iGait ASD - Monorepo

This monorepo contains all the core components of the iGait ASD system:

• igait-backend - Backend server handling submission, upload, storage, and processing
• igait-lib - Shared library code used across the project
• igait-pipeline - Gait analysis pipeline for processing video data

Documentation

It is highly recommended to read the documentation as onboarding before contributing to this project.

Installation & Usage

Installation (One-Time Setup)

To install the iGait pipeline on your account, run:

/lstr/sahara/zwlab/jw/igait-pipeline/install.sh

The installation script will:

1. Verify the pipeline is built and ready
2. Set up the module system for your account
3. Update your ~/.bashrc automatically

After installation, either:

• Run source ~/.bashrc to activate immediately, OR
• Start a new terminal session

Using the Pipeline

Once installed, you can run the pipeline from anywhere:

# Load the module (only needed once per session)
module load igait

# Run the pipeline
igait-pipeline --input-path-front front.mp4 --input-path-side side.mp4 --output-dir-path ./output

Example:

module load igait
igait-pipeline \
  --input-path-front /path/to/front_video.mp4 \
  --input-path-side /path/to/side_video.mp4 \
  --output-dir-path /path/to/output

The pipeline will process the videos through 7 stages:

1. Media Conversion
2. Validity Check
3. Reframing
4. Pose Estimation (OpenPose)
5. Cycle Detection
6. Prediction (ML Model)
7. Archive

Additional Options

# Skip to a specific stage (useful for debugging)
igait-pipeline --skip-to-stage 5 --input-path-front ... --input-path-side ... --output-dir-path ...

# Show help
igait-pipeline --help

---

For Developers

Building from Source

# Build all workspace members
cargo build --release

# Build specific component
cargo build --release -p igait-pipeline
cargo build --release -p igait-backend

# Run tests
cargo test --workspace

Using Nix Flakes

If you have Nix with flakes enabled:

# Build the backend
nix build .#igait-backend

# Build the Docker image
nix build .#docker
docker load < result

# Enter development shell
nix develop

Technologies Used

• ⚡ - Rust and Cargo
• 🔭 - Nix Flakes
• 🌱 - Docker
• 🌟 - Cloudflare
• 🕖 - Nginx
• ⚡ - Google Firebase Authentication
• 🎹 - Google Realtime DB
• 🔭 - Amazon S3 and EC2 (m6in.large)

Service Map

Development Team

• John W - Head Backend and Systems Engineer
• Michael S - Head Frontend Engineer