guoju0808/ry-ocr/app/utils/amount_utils.py

"""金额工具：中文大写金额解析 + 与小写金额比对"""
from __future__ import annotations

import re
from typing import Optional

import cn2an

# 小写金额正则
_NUM_PATTERN = re.compile(r"¥?\s*(\d{1,8}(?:,\d{3})*\.\d{2})")
_TAX_AMOUNT_PATTERN = re.compile(r"税\s*额\s*[¥:]?\s*(\d{1,8}(?:,\d{3})*\.\d{2})")
_PRETAX_PATTERN = re.compile(r"(?:不合?税价|不含税)\s*[¥:]?\s*(\d{1,8}(?:,\d{3})*\.\d{2})")
_TOTAL_PATTERN = re.compile(r"价税合计[^\d]*[¥]?\s*(\d{1,8}(?:,\d{3})*\.\d{2})")


def normalize_cn_amount(text: str) -> str:
    """中文金额归一化：圆→元、〇→零"""
    if not text:
        return ""
    return text.replace("圆", "元").replace("〇", "零").replace(" ", "")


def extract_cn_amount(text: str) -> Optional[str]:
    """从原文里匹配出第一段大写金额字符串

    策略（按优先级）：
    1. "价税合计" 后面括号内
    2. 任意中括号里的中文金额
    3. 含"元"或"圆"的最长中文字符串
    """
    if not text:
        return None

    # 1. 价税合计之后括号里
    m = re.search(r"价税合计[^\(（]*[\(（]([零壹贰叁肆伍陆柒捌玖拾佰仟万亿圆元角分整]+)[\)）]", text)
    if m:
        return m.group(1)

    # 2. 任意中括号里的中文金额
    m = re.search(r"[\(（]([零壹贰叁肆伍陆柒捌玖拾佰仟万亿圆元角分整]{3,30})[\)）]", text)
    if m:
        return m.group(1)

    # 3. 含"元"/"圆"的中文片段
    for cand in re.findall(r"[零壹贰叁肆伍陆柒捌玖拾佰仟万亿圆元角分整]{3,30}", text):
        if "元" in cand or "圆" in cand:
            return cand
    return None


def parse_cn_amount(cn_text: str) -> Optional[float]:
    """把中文大写金额转 float，例如 '贰佰元整' → 200.0"""
    if not cn_text:
        return None
    try:
        s = normalize_cn_amount(cn_text)
        # cn2an 要求带 '元' 或 '圆' 结尾
        s = s.rstrip("整")
        if not s.endswith("元"):
            s += "元"
        value = cn2an.cn2an(s, "smart")
        return float(value)
    except Exception:
        return None


def amount_consistent(cn_text: Optional[str], num_amount: Optional[float]) -> Optional[bool]:
    """大写 vs 小写金额比对"""
    if cn_text is None or num_amount is None:
        return None
    cn_value = parse_cn_amount(cn_text)
    if cn_value is None:
        return None
    return abs(cn_value - num_amount) < 0.011


def extract_num_amount(text: str) -> Optional[float]:
    """从文本里提取第一个形如 1234.56 或 ¥1,234.56 的金额"""
    if not text:
        return None
    m = _NUM_PATTERN.search(text)
    if not m:
        return None
    try:
        return float(m.group(1).replace(",", ""))
    except ValueError:
        return None


def extract_total_amount(text: str) -> Optional[float]:
    m = _TOTAL_PATTERN.search(text)
    if m:
        try:
            return float(m.group(1).replace(",", ""))
        except ValueError:
            pass
    return extract_num_amount(text)


def extract_tax_amount(text: str) -> Optional[float]:
    m = _TAX_AMOUNT_PATTERN.search(text)
    if m:
        try:
            return float(m.group(1).replace(",", ""))
        except ValueError:
            pass
    return None


def extract_pretax_amount(text: str) -> Optional[float]:
    m = _PRETAX_PATTERN.search(text)
    if m:
        try:
            return float(m.group(1).replace(",", ""))
        except ValueError:
            pass
    return None