G-219: Select Leftmost Shape Data Generator

Generates synthetic tasks where multiple shapes are displayed. The goal is to identify and circle the leftmost shape.

Each sample pairs a task (first frame + prompt describing what needs to happen) with its ground truth solution (final frame showing the result + video demonstrating how to achieve it). This structure enables both model evaluation and training.

---

📌 Basic Information

Property	Value
Task ID	G-219
Task	Select Leftmost Shape
Category	Spatiality
Resolution	1024×1024 px
FPS	16 fps
Duration	~5.6 seconds
Output	PNG images + MP4 video

---

🚀 Usage

Installation

# Clone the repository
git clone https://github.com/VBVR-DataFactory/G-219_select_leftmost_shape_data-generator.git
cd G-219_select_leftmost_shape_data-generator

# Install dependencies
pip install -r requirements.txt

Generate Data

# Generate 100 samples
python examples/generate.py --num-samples 100

# Generate with specific seed
python examples/generate.py --num-samples 100 --seed 42

# Generate without videos
python examples/generate.py --num-samples 100 --no-videos

# Custom output directory
python examples/generate.py --num-samples 100 --output data/my_output

Command-Line Options

Argument	Type	Description	Default
--num-samples	int	Number of samples to generate	100
--seed	int	Random seed for reproducibility	Random
--output	str	Output directory	data/questions
--no-videos	flag	Skip video generation	False

---

📖 Task Example

Prompt

Multiple shapes are shown. Circle the leftmost one. Do not change anything else.

Visual

Initial Frame Multiple shapes displayed	Animation Red circle expands around leftmost shape	Final Frame Leftmost shape circled

---

📖 Task Description

Objective

Identify the shape with the leftmost position (minimum x-coordinate) among multiple shapes and circle it.

Task Setup

• Shape Count: Multiple shapes per image (typically 3-8)
• Shape Types: Various geometric shapes (circles, squares, triangles, polygons)
• Positioning: Shapes distributed across the canvas with varying horizontal positions
• Marking Method: Red circle expands around the leftmost shape
• Criterion: Leftmost determined by minimum x-coordinate (left edge or center)
• Unique Solution: Only one shape is leftmost per image

Key Features

• Spatial reasoning: Tests ability to identify leftmost position along horizontal axis
• Position comparison: Requires comparing x-coordinates of multiple objects
• Shape-agnostic: Task focuses on position, not shape type
• Clear animation: Red circle smoothly expands around target shape
• Unambiguous answer: Only one shape occupies the leftmost position
• Visual clarity: Clean layout ensures easy position comparison

---

📦 Data Format

data/questions/select_leftmost_shape_task/select_leftmost_shape_00000000/
├── first_frame.png      # Initial state (shapes without marking)
├── final_frame.png      # Goal state (leftmost shape circled)
├── prompt.txt           # Task instructions
├── ground_truth.mp4     # Solution video (16 fps)
└── question_metadata.json # Task metadata

File specifications: Images are 1024×1024 PNG. Videos are MP4 at 16 fps, approximately 5.6 seconds long.

---

🏷️ Tags

spatial-reasoning position-identification leftmost-selection coordinate-comparison shape-selection directional-reasoning

---