I have tried to document everything in the file itself with comments.

• config.ini - configuration file for several different scripts... The most important part is the txt filename, as the rest has gotten a bit scrambled in the debugging process.
• event_simulator_2d_plane.py - warp class necessary for the entire project... probably most important file
• loss_plot_figs.ipynb - file to view the generated loss plots
• loss_plots_h5.py - generates the data needed for the loss plots
• ring_loss.py - generates the annular loss data (do this after running the loss_plots_h5.py script)
• utils/extract_events_from_rosbag.py - I think I got this from GitHub I can't even remember now but it goes from .bag to .txt

I used this repo setup in Docker. I used docker compose to set up ros then I used a Dockerfile to make it Kinetic ros with this image.

If using real data, skip down to the next section.

Here's the general process I followed to generate the rosbags:

1. Generate mkv file with blender
2. Use ffmpeg to make a folder of images from that video
3. Run something like the following in that docker container:

ssim
roscd esim_ros
python scripts/generate_stamps_file.py -i /root/evodo/plane_depth_2_pitch_-5_roll_15/imgs_wavy/ -r 1200
rosrun esim_ros esim_node \
 --data_source=2 \
 --path_to_output_bag=/root/evodo/plane_depth_2_pitch_-5_roll_15/plane_depth_2_pitch_-5_roll_15_wavy.bag \
 --path_to_data_folder=/root/evodo/plane_depth_2_pitch_-5_roll_15/imgs_wavy/ \
 --ros_publisher_frame_rate=30 \
 --exposure_time_ms=10.0 \
 --use_log_image=1 \
 --log_eps=0.1 \
 --contrast_threshold_pos=0.2 \
 --contrast_threshold_neg=0.2

1. Run loss_plots_h5.py. This will generate the warped histograms.
2. Run ring_loss.py. This will calculate the loss and ring masks.
3. View results using loss_plot_figs.ipynb.