extractor.py

import re


# -----------------------------
# INTEREST RATE EXTRACTION
# -----------------------------

def extract_interest_rate(text):
    """
    Extract interest rate using semantic priority.
    The document can contain multiple rates.
    We select the most authoritative one.
    """

    priority_patterns = [
        # 1. Comparison rate (most stable & always present)
        r"שיעור\s+הריבית\s+לצרכי\s+השוואה\s*[:\-]?\s*(\d+\.\d+)\s*%",

        # 2. Forecast total interest rate
        r"הריבית\s+הכוללת\s+החזויה\s*[:\-]?\s*(\d+\.\d+)\s*%",

        # 3. Adjusted interest rate
        r"שיעור\s+הריבית\s+המתואמת\s*[:\-]?\s*(\d+\.\d+)\s*%",

        # 4. Base interest rate
        r"שיעור\s+הריבית\s*[:\-]?\s*(\d+\.\d+)\s*%"
    ]

    for pattern in priority_patterns:
        match = re.search(pattern, text)
        if match:
            try:
                return float(match.group(1))
            except ValueError:
                continue

    # Special case: Bank of Israel 0% loans
    if "מתווה בנק ישראל" in text or "ריבית 0" in text:
        return 0.0

    return None


# -----------------------------
# LOAN AMOUNT EXTRACTION
# -----------------------------

def extract_loan_amount(text):
    """
    Extract loan amount ONLY from execution amount.
    Never guess from balances, totals, or monthly values.
    """

    priority_patterns = [
        # Canonical execution amount
        r"סכום\s+חלק\s+זה\s+בעת\s+הביצוע\s*[:\-]?\s*([\d,]+(?:\.\d{2})?)",

        # Variant formatting sometimes seen
        r"סכום\s+ההלוואה\s+בעת\s+הביצוע\s*[:\-]?\s*([\d,]+(?:\.\d{2})?)"
    ]

    for pattern in priority_patterns:
        match = re.search(pattern, text)
        if match:
            value = match.group(1).replace(",", "")
            try:
                return float(value)
            except ValueError:
                continue

    return None