number_to_words() with symbols

If the functionality of number_to_words() is expected to expand, I suggest including some awareness of symbols, especially with those symbols most commonly used with numbers, such as '$', '%', 'º', '/', ':', and the quotes.

### Positive/Negative value fixes
Negative values should have a way to alter the prefix and/or add a suffix:
`>>> number_to_words(-33)`
`'minus thirty-three'`
`>>> number_to_words("-33")`
`'minus thirty-three'`
Optional returns could include 'negative thirty-three', 'thirty-three down'

Similarly, for the leading 'plus':
`>>> number_to_words("+33")`
`'plus thirty-three'`
Optional returns could include 'positive thirty-three', 'thirty-three up', or nothing.

### Common symbol translations other than 'plus' or 'minus'
Values with unit symbols simply lose them:
`>>> number_to_words("33%")`
`'thirty-three'` ... 'percent' suffix lost
`>>> number_to_words("$33")`
`'thirty-three'` ... 'dollars' qualifier lost
`number_to_words("33º")`
`'thirty-three'` ... 'degrees' suffix lost
`>>> number_to_words("33#")`
`'thirty-three'` ... 'pounds' suffix lost
`>>> number_to_words("#33")`
`'thirty-three'` ... 'number' or 'hashtag' prefix lost
`>>> number_to_words("+/-33")`
`'plus thirty-three'` ... 'or minus' portion of prefix lost
`>>> number_to_words("33'")`
`'thirty-three'` ... 'feet' or 'minutes' suffix lost
`>>> number_to_words("33""")`
`'thirty-three'` ... 'inches' or 'seconds' suffix lost

Although, not as critical since they are separate numbers, these situations can also arise:
`>>> number_to_words("3:2")`
`'thirty-two'` ... 'to', possible 'ratio' or 'odds' suffix, and individuality of numbers are lost
`>>> number_to_words("3 + 2")`
`'thirty-two'` ... 'plus' and individuality of numbers are lost
`>>> number_to_words("33º22'11""", andword="")` (no 'and' to get proper EN_US version)
`'three hundred thirty-two thousand, two hundred eleven'`
... 'degrees', 'minutes', 'seconds' as well as individuality of numbers are lost

I understand why it ignores spaces and commas, but other than '+', '-', and '.', it treats all symbols as meaningless, which they are not, and does not handle repetitious or multiple symbols, such as used in dates:

`>>> number_to_words("22.11.33")`
`'twenty-two point one one three three'` ... second '.' is ignored
`>>> number_to_words("11/22/2033")`
`'eleven million, two hundred and twenty-two thousand and thirty-three'` ... slashes are ignored
Dates themselves, of course, are a whole other area of complication, but the symbols are still being lost.

### Other usages lost
`>>> number_to_words("2a12")`
`'two hundred and twelve'` ... hex digits are lost
`>>> number_to_words("0010.0011.0100.0111")`
`'ten point zero zero one one zero one zero zero zero one one one'` ... leading zeros and other 'dots' are lost
`>>> number_to_words("2/3")`
`'twenty-three'` ... 'two thirds' is misinterpreted or could have been a mathematical division
`>>> number_to_words("1-1/2")`
`'one hundred and twelve'` ... whole and fractional values misinterpreted
`'>>> number_to_words("4!")`
`'four'` ... 'factorial' suffix lost

### Final thoughts
The symbols that might benefit best from this library are those with pluralized written forms: dollars, feet, inches, degrees, seconds, minutes, pounds, etc., where the numeric value would be significant, e.g., "33º", '12"', "1¢", "$1", and so on.

The translation that might be most useful would be the option to translate fractional values or ratios, e.g. "2/3", "1-1/2", "3:2", "1000:1", etc.

Translating mathematical symbolism could open up another whole toolbox.

The issue names the number_to_words() entry point but no files or tests. Start by locating that API and its current parsing behavior, then narrow the broad symbol, unit, fraction, and ratio requests into a defined scope. Done would require an agreed set of supported inputs and expected outputs, with coverage for those cases.

Written by the indexing model from the issue text.