Fixed MNTM's implementation of read_input_as_ntm

[CamiloMartinezM]

Completed also test_read_input_as_ntm() to be able to detect whether the tapes are different after running an MNTM normally and as an NTM (which would have detected that both mntm.accepts_input(input_str)) and mntm.read_input_stepwise(input_str) weren't working in unison.

Now, the following code does work:

from automata.tm.mntm import MNTM

# MNTM which accepts all strings of the form ww, where w is in {0, 1}*
mntm = MNTM(
    states={"q0", "q1", "q2", "q3", "q4"},
    input_symbols={"0", "1"},
    tape_symbols={"0", "1", "$", "#"},
    n_tapes=3,
    transitions={
        "q0": {
            ("0", "#", "#"): [("q1", (("0", "N"), ("$", "R"), ("$", "R")))],
            ("1", "#", "#"): [("q1", (("1", "N"), ("$", "R"), ("$", "R")))],
        },
        "q1": {
            ("0", "#", "#"): [
                ("q1", (("0", "R"), ("0", "R"), ("#", "N"))),
                ("q2", (("0", "R"), ("#", "N"), ("0", "R"))),
            ],
            ("1", "#", "#"): [
                ("q1", (("1", "R"), ("1", "R"), ("#", "N"))),
                ("q2", (("1", "R"), ("#", "N"), ("1", "R"))),
            ],
        },
        "q2": {
            ("0", "#", "#"): [("q2", (("0", "R"), ("#", "N"), ("0", "R")))],
            ("1", "#", "#"): [("q2", (("1", "R"), ("#", "N"), ("1", "R")))],
            ("#", "#", "#"): [("q3", (("#", "N"), ("#", "L"), ("#", "L")))],
        },
        "q3": {
            ("#", "0", "0"): [("q3", (("#", "N"), ("0", "L"), ("0", "L")))],
            ("#", "1", "1"): [("q3", (("#", "N"), ("1", "L"), ("1", "L")))],
            ("#", "$", "$"): [("q4", (("#", "N"), ("$", "N"), ("$", "N")))],
        },
    },
    initial_state="q0",
    blank_symbol="#",
    final_states={"q4"},
)

print(mntm.accepts_input("0101"))  # This correctly returns True
for conf in mntm.read_input_as_ntm("0101"):  # This previously threw the exception
    final_config = conf.pop()
    print(final_config)

The final NTM output is the following: TMConfiguration('q4', TMTape('0101#^_$^01#_$^01#_', '#', 18))
Whereas the final MNTM output is: MTMConfiguration('q4', (TMTape('0101#', '#', 4), TMTape('$01#', '#', 0), TMTape('$01#', '#', 0))).

If we concatenate either of them, we get: 0101#^_$^01#_$^01#_ (which is what test_read_input_as_ntm() now tests).

[coveralls]

coverage: 99.613% (+0.001%) from 99.612%
when pulling 281d495 on CamiloMartinezM:fixed-mntm-read_input_as_ntm
into 9993dd1 on caleb531:develop.

[coveralls]

coverage: 99.613%. remained the same
when pulling 4042a47 on CamiloMartinezM:fixed-mntm-read_input_as_ntm
into 9ab1a1c on caleb531:main.

[CamiloMartinezM]

The linting errors are caused by doing: configs[-1].pop(), saying "AbstractSet[TMConfiguration]" has no attribute "pop"mypy[attr-defined](https://mypy.readthedocs.io/en/stable/_refs.html#code-attr-defined). Any advice on how to do this properly? @eliotwrobson
Also, can you link the PR to the #237 issue? :)

[eliotwrobson]

@CamiloMartinezM sorry for the delay! It's been a busy time of year for me. I've fixed the linting issue by adding a cast, and it looks like everything is green now. If the current changeset looks good to you still, please comment below and I can go ahead and merge this.

[CamiloMartinezM]

@eliotwrobson LGTM!