我需要解析可能包含字母数字的文本。例如
"I`ve got sixty six tasks"
或
"There is four people"
我的目标是获取子字符串sixty six 和four
在互联网上有很多方法将数字字符串表示形式转换为整数，而无需额外的文本。但我需要下一个结果：
find_numbers("Hello world") -> []
find_numbers("Hello five world") -> ['five']
【问题讨论】：
标签： python-3.x parsing text nltk
您需要为此使用 2 个库：
word2number 将从字符串中提取数字。
例如
>>> print(w2n.word_to_num("Hello five world"))
5
然后您可以使用num2words 库将输出转换回单词：
>>> print(num2words(5))
five
【讨论】：
i have three apples and two bananas w2n return 5
sixty six、three million、three hundred twenty eight
由于无聊，我修改了word_to_num的一个版本。一些错误检查不包括在内，但您可以根据需要添加。
我不会详细介绍它的工作原理，但本质上，它将数字分成组，然后将这些组中的每一个输入word2numbers 算法。
american_number_system = {
'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,
'hundred': 100,
'thousand': 1000,
'million': 1000000,
'billion': 1000000000
}
def number_formation(number_words):
numbers = []
for number_word in number_words:
numbers.append(american_number_system[number_word])
if len(numbers) == 4:
return (numbers[0] * numbers[1]) + numbers[2] + numbers[3]
elif len(numbers) == 3:
return numbers[0] * numbers[1] + numbers[2]
elif len(numbers) == 2:
if 100 in numbers:
return numbers[0] * numbers[1]
else:
return numbers[0] + numbers[1]
else:
return numbers[0]
def get_decimal_sum(decimal_digit_words):
decimal_number_str = []
for dec_word in decimal_digit_words:
if(dec_word not in decimal_words):
return 0
else:
decimal_number_str.append(american_number_system[dec_word])
final_decimal_string = '0.' + ''.join(map(str,decimal_number_str))
return float(final_decimal_string)
def to_num(string):
string = string.replace('-', ' ')
string = string.replace(',', ' ')
words = string.strip().split()
number_groups = []
current_word = []
numbers = []
for word in words:
if word in american_number_system:
current_word.append(word)
elif word.lower() != 'and' and len(current_word):
number_groups.append(current_word)
current_word = []
if len(current_word):
number_groups.append(current_word)
for clean_numbers in number_groups:
clean_decimal_numbers = []
total_sum = 0
if clean_numbers.count('point') == 1:
clean_decimal_numbers = clean_numbers[clean_numbers.index('point')+1:]
clean_numbers = clean_numbers[:clean_numbers.index('point')]
billion_index = clean_numbers.index('billion') if 'billion' in clean_numbers else -1
million_index = clean_numbers.index('million') if 'million' in clean_numbers else -1
thousand_index = clean_numbers.index('thousand') if 'thousand' in clean_numbers else -1
if len(clean_numbers) == 1:
total_sum += american_number_system[clean_numbers[0]]
else:
if billion_index > -1:
billion_multiplier = number_formation(clean_numbers[0:billion_index])
total_sum += billion_multiplier * 1000000000
if million_index > -1:
if billion_index > -1:
million_multiplier = number_formation(clean_numbers[billion_index+1:million_index])
else:
million_multiplier = number_formation(clean_numbers[0:million_index])
total_sum += million_multiplier * 1000000
if thousand_index > -1:
if million_index > -1:
thousand_multiplier = number_formation(clean_numbers[million_index+1:thousand_index])
elif billion_index > -1 and million_index == -1:
thousand_multiplier = number_formation(clean_numbers[billion_index+1:thousand_index])
else:
thousand_multiplier = number_formation(clean_numbers[0:thousand_index])
total_sum += thousand_multiplier * 1000
if thousand_index > -1 and thousand_index != len(clean_numbers)-1:
hundreds = number_formation(clean_numbers[thousand_index+1:])
elif million_index > -1 and million_index != len(clean_numbers)-1:
hundreds = number_formation(clean_numbers[million_index+1:])
elif billion_index > -1 and billion_index != len(clean_numbers)-1:
hundreds = number_formation(clean_numbers[billion_index+1:])
elif thousand_index == -1 and million_index == -1 and billion_index == -1:
hundreds = number_formation(clean_numbers)
else:
hundreds = 0
total_sum += hundreds
if len(clean_decimal_numbers) > 0:
decimal_sum = get_decimal_sum(clean_decimal_numbers)
total_sum += decimal_sum
numbers.append(total_sum)
return numbers
tests = []
tests.append(to_num("I`ve got sixty six tasks"))
tests.append(to_num("There is four people"))
tests.append(to_num("Hello world"))
tests.append(to_num("Hello five world"))
tests.append(to_num("i have three apples and two bananas"))
tests.append(to_num("three hundred twenty eight"))
print(tests)
从这里您可以使用num2words 来反转结果。
编辑：
实际上，重新阅读您的问题，比这容易得多。你只需要找到这些数字的位置并提取它们。
american_number_system = {
'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,
'hundred': 100,
'thousand': 1000,
'million': 1000000,
'billion': 1000000000
}
def extract_num(raw_string):
string = raw_string.replace('-', ' ')
string = string.replace(',', ' ')
words = string.strip().split()
word_pos = False
numbers = []
current_pos = 0
for word in words:
if word in american_number_system:
if word_pos:
length = len(word) + 1
word_pos = (word_pos[0], word_pos[1] + length)
else:
length = len(word)
word_pos = (current_pos, current_pos + length)
elif word.lower() == 'and' and word_pos:
word_pos = (word_pos[0], word_pos[1] + 4)
elif word_pos:
numbers.append(raw_string[word_pos[0]:word_pos[1]])
word_pos = False
current_pos += len(word) + 1
if word_pos:
numbers.append(raw_string[word_pos[0]:])
return numbers
tests = []
tests.append(extract_num("I`ve got sixty six tasks"))
tests.append(extract_num("There is four people"))
tests.append(extract_num("Hello world"))
tests.append(extract_num("Hello five world"))
tests.append(extract_num("i have three apples and two bananas"))
tests.append(extract_num("three hundred twenty eight"))
print(tests)
【讨论】：