Hangman('E', 'everything') ==> 'e_e_______'

Random tests consistently generate cases where at least two characters are not present. In fact, most of the test cases generate strings composed of underscores ('_') with random lengths.

Problems with typography and Markdown formatting:

• Use # heading not ###heading###
• No need to write headings in uppercase and followed by colon

The description should say what to do about corner cases:

• empty string given as input
• invalid locomotive position or multiple locomotives e.g. "_A__B"
• other characters in the input

I think the description should simply say:

You can assume that input will only contain valid trains (i.e. one locomotive followed by coaches).

It is not clear from the description that the tests expect the output to be entirely lowercase regardless of the case of the input. I would suggest clarifying this in the description.

The trays are collor coded

collor -> color

No random tests

Traceback (most recent call last):
  File "tests.py", line 1, in <module>
    from solution import Hangman
ImportError: cannot import name 'Hangman' from 'solution' (/workspace/default/solution.py)

Why is the function called bob ?

This series converges to e-1.

e = \sum_{k=0}^{\inf} 1/k! starting fom 0 (not 1).

Main issue:

Floats have limited precision, and are unable to exactly represent all needed values (only 18 different "e values" can be represented).
https://docs.codewars.com/authoring/kata/floating-point/
https://docs.python.org/3/tutorial/floatingpoint.html

It's better to use Fraction or Decimal.

Other issues:

• Floats should be compared with test.assert_approx_equals
• Floats should be summed with fsum, or at least from smaller to bigger.

purposley should be purposely

Even though I think that your disclaimer is a bit useless, anyone who wants more test cases can just run the random tests.

Binary is a string made of integers, not a decimal number

1. Having multiple input types is bad. (List[Any] vs List[List[Any]])
2. It is not mentioned, if items of input of type List[Any] can be always enumerated. Some solutions don't check that.
3. Random tests only have List[List[int]], which is inconsistent with sample tests

In general, it would be beneficial to have inputs of one type and to have type definitions in initial solution.
Though, there is 99% probability that this kata will receive issue trivial filter/map/reduce is not a novel idea :P

There are no example tests in the attempt section

There is only one value in the initial solution