…ix Python 3 multi-exception syntax in instagram_crawler

…xception — int() raises ValueError on non-numeric input, not a generic Exception

…emoved in Python 3, use parentheses instead

…oat*float can only raise TypeError for incompatible types; the bare Exception was masking real bugs from the caller

…t_to_negative

The bare 'except Exception' around mass * acceleration was catching
RecursionError, MemoryError, KeyboardInterrupt, and any unrelated bug
in surrounding code. The multiplication of two Python objects can only
raise TypeError (wrong operand types) or OverflowError (very large
floats), so narrowing to TypeError matches the documented error path
and lets other exceptions propagate as designed.

Also adds a doctest that exercises the error path with a None
argument, so the new behavior is locked in by the test suite.

Two real bugs in maths/area.py:

1. area_reg_polygon()'s isinstance(sides, int) check was passing for bool,
   because bool is a subclass of int in Python. area_reg_polygon(True, 5)
   would silently compute cot(pi/1) for a '1-sided regular polygon' and
   return a real number, when the function is meant to reject anything
   that is not a genuine integer count of sides >= 3. Added an explicit
   isinstance(sides, bool) check and a docstring note explaining the
   subclass relationship.

   Also split the original  so the type check on sides
   and the non-negative check on length are no longer mutually exclusive
   branches. Previously, sides=True and length=-1 would silently pass the
   type guard, then hit the elif and raise the wrong error.

   Added doctest cases for area_reg_polygon(True, 5) and
   area_reg_polygon(3.0, 5) so the new behavior is locked in.

2. surface_area_cuboid()'s existing test coverage happened to miss
   triples where the negative argument is the *middle* one
   (length=1, breadth=-1, height=1). Added a doctest case for that
   exact input. The if-statement itself was already correct (short-
   circuit-or with three comparands binds tighter than the
   not-yet-existing precedence issue), but the missing test meant
   nothing was actually exercising the breadth < 0 branch in
   isolation. The new case locks it in.

Verified: `python3 -m doctest maths/area.py -v` reports
'116 passed and 0 failed'.