num2words | words converter

There are two problems:

1. Your regular expression isn't correctly matching the suffix, only the first letter of the suffix. [st|nd|rd|th] matches exactly one of the characters inside the brackets; duplicates are ignored, so it's equivalent to [st|ndrh], with the | treated as a character to match like each of the letters. Use r"(\d+ )(?:st|nd|rd|th)" instead; the | inside the non-capture group (?:...) does work to separate the 4 patterns st, nd, rd, and th.

2. The callable passed to re.sub takes a Match object as its argument. You need to use its group method to extract the captured sting. lambda x: num2words.num2words(x.group(1), ordinal=True).

There is a very quick way to do it in one line using regex to match digits and replace them in string:

With perl,

You can try this approach

• get all the digits in a string using regex [\d]+
• cast the match to integer
• use modulo(%) operator to get the digits from the number
• finally use your dictionary to get the number in word

Alternate solution

• use re.findall(r'[\d]', my_string)
• this will give you all possible digits
• next just use my_string.replace(digit, f' {conversion_dict[digit]} ' )

This is a known issue with cpython on windows. The error occurs when black tries to run multiple workers on >60 core machines because the default number of process workers given by os.cpu_count() breaks some other windows limit (number of waiting processes? I'm not quite sure). Black >=19.10b0 has a fix for this, so try updating the version of black in your pre-commit config if you can?

• Python bug report: https://bugs.python.org/issue26903
• Fix applied by this PR in black: https://github.com/psf/black/pull/838

Writing rupees, paise = x.split('.') only works if x contains exactly one '.'. You can simply ignore NaNs:

You can use num2words module -

You can simplify your replace_ordinal_numbers function by using re.sub and calling num2words in a lambda function as the replacement. Then just use DataFrame.apply to run the function over the column:

Your program is missing a lot of information. This is not a problem because I have been where you are. You are missing some lines of code. Instead of importing things like the say function or response, here is a working and simpler alternative.