# Motion Analysis LLM Prompt Template
# Converts DSL motion data into a concise, LLM-friendly format

You are analyzing motion tracking data from a video stream.

## Scene Summary
- Duration: {duration}s ({frame_count} frames)
- Camera: {camera_info}
- Average motion: {avg_motion}%

## Tracked Objects ({object_count} total)
{object_summaries}

## Movement Timeline
{timeline}

## Key Events
{key_events}

## Question
Based on this motion data, {question}

---
# Example filled prompt:

You are analyzing motion tracking data from a video stream.

## Scene Summary
- Duration: 20s (4 frames at 5s intervals)
- Camera: Security camera, indoor
- Average motion: 27.1%

## Tracked Objects (5 total)
- Object #1: Moved RIGHT→DOWN→LEFT, traveled 0.19 units, avg speed 0.047
- Object #2: Appeared F2, moved DOWN→UP, traveled 0.08 units
- Object #3: Appeared F2, moved DOWN→UP, small displacement
- Object #4: Appeared F4, stationary
- Object #5: Appeared F4, stationary

## Movement Timeline
F1 (0s): Large motion detected (100%), 1 blob appeared
F2 (10s): Motion 2.3%, blob#1 moved RIGHT, 2 new objects appeared
F3 (15s): Motion 4.3%, all 3 objects moving DOWN
F4 (20s): Motion 1.9%, objects moving UP, 2 new stationary objects

## Key Events
- APPEAR: 5 objects total (1 at F1, 2 at F2, 2 at F4)
- MOVEMENT: Primary object #1 traveled longest path (0.19 units)
- DIRECTION: Initially RIGHT, then DOWN, finally LEFT (possible person walking)

## Question
Based on this motion data, describe what is likely happening in the scene. Is there a person present? What are they doing?
