Rosbag Performance Benchmark

Result

Dependencies

Ensure you have the following Python packages installed:

pip install rosbag matplotlib pandas tqdm numpy

The rosbag_io_py package is required for testing the C++ implementation.

Usage

Single Benchmark Run

1. Run the benchmark:

python rosbag_benchmark.py path/to/your/bagfile.bag [--output results.json]

2. Visualize results:

python visualize.py results.json [--output-dir benchmark_results]

Multiple Runs with Statistical Analysis

Run multiple benchmarks and generate statistical analysis:

python run_benchmarks.py path/to/your/bagfile.bag [--runs N] [--output-dir DIR]

Options:

• --runs: Number of benchmark runs (default: 5)
• --output-dir: Directory to save results (default: benchmark_stats)

Output Files

Single Run Output

1. Detailed test results in JSON format
2. Processed bag file using Python implementation
3. Processed bag file using C++ implementation (if available)
4. Visualization results:
   • Processing time comparison chart
   • Message throughput comparison chart
   • Detailed Markdown report

Statistical Analysis Output

1. Individual JSON results for each run
2. Statistical report including:
   • Mean processing time
   • Standard deviation
   • Min/max times
   • Median time
3. Box plot showing time distribution
4. Performance comparison between implementations

Example

# Run a single benchmark
python rosbag_benchmark.py demo.bag

# Run multiple benchmarks with statistics
python run_benchmarks.py demo.bag --runs 10

# Visualize single run results
python visualize.py benchmark_results.json

Notes

1. Ensure sufficient disk space as output files will be similar in size to input
2. Processing large bag files may take considerable time
3. If rosbag_io_py is not available, only the Python implementation will be tested