SeisFlowBench: A Benchmarking Dataset for Seismic and Flow Simulations

This repository provides a reproducible scientific project built using Julia and DrWatson.jl.
It is designed for benchmarking seismic and flow simulations, ensuring consistency across computational environments.

📌 Project Overview

• Project Name: SeisFlowBench
• Authors: Haoyun Li and Abhinav Gahlot
• Focus Areas:
  • Seismic Wave Propagation
  • Fluid Flow in Porous Media
  • Seismic-Flow Coupling Simulations
• Technologies Used:
  • Julia – A high-performance programming language for scientific computing.
  • DrWatson – A framework for reproducible scientific projects.

---

🚀 Getting Started

To reproduce this project locally, follow these steps:

1️⃣ Clone the Repository

Download the codebase using Git:

git clone https://github.com/your-repo/SeisFlowBench.git
cd SeisFlowBench

🚨 Note: Raw data files are not included in version control (Git) and may need to be downloaded separately.

2️⃣ Set Up the Julia Environment

Open a Julia REPL and run the following commands:

using Pkg
Pkg.add("DrWatson")   # Install globally (needed for `quickactivate`)
Pkg.activate(".")      # Activate this project's environment
Pkg.instantiate()      # Install all required dependencies

✅ This will install all necessary Julia packages for the project to run correctly.

---

🛠 Using the Project

Many scripts in this repository begin with:

using DrWatson
@quickactivate "SeisFlowBench"

This ensures that:

• The correct project environment is activated.
• Local paths are managed automatically using DrWatson.

After activation, you can run any script or experiment without manually setting paths.

---

📂 Project Structure

SeisFlowBench/
│── data/           # Processed and simulation data (not included in Git)
│── scripts/        # Main simulation and analysis scripts
│── src/            # Core project modules and functions
│── notebooks/      # Jupyter/Pluto notebooks for interactive analysis
│── Project.toml    # Julia package dependencies
│── Manifest.toml   # Exact dependency versions for reproducibility
│── README.md       # This file

---

🔄 Converting JLD2 to HDF5

Simulation results are stored in JLD2 format, but you may need them in HDF5 for better compatibility.

📌 Convert JLD2 to HDF5

To convert all samples into a single HDF5 file, use the function provided in:
📄 scripts/convert_jld2_hdf5.jl

Run the script as follows:

include("scripts/convert_jld2_hdf5.jl")
convert_all_samples_to_hdf5("2D_perm_perturb", 1:10, "all_samples.h5")

✅ This will save all data in one HDF5 file under /data/2D_perm_perturb/all_samples.h5.

---

📥 Loading Data from HDF5

To load and inspect simulation data from HDF5, use the provided function in:
📄 scripts/load_hdf5.jl

Example usage:

include("scripts/load_hdf5.jl")
hdf5_file = datadir("2D_perm_perturb", "all_samples.h5")
data = load_hdf5_data(hdf5_file)

# Accessing a specific variable
S_array_sample_1 = data["sample_1"]["S_array"]
println("Loaded S_array from sample_1: Size = ", size(S_array_sample_1))

✅ This allows easy retrieval of seismic and flow simulation results.

---

📌 Additional Notes

• If any dependency is missing, run Pkg.instantiate() again.
• For issues, create a GitHub issue or contact the authors.
• This project follows reproducibility best practices using DrWatson.

---

✅ SeisFlowBench is now ready for seismic and flow simulation benchmarking! 🚀