萬盛學電腦網

您现在的位置: 萬盛學電腦網 >> 程式設計 >  >> 網路程式設計 >> 程式語言綜合

Python 將英文數字單字轉為阿拉伯數字:支援複合表達與邊界處理

問題定義

將如 "two hundred thirty two"、"nineteen eighty nine"、"one thousand nine hundred and eighty nine" 等英文敘述轉為整數 232、1989、1989。

核心演算法:逆向掃描 + 位值累加

import re

_KNOWN = {
    'zero': 0, 'one': 1, 'two': 2, 'three': 3, 'four': 4,
    'five': 5, 'six': 6, 'seven': 7, 'eight': 8, 'nine': 9,
    'ten': 10, 'eleven': 11, 'twelve': 12, 'thirteen': 13,
    'fourteen': 14, 'fifteen': 15, 'sixteen': 16, 'seventeen': 17,
    'eighteen': 18, 'nineteen': 19, 'twenty': 20,
    'thirty': 30, 'forty': 40, 'fifty': 50, 'sixty': 60,
    'seventy': 70, 'eighty': 80, 'ninety': 90
}

def spoken_word_to_number(text: str) -> int:
    text = text.lower().strip()
    if text in _KNOWN:
        return _KNOWN[text]

    # 切分:空格或連字號
    words = re.split(r'[ -]', text)
    words = [w for w in words if w not in ('and', 'minus', 'negative')]

    # 處理 hundred/thousand 結尾補零
    if words[-1] == 'hundred':
        words += ['zero', 'zero']
    elif words[-1] == 'thousand':
        words += ['zero', 'zero', 'zero']
    if words[0] in ('hundred', 'thousand'):
        words.insert(0, 'one')

    output = 0
    position = 'unit'  # unit → ten → hundred → thousand

    for word in reversed(words):
        if position == 'unit':
            num = _KNOWN[word]
            output += num
            position = 'hundred' if num > 9 else 'ten'

        elif position == 'ten':
            if word != 'hundred':
                num = _KNOWN[word]
                output += num * 10 if num < 10 else num
            position = 'hundred'

        elif position == 'hundred':
            if word not in ('hundred', 'thousand'):
                num = _KNOWN[word]
                output += num * 100
            position = 'thousand' if word == 'thousand' else 'hundred'

        elif position == 'thousand':
            if word != 'thousand':
                num = _KNOWN[word]
                output += num * 1000
            # 千位以上不再處理 (可擴充 million)

    return output


# 測試驗證
assert spoken_word_to_number("nine hundred") == 900
assert spoken_word_to_number("one hundred") == 100
assert spoken_word_to_number("eleven") == 11
assert spoken_word_to_number("twenty two") == 22
assert spoken_word_to_number("thirty-two") == 32
assert spoken_word_to_number("two hundred thirty two") == 232
assert spoken_word_to_number("nineteen hundred eighty nine") == 1989
assert spoken_word_to_number("one thousand nine hundred and eighty nine") == 1989
assert spoken_word_to_number("nine thousand nine hundred") == 9900
print("All tests passed!")

擴充支援 Million/Billion


_SCALES = [
    (10**3, 'thousand'),
    (10**6, 'million'),
    (10**9, 'billion'),
    (10**12, 'trillion'),
]

def parse_large_number(words):
    # 依 scale 由大到小切分
    total = 0
    current = 0
    for word in words:
        if word in _KNOWN:
            current += _KNOWN[word]
        elif word == 'hundred':
            current *= 100
        else:  # thousand, million...
            scale_val = next(v for v, n in _SCALES if n == word)
            total += current * scale_val
            current = 0
    return total + current

現成套件推薦

  • word2number:pip install word2number,from word2number import w2n; w2n.word_to_num("two million three thousand")
  • num2words:反向轉換 (數字→單字),支援多語言。

應用場景

  1. 語音辨識後處理 (ASR 輸出轉數值)
  2. 自然語言理解 (NLU) 槽位填充
  3. 文件 OCR 文字數值提取
  4. 聊天機器人解析用戶輸入金額/數量
copyright © 萬盛學電腦網 all rights reserved