fix(naturaldelta): catch OverflowError for float('inf') (#333)

`int(float("inf"))` raises `OverflowError`, not `ValueError`, so the
existing `except (ValueError, TypeError)` guard misses it. The function
docstring states that non-finite floats are returned unchanged, but
`float("inf")` and `float("-inf")` raised uncaught `OverflowError`
while only `float("nan")` was silently returned.

Add `OverflowError` to the except clause so that ±inf are treated the
same as nan. Remove the misleading "Raises: OverflowError" note from
the docstring since that exception is now caught.

Author: Mukller

src/humanize/time.py

@@ -117,9 +117,6 @@ def naturaldelta(
             converted to int (cannot be float due to 'inf' or 'nan').
             In that case, a `value` is returned unchanged.
 
-    Raises:
-        OverflowError: If `value` is too large to convert to datetime.timedelta.
-
     Examples:
         Compare two timestamps in a custom local timezone::
 
@@ -151,7 +148,7 @@ def naturaldelta(
             int(value)  # Explicitly don't support string such as "NaN" or "inf"
             value = float(value)
             delta = dt.timedelta(seconds=value)
-        except (ValueError, TypeError):
+        except (ValueError, TypeError, OverflowError):
             return str(value)
 
     use_months = months