ACG

Demo Creation Process & Reproduction Guide:

First, run run_simulation.py using the following command:

python run_simulation.py --scene ./data/scenes/fluid_rigid_coupling1.json

This will produce a directory called fluid_rigid_coupling1_output and in each subdirectory of it there is a .png file.

We can then generate a .mp4 file using make_video.py with the following command:

python make_video.py --input_dir ./fluid_rigid_coupling1_output --image_name raw_view.png --output_path video.mp4 --fps 30

This will produce a file called video.mp4.

We change the file name to video_for_fluid_rigid_coupling1.mp4 in the demo directory.

We can also produce .obj files for rendering. To do this, enter the data/scenes/fluid_rigid_coupling1.json file and in "Configuration" change "exportPly" and "exportObj" to be true. Then run the first command again. After that, run surface_reconstruction.py using the following command:

python surface_reconstruction.py --input_dir ./fluid_rigid_coupling1_output

After that, make a scene using blender. Since this scene.blend file we configured is rather large (over 100MB) and cannot be pushed to GitHub, we only reserve it in our local directories.

Run render.py to produce a directory output_frames.

python render.py

Finally, run make_video_with_blender.py:

python make_video_with_blender.py --input_dir ./output_frames --output_path rendered_video_for_fluid_rigid_coupling1.mp4

This produces the output video rendered_video_for_fluid_rigid_coupling1.mp4 in demo.

We can also run the second demo:

python run_simulation.py --scene ./data/scenes/fluid_rigid_coupling2.json
python make_video.py --input_dir ./fluid_rigid_coupling2_output --image_name raw_view.png --output_path video.mp4 --fps 30

We change the name of the video and get video_for_fluid_rigid_coupling2.mp4 in demo.

Declarations:

Part of the data/models and data/scenes in this project uses the data from jason-huang03/SPH_Project.

The run_simulation.py and config_builder in this project references the implementation from erizmr/SPH_Taichi.

The surface_reconstruction.py in this project uses splashsurf.

The make_video.py and boundary.py in this project references the implementation from jason-huang03/SPH_Project.

The whole structure idea was referencing the framework of jason-huang03/SPH_Project, and we implement the codes based on the paper or tutorial Smoothed Particle Hydrodynamics: Techniques for the Physics Based Simulation of Fluids and Solids. Later advanced implementation and functionalities will be learned based on the papers suggested in this tutorial.