Skip to content

fix(aggregators): report original keyframe buffer size in VLM diff summary#800

Open
Bartok9 wants to merge 1 commit into
RobotecAI:mainfrom
Bartok9:fix/img-vlm-diff-keyframe-count
Open

fix(aggregators): report original keyframe buffer size in VLM diff summary#800
Bartok9 wants to merge 1 commit into
RobotecAI:mainfrom
Bartok9:fix/img-vlm-diff-keyframe-count

Conversation

@Bartok9

@Bartok9 Bartok9 commented Jul 14, 2026

Copy link
Copy Markdown

Purpose

ROS2ImgVLMDiffAggregator.get() reported keyframes as selected from len(selected) twice, so when the buffer had more than 3 images the status text wrongly claimed e.g. "3 keyframes selected from 3 last images".

Proposed Changes

  • Capture original buffer length before first/middle/last selection.
  • Docstring typo: midden → middle.
  • Unit tests for get_key_elements and the five-image summary string.

Testing

  • New unit tests in tests/aggregators/test_ros2_logs_aggregator.py (ROS env / existing aggregator fixture conventions).
  • Spec: aerial dual-agent specs/rai/2026-07-14-1.md (Composer 2.5 Standard posture; Grok-spec + review micro-diff).

Duplicate check

No open PR for this summary-string fix.

…mmary

Keep original image-buffer length before first/middle/last selection so
status text does not claim keyframes were selected from the post-downsample
count. Fix docstring typo midden→middle. Add unit coverage for
get_key_elements and the five-image case.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant