options?: { maxResults?: number; includeFinancialData?: boolean },
+  ): Promise<ScrapeResult[]> {
+    const maxResults = options?.maxResults ?? 0;
+    const includeFinancial = options?.includeFinancialData ?? true;
+
+    const results = await this.search(query);
+    const limited = maxResults > 0 ? results.slice(0, maxResults) : results;
+
+    const output: ScrapeResult[] = [];
+    for (const result of limited) {
+      try {
+        const scraped = await this.scrapeCompany(result.url);
+        output.push(scraped);
+      } catch (err) {
+        console.error(`Failed to scrape ${result.name}:`, err);
+        output.push({
+          company: {
+            name: result.name,
+            fullAddress: "",
+            phone: "",
+            website: "",
+            email: "",
+            taxNumber: result.taxNumber,
+            registrationNumber: result.registrationNumber,
+            vatPayer: false,
+            vatNumber: "",
+            skis: "",
+            dateOfEntry: "",
+            activity: result.activity,
+            keyPeople: { nadzorniki: [], ustanovitelji: [], zastopniki: [] },
+            bankAccounts: [],
+            financingFromBudget: {},
+          },
+          financialData: [],
+        });
+      }
+    }
+
+    return output;
+  }
+
+  // -----------------------------------------------------------------------
+  // Private helpers
+  // -----------------------------------------------------------------------
+
+  /** Statusna sporočila, ki niso imena oseb. */
+  static #statusPhrases = [
+    'ni vpisan', 'ni vpisanih', 'ni aktivnih',
+    'trenutno ni', 'ni podatka', 'ni zabeleženih',
+    'ni najdenih', 'brez podatka',
+  ];
+
+  static #isStatus(text: string): boolean {
+    const lower = text.toLowerCase();
+    return BizisiScraper.#statusPhrases.some((p) => lower.includes(p));
+  }
+
+  /**
+   * Razčleni HTML sekcije "Ključne osebe" in vrne KeyPeople z ImageGenerator URL-ji.
+   * Vsaka oseba je v svojem <span class="...b-attr-value...">.
+   */
+  #parseKeyPeople(html: string): KeyPeople {
+    const result: KeyPeople = { nadzorniki: [], ustanovitelji: [], zastopniki: [] };
+    if (!html) return result;
+
+    const cats = [
+      { label: 'Nadzorniki:', key: 'nadzorniki' as const },
+      { label: 'Ustanovitelji:', key: 'ustanovitelji' as const },
+      { label: 'Zastopniki:', key: 'zastopniki' as const },
+    ];
+
+    for (let ci = 0; ci < cats.length; ci++) {
+      const { label, key } = cats[ci]!;
+      const labelIdx = html.indexOf(label);
+      if (labelIdx < 0) continue;
+
+      const nextLabel = ci + 1 < cats.length ? cats[ci + 1]!.label : null;
+      const nextIdx = nextLabel ? html.indexOf(nextLabel, labelIdx + label.length) : html.length;
+      const sectionHtml = html.substring(labelIdx + label.length, nextIdx > 0 ? nextIdx : html.length);
+
+      // Vsak <span class="...b-attr-value..."> je ena oseba z 1+ slikami
+      const personSpans = Array.from(
+        sectionHtml.matchAll(/<span[^>]*class="[^"]*b-attr-value[^"]*"[^>]*>([\s\S]*?)<\/span>/gi),
+      );
+
+      for (const span of personSpans) {
+        const imgs = Array.from(span[1]!.matchAll(/<img[^>]*src="([^"]*ImageGenerator[^"]*)"[^>]*>/gi));
+        const urls = imgs.map((m) => m[1]!);
+        if (urls.length > 0) result[key].push(urls.join('|||')); // delimiter za OCR
+      }
+    }
+
+    return result;
+  }
+
+  async #ocrKeyPeople(raw: KeyPeople): Promise<KeyPeople> {
+    const result: KeyPeople = { nadzorniki: [], ustanovitelji: [], zastopniki: [] };
+
+    for (const group of ['nadzorniki', 'ustanovitelji', 'zastopniki'] as const) {
+      for (const entry of raw[group]) {
+        const urls = entry.split('|||');
+        if (urls.length === 0) continue;
+
+        const texts = await Promise.all(urls.map((url: string) => this.#ocrImage(url)));
+        const joined = texts.join(' ').trim();
+
+        if (!joined || BizisiScraper.#isStatus(joined)) continue;
+
+        // Odstrani vloge (direktor, prokurist, ...)
+        const cleaned = joined.replace(/\s*,?\s*(?:direktor|prokurist|član|predsednik|namestnik|vodja)\s*,?\s*/gi, '').trim();
+        if (cleaned) result[group].push(cleaned);
+      }
+    }
+
+    return result;
+  }
+
+  #parseFinancing(text: string): Record<number, number> {
+    const result: Record<number, number> = {};
+    const regex = /\b(20[0-9]{2})\D{1,20}?(\d{1,3}(?:\.\d{3})*(?:,\d{2})?)\s*(?:EUR|€)?/g;
+    let m: RegExpExecArray | null;
+    while ((m = regex.exec(text)) !== null) {
+      const year = parseInt(m[1]!, 10);
+      const amount = parseSlovenianNumber(m[2] ?? null);
+      if (year >= 2010 && year <= 2030 && amount !== null) {
+        result[year] = amount;
+      }
+    }
+    return result;
+  }
+}