isnumb - a statistical approach to Python's isnumeric

Is "$100,000,001" a number?

I say True, Python says...

>>> "$100,000,001".isnumeric()
False

Approach to implementation

I imagine two ways are possible:

1. supervised learning by training

2. model by heuristics

By training

You have a training (.txt) file of what humans consider to be "numbers" :

$100,000,001
185774
123,456,789
1.01
%99.991
$9.11
100,000.00
%10

If classification resolution is desired, we could create multiple training files where each file has numbers of specific classes:

# comma separated (with decimals)
123,456,789
100,000.00

# currencies
$100,000,001
$9.11

# percentages
%99.991
%10

By heuristics

With heuristics, the task changes from having to create training files to having to describe rules that classify numbers.

For comma separated values, simply replacing commas with empty strings can suffice.

For currencies, we can conditionally check for currency symbols (ie. $, €, etc.). The same principle can be applied for percentages.