v2.32.2

v2.32.2

Fixes fusion_comp(action="get_keyframes") serialization — a follow-up to the add_keyframe fix in v2.32.1.

Fixed

• get_keyframes iterated Input.GetKeyFrames() as if it returned {time: value}, but Fusion returns {1-based index: frame_position}. The result put the keyframe index in time and the frame position in value — the actual keyframed values were never reported. The handler now treats the dict values as frame positions and reads each keyframed value back via GetInput(input_name, frame), returning [{"time": <frame>, "value": <value>}, ...] in frame order.

Validation

• Static checks (import, API parity, CLI --version, npm pack --dry-run, whitespace) clean.
• Unit tests pass (tests/test_fusion_comp_targeting.py, 10 tests).
• Live-validated on DaVinci Resolve Studio 21.0.0: a Transform Size keyed 1.0@f0 / 1.4@f75 now serializes as [{"time": 0.0, "value": 1.0}, {"time": 75.0, "value": 1.4}].

Flagged by @sandypoli-boop in #56.