Update tests for monoid output for orientation preserving and reversing monoids

[MTWhyte]

This PR updates the tests for orientation_preserving_monoid and orientation_reversing_monoid to work with monoid presentation output, rather than semigroup presentation.

[MTWhyte]

The CI tanked, but we expected this, right?

[james-d-mitchell]

Thanks @MTWhyte, yes I think I'd expect the CI to tank. Perhaps the test can be made more robust using something similar to:

libsemigroups_pybind11/tests/test_present.py

Lines 631 to 641 in ef498fb

	if (
	compare_version_numbers(libsemigroups_version(), "2.5.2")
	or p.letter(0) == "a"
	):
	assert p.letter(0) == "a"
	assert p.letter(1) == "b"
	assert p.letter(2) == "c"
	else:
	assert p.letter(0) == "\x00"
	assert p.letter(1) == "\x01"
	assert p.letter(2) == "\x02"

This way we can merge and everything will be ok!

[MTWhyte]

Thanks @MTWhyte, yes I think I'd expect the CI to tank. Perhaps the test can be made more robust using something similar to:

libsemigroups_pybind11/tests/test_present.py

Lines 631 to 641 in ef498fb

	if (
	compare_version_numbers(libsemigroups_version(), "2.5.2")
	or p.letter(0) == "a"
	):
	assert p.letter(0) == "a"
	assert p.letter(1) == "b"
	assert p.letter(2) == "c"
	else:
	assert p.letter(0) == "\x00"
	assert p.letter(1) == "\x01"
	assert p.letter(2) == "\x02"

This way we can merge and everything will be ok!

I think we touched on this the other day, I understand presentation.make gives a string presentation, whereas fpsemigroup.make gives a word-type presentation. Just to check: since I'm using the latter here, should this test be independent of the version? Or am I mistaken and in fact make from both namespaces will use strings?