Remove absolute tolerance in floating-point comparison

The comparison of the max excess `take_from[1]` and `deficit` against
zero was initially done using absolute tolerance. However, as both
values are then compared against each other and known to be away from
zero, the absolute tolerance becomes unnecessary. The relative tolerance,
which is defined by default as 1e-09, is sufficient for this comparison.

Signed-off-by: Daniel Zullo <[email protected]>

Author: daniel-zullo-frequenz

src/frequenz/sdk/power/_distribution_algorithm.py

@@ -404,9 +404,7 @@ def _distribute_power(  # pylint: disable=too-many-arguments
                 take_from = max(excess_reserved.items(), key=lambda item: item[1])
                 if is_close_to_zero(take_from[1]) or take_from[1] < 0.0:
                     break
-                if take_from[1] >= -deficit or math.isclose(
-                    take_from[1], -deficit, abs_tol=1e-6
-                ):
+                if take_from[1] >= -deficit or math.isclose(take_from[1], -deficit):
                     excess_reserved[take_from[0]] += deficit
                     deficits[inverter_id] = 0.0
                     deficit = 0.0