An interactive, containerized web application for visualizing and exploring Hierarchical All-against-All (HAllA) association testing results.
This app provides a streamlined Streamlit interface to execute HAllA, dynamically explore the resulting similarity matrices using Plotly, zoom in on significant feature blocks, and export publication-ready PDFs.
To ensure complete reproducibility and avoid dependency conflicts across different operating systems, this application is fully containerized.
- Docker Desktop must be installed and running on your machine.
1. Clone the repository
git clone [https://github.com/BenaroyaResearch/halla_interactive.git](https://github.com/BenaroyaResearch/halla_interactive.git)
cd halla_interactive2. Build the Docker Image
docker build -t halla_interactive:latest .3. Run the Application
docker run --rm -p 8501:8501 -v "$(pwd):/app" halla_interactive:latest streamlit run app.py4. View the App Open your web browser and navigate to http://localhost:8501.
-Upload Data: Use the sidebar to upload your tab-delimited X and Y datasets. (Leave blank to run the included toy datasets).
-Configure Settings: Select your correlation metric (Spearman, Pearson, or Xicor) and define the FDR Alpha threshold.
-Execute: Click "Run HAllA". The computation will securely run inside the Linux container.
-Explore: Use the interactive Plotly heatmap to hover over data points, drag to zoom into specific clusters, or use the "Top Blocks" sub-setting feature to isolate the most significant associations.
-Export: Download the summary statistics or the native, publication-ready PDF and PNG hallagrams directly from the UI.
This application utilizes a Linux miniconda3 base image to bridge the R-Python execution requirements of the core halla library, completely bypassing macOS dynamic library linking issues. The frontend is built entirely in Streamlit.