nordabiz/krs_audit_service.py

#!/usr/bin/env python3
"""
KRS Audit Service - Full data extraction from KRS PDF files.

Downloads PDF documents from EKRS and extracts complete company data:
- Basic info: KRS, NIP, REGON, company name, legal form
- Address: full address with email, website
- Capital: share capital amount, shares count, nominal value
- People: management board, shareholders, procurators
- PKD codes: main and secondary business activities
- Financial reports: filing dates
- Representation rules

Author: Maciej Pienczyn, InPi sp. z o.o.
Created: 2026-01-13
"""

import re
import os
import json
import logging
from pathlib import Path
from datetime import datetime, date
from dataclasses import dataclass, field, asdict
from typing import List, Optional, Dict, Any, Tuple
from decimal import Decimal

# Setup logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

try:
    import pdfplumber
except ImportError:
    logger.error("Required library pdfplumber. Install: pip install pdfplumber")
    raise


# ============================================================
# DATA CLASSES
# ============================================================

@dataclass
class KRSPerson:
    """Person from KRS (board member, shareholder, procurator)"""
    nazwisko: str
    imiona: str
    pesel: Optional[str] = None
    rola: str = ""  # PREZES ZARZĄDU, CZŁONEK ZARZĄDU, WSPÓLNIK, PROKURENT
    rola_kategoria: str = ""  # zarzad, wspolnik, prokurent

    # For shareholders
    udzialy_liczba: Optional[int] = None
    udzialy_wartosc: Optional[Decimal] = None
    udzialy_procent: Optional[Decimal] = None

    def full_name(self) -> str:
        return f"{self.imiona} {self.nazwisko}"


@dataclass
class KRSPKD:
    """PKD code from KRS"""
    kod: str  # e.g., "62.03.Z"
    opis: str
    jest_przewazajacy: bool = False


@dataclass
class KRSFinancialReport:
    """Financial report filing info from KRS"""
    okres_od: Optional[date] = None
    okres_do: Optional[date] = None
    data_zlozenia: Optional[date] = None


@dataclass
class KRSAddress:
    """Company address from KRS"""
    ulica: Optional[str] = None
    numer_domu: Optional[str] = None
    numer_lokalu: Optional[str] = None
    miejscowosc: Optional[str] = None
    kod_pocztowy: Optional[str] = None
    poczta: Optional[str] = None
    wojewodztwo: Optional[str] = None
    powiat: Optional[str] = None
    gmina: Optional[str] = None
    email: Optional[str] = None
    www: Optional[str] = None


@dataclass
class KRSFullData:
    """Complete data extracted from KRS PDF"""
    # Identifiers
    krs: str
    nip: Optional[str] = None
    regon: Optional[str] = None

    # Basic info
    nazwa: str = ""
    nazwa_skrocona: Optional[str] = None
    forma_prawna: Optional[str] = None

    # Dates
    data_rejestracji: Optional[date] = None
    data_umowy_spolki: Optional[date] = None
    czas_trwania: Optional[str] = None  # "NIEOZNACZONY" or date

    # Address
    siedziba: Optional[KRSAddress] = None

    # Capital
    kapital_zakladowy: Optional[Decimal] = None
    waluta: str = "PLN"
    liczba_udzialow: Optional[int] = None
    wartosc_nominalna_udzialu: Optional[Decimal] = None

    # Representation
    sposob_reprezentacji: Optional[str] = None

    # PKD codes
    pkd_przewazajacy: Optional[KRSPKD] = None
    pkd_pozostale: List[KRSPKD] = field(default_factory=list)

    # People
    zarzad: List[KRSPerson] = field(default_factory=list)
    wspolnicy: List[KRSPerson] = field(default_factory=list)
    prokurenci: List[KRSPerson] = field(default_factory=list)
    rada_nadzorcza: List[KRSPerson] = field(default_factory=list)

    # Status
    czy_opp: bool = False
    zaleglosci: Optional[str] = None
    wierzytelnosci: Optional[str] = None

    # Financial reports
    sprawozdania_finansowe: List[KRSFinancialReport] = field(default_factory=list)

    # Metadata
    zrodlo: str = "ekrs.ms.gov.pl"
    data_pobrania: Optional[datetime] = None

    def to_dict(self) -> Dict[str, Any]:
        """Convert to dictionary for JSON serialization"""
        def convert_value(v):
            if isinstance(v, (date, datetime)):
                return v.isoformat()
            if isinstance(v, Decimal):
                return float(v)
            if hasattr(v, '__dict__'):
                return {k: convert_value(val) for k, val in asdict(v).items()}
            if isinstance(v, list):
                return [convert_value(item) for item in v]
            return v

        return {k: convert_value(v) for k, v in asdict(self).items()}


# ============================================================
# PDF PARSING FUNCTIONS
# ============================================================

def extract_text_from_pdf(pdf_path: str) -> str:
    """Extract full text from PDF"""
    with pdfplumber.open(pdf_path) as pdf:
        text = ""
        for page in pdf.pages:
            page_text = page.extract_text()
            if page_text:
                text += page_text + "\n"
    return text


def parse_date(date_str: str) -> Optional[date]:
    """Parse date from various formats"""
    if not date_str:
        return None

    # Try DD.MM.YYYY format
    match = re.search(r'(\d{2})\.(\d{2})\.(\d{4})', date_str)
    if match:
        try:
            return date(int(match.group(3)), int(match.group(2)), int(match.group(1)))
        except ValueError:
            pass

    # Try YYYY-MM-DD format
    match = re.search(r'(\d{4})-(\d{2})-(\d{2})', date_str)
    if match:
        try:
            return date(int(match.group(1)), int(match.group(2)), int(match.group(3)))
        except ValueError:
            pass

    return None


def parse_money(money_str: str) -> Optional[Decimal]:
    """Parse money value from Polish format (e.g., '4.000,00' or '5 000,00')"""
    if not money_str:
        return None

    # Remove spaces
    cleaned = money_str.replace(' ', '')

    # Handle Polish format: dot as thousands separator, comma as decimal separator
    # e.g., "4.000,00" should become "4000.00"
    # First, check if we have Polish format (comma present)
    if ',' in cleaned:
        # Remove dots (thousands separators) and replace comma with dot (decimal separator)
        cleaned = cleaned.replace('.', '').replace(',', '.')
    else:
        # US format or simple integer - dots are decimal separators
        pass

    # Extract number
    match = re.search(r'([\d\.]+)', cleaned)
    if match:
        try:
            return Decimal(match.group(1))
        except:
            pass

    return None


def parse_shares_info(shares_str: str) -> Tuple[Optional[int], Optional[Decimal]]:
    """Parse shares count and value from string like '80 UDZIAŁÓW O ŁĄCZNEJ WARTOŚCI 4.000,00 ZŁ'"""
    count = None
    value = None

    # Count pattern
    count_match = re.search(r'(\d+)\s+UDZIAŁ', shares_str, re.IGNORECASE)
    if count_match:
        count = int(count_match.group(1))

    # Value pattern
    value_match = re.search(r'WARTOŚCI\s+([\d\s,\.]+)\s*ZŁ', shares_str, re.IGNORECASE)
    if value_match:
        value = parse_money(value_match.group(1))

    return count, value


def extract_person_block(lines: List[str], start_idx: int, role_category: str) -> Optional[KRSPerson]:
    """Extract person data from PDF text block"""
    person = KRSPerson(nazwisko="", imiona="", rola_kategoria=role_category)
    found_nazwisko = False
    found_imiona = False

    for i in range(start_idx, min(start_idx + 15, len(lines))):
        line = lines[i].strip()

        # Stop at next person block
        if i > start_idx and ('1.Nazwisko' in line or 'Nazwisko / Nazwa' in line):
            break

        # Nazwisko
        if not found_nazwisko and 'Nazwisko' in line and ' - ' in line:
            match = re.search(r' - ([A-ZĄĆĘŁŃÓŚŹŻ\-]+)$', line)
            if match:
                person.nazwisko = match.group(1)
                found_nazwisko = True

        # Imiona
        if not found_imiona and 'Imiona' in line and ' - ' in line:
            match = re.search(r' - ([A-ZĄĆĘŁŃÓŚŹŻ ]+)$', line)
            if match:
                person.imiona = match.group(1).strip()
                found_imiona = True

        # PESEL
        if 'PESEL' in line and ' - ' in line:
            match = re.search(r' - (\d{11})', line)
            if match:
                person.pesel = match.group(1)

        # Function (for board members)
        if 'Funkcja' in line and ' - ' in line:
            match = re.search(r' - ([A-ZĄĆĘŁŃÓŚŹŻ ]+)$', line)
            if match:
                person.rola = match.group(1).strip()

        # Shares (for shareholders)
        if 'udziały' in line.lower() and ' - ' in line:
            match = re.search(r' - (.+)$', line)
            if match:
                shares_str = match.group(1)
                count, value = parse_shares_info(shares_str)
                person.udzialy_liczba = count
                person.udzialy_wartosc = value

    if person.nazwisko and person.imiona:
        return person
    return None


def parse_krs_pdf_full(pdf_path: str) -> KRSFullData:
    """
    Parse KRS PDF and extract all available data.

    Args:
        pdf_path: Path to KRS PDF file

    Returns:
        KRSFullData object with all extracted data
    """
    text = extract_text_from_pdf(pdf_path)
    lines = text.split('\n')

    data = KRSFullData(krs="", data_pobrania=datetime.now())

    # === Basic identifiers ===

    # KRS number
    krs_match = re.search(r'Numer KRS:\s*(\d{10})', text)
    if krs_match:
        data.krs = krs_match.group(1)

    # NIP and REGON
    nip_match = re.search(r'NIP:\s*(\d{10})', text)
    if nip_match:
        data.nip = nip_match.group(1)

    regon_match = re.search(r'REGON:\s*(\d{9,14})', text)
    if regon_match:
        data.regon = regon_match.group(1)

    # === Company name and form ===

    # Company name
    nazwa_match = re.search(r'3\.Firma,?\s+pod którą spółka działa\s+\d+\s+-\s+([^\n]+)', text)
    if nazwa_match:
        data.nazwa = nazwa_match.group(1).strip()

    # Legal form
    forma_match = re.search(r'1\.Oznaczenie formy prawnej\s+\d+\s+-\s+([^\n]+)', text)
    if forma_match:
        data.forma_prawna = forma_match.group(1).strip()

    # OPP status
    opp_match = re.search(r'status organizacji\s*pożytku publicznego\?\s+\d+\s+-\s+(TAK|NIE)', text, re.IGNORECASE)
    if opp_match:
        data.czy_opp = opp_match.group(1).upper() == 'TAK'

    # === Registration date ===

    # Find first entry date
    reg_match = re.search(r'Nr wpisu 1 Data dokonania wpisu (\d{2}\.\d{2}\.\d{4})', text)
    if reg_match:
        data.data_rejestracji = parse_date(reg_match.group(1))

    # Company agreement date
    umowa_match = re.search(r'Informacja o zawarciu.+?(\d{2}\.\d{2}\.\d{4})', text, re.DOTALL)
    if umowa_match:
        data.data_umowy_spolki = parse_date(umowa_match.group(1))

    # Duration
    czas_match = re.search(r'Czas,?\s+na jaki została utworzona spółka\s+\d+\s+-\s+([^\n]+)', text)
    if czas_match:
        data.czas_trwania = czas_match.group(1).strip()

    # === Address ===

    data.siedziba = KRSAddress()

    # Full address line
    adres_match = re.search(
        r'2\.Adres\s+\d+\s+-\s+ul\.\s*([^,]+),\s*nr\s*(\d+\w*)'
        r'(?:,\s*lok\.\s*([^,]+))?'
        r',\s*miejsc\.\s*([^,]+)'
        r',\s*kod\s*(\d{2}-\d{3})'
        r',\s*poczta\s*([^,]+)',
        text, re.IGNORECASE
    )
    if adres_match:
        data.siedziba.ulica = adres_match.group(1).strip()
        data.siedziba.numer_domu = adres_match.group(2).strip()
        if adres_match.group(3):
            data.siedziba.numer_lokalu = adres_match.group(3).strip()
        data.siedziba.miejscowosc = adres_match.group(4).strip()
        data.siedziba.kod_pocztowy = adres_match.group(5).strip()
        data.siedziba.poczta = adres_match.group(6).strip()

    # Siedziba (province info)
    siedziba_match = re.search(
        r'1\.Siedziba\s+\d+\s+-\s+kraj\s+POLSKA,\s*woj\.\s*([^,]+),\s*powiat\s*([^,]+),\s*gmina\s*([^,]+)',
        text, re.IGNORECASE
    )
    if siedziba_match:
        data.siedziba.wojewodztwo = siedziba_match.group(1).strip()
        data.siedziba.powiat = siedziba_match.group(2).strip()
        data.siedziba.gmina = siedziba_match.group(3).strip()

    # Email
    email_match = re.search(r'3\.Adres poczty elektronicznej\s+\d+\s+-\s+([^\n]+)', text)
    if email_match:
        data.siedziba.email = email_match.group(1).strip()

    # Website
    www_match = re.search(r'4\.Adres strony internetowej\s+\d+\s+-\s+([^\n]+)', text)
    if www_match:
        data.siedziba.www = www_match.group(1).strip()

    # === Capital ===

    kapital_match = re.search(r'1\.Wysokość kapitału zakładowego\s+\d+\s+-\s+([\d\s,\.]+)\s*ZŁ', text, re.IGNORECASE)
    if kapital_match:
        data.kapital_zakladowy = parse_money(kapital_match.group(1))

    # === Representation rules ===

    repr_match = re.search(r'2\.Sposób reprezentacji podmiotu\s+\d+\s+-\s+([^\n]+(?:\n[^\n]*)?)', text)
    if repr_match:
        # Clean up multiline representation
        repr_text = repr_match.group(1).strip()
        # Remove line breaks and extra spaces
        repr_text = ' '.join(repr_text.split())
        data.sposob_reprezentacji = repr_text

    # === PKD codes ===

    # Find the PKD section first
    pkd_section_match = re.search(
        r'Przedmiot działalności.*?(?=Rubryka 2|Dział 4|$)',
        text,
        re.DOTALL | re.IGNORECASE
    )

    if pkd_section_match:
        pkd_section = pkd_section_match.group(0)

        # Main PKD (przeważający)
        pkd_glowny_match = re.search(
            r'1\.Przedmiot przeważającej.*?(\d{2}),\s*(\d{2}),\s*([A-Z]),\s*([A-ZĄĆĘŁŃÓŚŹŻ][^\n]*)',
            pkd_section,
            re.DOTALL
        )
        if pkd_glowny_match:
            kod = f"{pkd_glowny_match.group(1)}.{pkd_glowny_match.group(2)}.{pkd_glowny_match.group(3)}"
            opis = pkd_glowny_match.group(4).strip()
            # Clean up multi-line descriptions
            opis = ' '.join(opis.split())
            data.pkd_przewazajacy = KRSPKD(kod=kod, opis=opis, jest_przewazajacy=True)

        # Secondary PKDs (pozostałe) - find the section first
        pkd_pozostale_section = re.search(
            r'2\.Przedmiot pozostałej działalności.*?(?=Rubryka|Dział|$)',
            pkd_section,
            re.DOTALL
        )

        if pkd_pozostale_section:
            pozostale_text = pkd_pozostale_section.group(0)

            # Find all PKD codes in this section
            # Pattern: number(s) followed by PKD code (XX, XX, Z, DESCRIPTION)
            pkd_entries = re.findall(
                r'(?:^|\s)(\d{2}),\s*(\d{2}),\s*([A-Z]),\s*([A-ZĄĆĘŁŃÓŚŹŻ][^\n]*?)(?=\n\d|\n[A-Z]|\Z)',
                pozostale_text,
                re.MULTILINE
            )

            # If the above pattern doesn't work well, try a simpler one
            if len(pkd_entries) < 2:
                pkd_entries = re.findall(
                    r'(\d{2}),\s*(\d{2}),\s*([A-Z]),\s*([A-ZĄĆĘŁŃÓŚŹŻ][A-ZĄĆĘŁŃÓŚŹŻ\s\-,]+)',
                    pozostale_text
                )

            # Deduplicate by PKD code
            seen_codes = set()
            if data.pkd_przewazajacy:
                seen_codes.add(data.pkd_przewazajacy.kod)

            for match in pkd_entries:
                kod = f"{match[0]}.{match[1]}.{match[2]}"
                if kod not in seen_codes:
                    seen_codes.add(kod)
                    opis = match[3].strip()
                    # Clean up multi-line descriptions
                    opis = ' '.join(opis.split())
                    # Truncate very long descriptions
                    if len(opis) > 200:
                        opis = opis[:200] + '...'
                    data.pkd_pozostale.append(KRSPKD(kod=kod, opis=opis, jest_przewazajacy=False))

    # === People ===

    in_zarzad = False
    in_wspolnicy = False
    in_prokurenci = False
    in_rada = False

    for i, line in enumerate(lines):
        line_stripped = line.strip()

        # Section detection
        if 'ZARZĄD' in line_stripped.upper() and 'Nazwa organu' in line_stripped:
            in_zarzad = True
            in_wspolnicy = False
            in_prokurenci = False
            in_rada = False
            continue

        if 'Dane wspólników' in line_stripped or ('Rubryka 7' in line_stripped and 'wspólników' in line_stripped.lower()):
            in_wspolnicy = True
            in_zarzad = False
            in_prokurenci = False
            in_rada = False
            continue

        if 'Prokurenci' in line_stripped:
            in_prokurenci = True
            in_zarzad = False
            in_wspolnicy = False
            in_rada = False
            continue

        if 'Organ nadzoru' in line_stripped:
            in_rada = True
            in_zarzad = False
            in_wspolnicy = False
            in_prokurenci = False
            continue

        # New section - reset
        if 'Dział' in line_stripped and re.match(r'Dział \d+', line_stripped):
            if 'Dział 2' not in line_stripped and 'Dział 1' not in line_stripped:
                in_zarzad = False
                in_wspolnicy = False
                in_prokurenci = False
                in_rada = False

        # Parse person when we find "Nazwisko"
        if '1.Nazwisko' in line_stripped or 'Nazwisko / Nazwa' in line_stripped:
            if in_zarzad:
                person = extract_person_block(lines, i, 'zarzad')
                if person:
                    if not person.rola:
                        person.rola = "CZŁONEK ZARZĄDU"
                    data.zarzad.append(person)
            elif in_wspolnicy:
                person = extract_person_block(lines, i, 'wspolnik')
                if person:
                    person.rola = "WSPÓLNIK"
                    data.wspolnicy.append(person)
            elif in_prokurenci:
                person = extract_person_block(lines, i, 'prokurent')
                if person:
                    person.rola = "PROKURENT"
                    data.prokurenci.append(person)
            elif in_rada:
                person = extract_person_block(lines, i, 'rada_nadzorcza')
                if person:
                    person.rola = "CZŁONEK RADY NADZORCZEJ"
                    data.rada_nadzorcza.append(person)

    # Calculate share percentage for shareholders
    if data.kapital_zakladowy and data.wspolnicy:
        for wspolnik in data.wspolnicy:
            if wspolnik.udzialy_wartosc:
                wspolnik.udzialy_procent = Decimal(str(
                    round(float(wspolnik.udzialy_wartosc) / float(data.kapital_zakladowy) * 100, 2)
                ))

    # Calculate total shares
    total_shares = sum(w.udzialy_liczba or 0 for w in data.wspolnicy)
    if total_shares > 0:
        data.liczba_udzialow = total_shares
        if data.kapital_zakladowy:
            data.wartosc_nominalna_udzialu = Decimal(str(
                round(float(data.kapital_zakladowy) / total_shares, 2)
            ))

    # === Financial reports ===

    # Parse financial report filings - multiple patterns needed
    # Format 1: "1.Wzmianka o złożeniu rocznego 1 4 - 22.06.2022 OD 05.02.2021 DO 31.12.2021"
    # Format 2: "2 6 - 20.06.2023 OD 01.01.2022 DO 31.12.2022" (continuation lines)
    report_matches = re.findall(
        r'(\d{2}\.\d{2}\.\d{4})\s+OD\s+(\d{2}\.\d{2}\.\d{4})\s+DO\s+(\d{2}\.\d{2}\.\d{4})',
        text
    )
    seen_reports = set()
    for match in report_matches:
        # Deduplicate by period
        key = (match[1], match[2])
        if key not in seen_reports:
            seen_reports.add(key)
            report = KRSFinancialReport(
                data_zlozenia=parse_date(match[0]),
                okres_od=parse_date(match[1]),
                okres_do=parse_date(match[2])
            )
            data.sprawozdania_finansowe.append(report)

    # Sort by period end date
    data.sprawozdania_finansowe.sort(key=lambda r: r.okres_do or date.min)

    return data


# ============================================================
# MAIN ENTRY POINT
# ============================================================

def parse_krs_pdf(pdf_path: str, verbose: bool = False) -> Dict[str, Any]:
    """
    Parse KRS PDF and return dictionary with all data.

    This is the main entry point for external use.

    Args:
        pdf_path: Path to KRS PDF file
        verbose: Print debug information

    Returns:
        Dictionary with all extracted data
    """
    if verbose:
        logger.info(f"Parsing: {pdf_path}")

    data = parse_krs_pdf_full(pdf_path)

    if verbose:
        logger.info(f"  KRS: {data.krs}")
        logger.info(f"  Nazwa: {data.nazwa}")
        logger.info(f"  NIP: {data.nip}, REGON: {data.regon}")
        logger.info(f"  Kapitał: {data.kapital_zakladowy} {data.waluta}")
        logger.info(f"  Zarząd: {len(data.zarzad)} osób")
        logger.info(f"  Wspólnicy: {len(data.wspolnicy)} osób")
        if data.pkd_przewazajacy:
            logger.info(f"  PKD główny: {data.pkd_przewazajacy.kod}")

    return data.to_dict()


def main():
    """CLI entry point for testing"""
    import argparse

    parser = argparse.ArgumentParser(description="Parse KRS PDF files (full data extraction)")
    parser.add_argument("--file", type=str, help="Single PDF file to parse")
    parser.add_argument("--dir", type=str, help="Directory with PDF files")
    parser.add_argument("--output", type=str, help="Output JSON file")
    parser.add_argument("--verbose", "-v", action="store_true", help="Verbose output")
    args = parser.parse_args()

    results = []

    if args.file:
        data = parse_krs_pdf(args.file, verbose=args.verbose)
        results.append(data)

        # Print summary
        print(f"\n{'='*60}")
        print(f"{data['nazwa']} (KRS: {data['krs']})")
        print(f"{'='*60}")
        print(f"NIP: {data['nip']}, REGON: {data['regon']}")
        print(f"Forma prawna: {data['forma_prawna']}")
        print(f"Data rejestracji: {data['data_rejestracji']}")

        if data['siedziba']:
            addr = data['siedziba']
            print(f"Adres: {addr.get('ulica', '')} {addr.get('numer_domu', '')}, {addr.get('kod_pocztowy', '')} {addr.get('miejscowosc', '')}")
            if addr.get('email'):
                print(f"Email: {addr['email']}")
            if addr.get('www'):
                print(f"WWW: {addr['www']}")

        print(f"\nKapitał zakładowy: {data['kapital_zakladowy']} {data['waluta']}")
        print(f"Liczba udziałów: {data['liczba_udzialow']}")
        print(f"Wartość nominalna: {data['wartosc_nominalna_udzialu']} {data['waluta']}")

        print(f"\nSposób reprezentacji: {data['sposob_reprezentacji']}")

        if data['pkd_przewazajacy']:
            print(f"\nPKD przeważający: {data['pkd_przewazajacy']['kod']} - {data['pkd_przewazajacy']['opis']}")

        if data['pkd_pozostale']:
            print("PKD pozostałe:")
            for pkd in data['pkd_pozostale']:
                print(f"  - {pkd['kod']}: {pkd['opis']}")

        print(f"\nZarząd ({len(data['zarzad'])} osób):")
        for p in data['zarzad']:
            print(f"  - {p['imiona']} {p['nazwisko']} - {p['rola']}")

        print(f"\nWspólnicy ({len(data['wspolnicy'])} osób):")
        for p in data['wspolnicy']:
            shares_info = ""
            if p.get('udzialy_liczba'):
                shares_info = f" ({p['udzialy_liczba']} udziałów, {p.get('udzialy_procent', '?')}%)"
            print(f"  - {p['imiona']} {p['nazwisko']}{shares_info}")

        if data['prokurenci']:
            print(f"\nProkurenci ({len(data['prokurenci'])} osób):")
            for p in data['prokurenci']:
                print(f"  - {p['imiona']} {p['nazwisko']}")

        if data['sprawozdania_finansowe']:
            print(f"\nSprawozdania finansowe ({len(data['sprawozdania_finansowe'])}):")
            for sf in data['sprawozdania_finansowe']:
                print(f"  - {sf['okres_od']} do {sf['okres_do']} (złożone: {sf['data_zlozenia']})")

    elif args.dir:
        pdf_dir = Path(args.dir)
        pdf_files = list(pdf_dir.glob("*.pdf"))
        print(f"Found {len(pdf_files)} PDF files")

        for pdf_file in pdf_files:
            try:
                data = parse_krs_pdf(str(pdf_file), verbose=args.verbose)
                results.append(data)
                print(f"  ✓ {data['nazwa']} (KRS: {data['krs']})")
            except Exception as e:
                print(f"  ✗ {pdf_file.name}: {e}")

    # Save results
    if args.output and results:
        with open(args.output, 'w', encoding='utf-8') as f:
            json.dump(results, f, ensure_ascii=False, indent=2)
        print(f"\nResults saved to: {args.output}")


if __name__ == "__main__":
    main()