Behavioral Annotation Data Conversion

A Python tool for converting behavioral annotation data from Drosophila fly videos (annotated via ImageJ) into a structured and analysis-friendly format.

Features

• Converts ImageJ CSV files to structured timeline and event data
• Generates visualizations of grooming events
• Produces summary statistics
• Provides batch processing capabilities

Installation

Prerequisites

• Python 3.x
• Conda (for environment management)

Setup

1. Clone this repository
2. Create the conda environment: conda env create -f environment.yml
3. Activate the environment: conda activate fly_analysis
4. Install required packages: conda install pandas numpy matplotlib pytest

Usage

Run the script with the following command:

python main.py --input <input_path> --output <output_path> [--total_frames ]

Arguments

• --input: Path to input CSV file or directory containing CSV files (required)
• --output: Path to output directory for storing results (required)
• --total_frames: Total number of frames to consider (default: 8999)

Example

python main.py --input data/annotations/ --output results/fly_analysis --total_frames 10000

Output Structure

For each processed file, the following outputs are generated:

• Timeline Table: <filename>_timeline.csv
• Event List: <filename>_events.csv
• Timeline Plot: <filename>_timeline.png
• Box Plot: <filename>_boxplot.png

Additionally, the tool generates:

• Consolidated summary report
• Error log (if applicable)

Testing

Run tests using pytest: