roa2web-service-auto/backend/scripts/whatsapp_import/process_and_save.py

#!/usr/bin/env python3
"""
Script complet: PDF → OCR → Oracle
Usage: python process_and_save.py <path_to_pdf> [--save]

Fără --save: doar arată ce ar salva (dry run)
Cu --save: salvează efectiv în Oracle
"""
import sys
sys.path.insert(0, "/workspace/roa2web")

import asyncio
import oracledb
from datetime import datetime
from decimal import Decimal
from pathlib import Path

ORACLE_CONFIG = {
    "user": "CONTAFIN_ORACLE",
    "password": "ROMFASTSOFT",
    "dsn": "10.0.20.121:1521/ROA"
}

CUI_TO_CONT = {
    "11201891": "6022",  # MOL
    "1590082": "6022",   # OMV Petrom
    "10562600": "6021",  # Dedeman
}

def get_cont_cheltuiala(cui: str) -> str:
    cui_clean = cui.upper().replace("RO", "").strip()
    return CUI_TO_CONT.get(cui_clean, "6028")

async def process_pdf(pdf_path: Path):
    from backend.modules.data_entry.services.ocr_service import ocr_service
    
    mime_type = "application/pdf" if pdf_path.suffix.lower() == ".pdf" else "image/jpeg"
    
    print(f"\n[OCR] Processing: {pdf_path.name}")
    print("-" * 50)
    
    success, message, result = await ocr_service.process_image(pdf_path, mime_type)
    
    if not success:
        print(f"ERROR: {message}")
        return None
    
    print(f"Partner: {result.partner_name}")
    print(f"CUI: {result.cui}")
    print(f"Data: {result.receipt_date}")
    print(f"Numar: {result.receipt_number}")
    print(f"Total: {result.amount}")
    print(f"TVA: {result.tva_total}")
    print(f"Confidence: {result.overall_confidence:.0%}")
    
    return result

def save_to_oracle(result, do_commit: bool = False):
    mode = "SAVE" if do_commit else "DRY RUN"
    print(f"\n[Oracle] {mode}")
    print("-" * 50)
    
    conn = oracledb.connect(**ORACLE_CONFIG)
    cursor = conn.cursor()
    
    try:
        an = result.receipt_date.year if result.receipt_date else datetime.now().year
        luna = result.receipt_date.month if result.receipt_date else datetime.now().month
        receipt_date = result.receipt_date or datetime.now().date()
        
        cursor.execute("SELECT NVL(MAX(COD), 0) + 1 FROM MARIUSM_AUTO.ACT WHERE AN = :an AND LUNA = :luna", an=an, luna=luna)
        cod = cursor.fetchone()[0]
        
        cui_clean = (result.cui or "").upper().replace("RO", "").strip()
        cursor.execute("SELECT ID_PART FROM MARIUSM_AUTO.NOM_PARTENERI WHERE COD_FISCAL = :cui OR COD_FISCAL = :cui2", 
                      cui=cui_clean, cui2="RO"+cui_clean)
        row = cursor.fetchone()
        id_part = row[0] if row else 0
        
        total = float(result.amount or 0)
        tva = float(result.tva_total or 0)
        fara_tva = total - tva
        nract = int(result.receipt_number) if result.receipt_number and result.receipt_number.isdigit() else 0
        cont = get_cont_cheltuiala(result.cui or "")
        expl = f"OCR: {result.partner_name or 'N/A'}"
        
        print(f"COD: {cod}")
        print(f"Partner ID: {id_part} (CUI: {cui_clean})")
        print(f"Cont: {cont}")
        
        lines = [
            (cont, "401", fara_tva, expl),
            ("401", "5311", total, f"Plata {expl}"),
        ]
        if tva > 0:
            lines.insert(1, ("4426", "401", tva, f"TVA {expl}"))
        
        for scd, scc, suma, e in lines:
            cursor.execute("""
                INSERT INTO MARIUSM_AUTO.ACT_TEMP (
                    LUNA, AN, COD, DATAIREG, DATAACT, NRACT,
                    EXPLICATIA, SCD, SCC, SUMA, ID_PARTC, ID_UTIL, DATAORA
                ) VALUES (
                    :luna, :an, :cod, TRUNC(SYSDATE), :dataact, :nract,
                    :expl, :scd, :scc, :suma, :id_part, 0, SYSDATE
                )
            """, luna=luna, an=an, cod=cod, dataact=receipt_date, nract=nract,
                expl=e, scd=scd, scc=scc, suma=suma, id_part=id_part)
            print(f"  {scd} = {scc}: {suma:.2f}")
        
        if do_commit:
            conn.commit()
            print(f"\nSAVED to Oracle (COD={cod})")
        else:
            conn.rollback()
            print(f"\nDRY RUN - not saved (would be COD={cod})")
        
        return cod
        
    except Exception as e:
        print(f"ERROR: {e}")
        conn.rollback()
        raise
    finally:
        cursor.close()
        conn.close()

async def main():
    if len(sys.argv) < 2:
        print("Usage: python process_and_save.py <pdf_path> [--save]")
        return
    
    pdf_path = Path(sys.argv[1])
    if not pdf_path.exists():
        pdf_path = Path(f"/workspace/roa2web/tests/fixtures/ocr-samples/{sys.argv[1]}")
    
    if not pdf_path.exists():
        print(f"File not found: {sys.argv[1]}")
        return
    
    do_save = "--save" in sys.argv
    
    print("=" * 50)
    print("PDF -> OCR -> Oracle")
    print("=" * 50)
    
    result = await process_pdf(pdf_path)
    if result:
        save_to_oracle(result, do_commit=do_save)
    
    print("\n" + "=" * 50)

if __name__ == "__main__":
    asyncio.run(main())