fix(matrix): make is_safe() return bool and prove count_islands' improper results

[priya-sundaram-dev]

Describe your change:

Follow-up to the discussion on #12691, as requested by @cclauss — a focused PR that (1) makes Matrix.is_safe() -> bool actually return a bool, and (2) adds count_islands() test cases that pin down where the algorithm was giving improper results.

The bug. is_safe ended in and self.graph[i][j], so on a safe cell it returned the raw cell value (an int), not a bool — hence the old doctests asserted 1/0. More than cosmetic: is_safe used a truthy test while count_islands seeds islands only on cells == 1. The two disagree on any cell whose value isn't 0/1.

The fix. End with and self.graph[i][j] == 1. This always returns a bool and aligns is_safe with count_islands' seeding rule.

Proof of improper results (new doctests). For [[1, 2, 1]] the two 1-islands are bridged by a non-island 2. The old truthy is_safe absorbed the 2 and merged them, returning 1; the corrected version returns 2. Also cover out-of-bounds, diagonal-only adjacency, and a lone 2.

Checklist:

• Add an algorithm? No — bug fix + tests.
• All new Python files are placed inside an existing directory.
• All filenames are in all lowercase characters with no spaces or dashes.
• All functions and variable names follow Python naming conventions.
• All function parameters and return values are annotated with Python type hints.
• All functions have doctests that pass the automated testing.
• All new algorithms include at least one URL that points to Wikipedia or another similar explanation.
• If this pull request resolves one or more open issues then the description above includes the issue number(s).

Contributes to #9943

@akiels