Reveal - Setup

📦 Data Setup

1. Preprocess Your Dataset

   Go to the Reveal Preprocessing Repository
   and follow the steps there to preprocess your dataset.

2. Organize the Backend Dataset Structure

   After completing preprocessing, navigate to your Reveal backend folder and create the following structure:

/back/
├── dataset/
│   ├── llm/
│   │   ├── processed/
│   │   │   ├── 0.jpg
│   │   │   ├── 1.jpg
│   │   │   └── ...
│   │   └── thumbnails/
│   │       ├── 0.jpg
│   │       ├── 1.jpg
│   │       └── ...
│   └── files/
│       ├── unique_words_final.json
│       ├── data_final.json
│       ├── multi_clip_images_embedding.pt
│       ├── multi_clip_words_embedding.pt
│       └── multi_clip_joint_embedding.pt

✅ Make sure:

• All .jpg files are named using their numeric IDs (0.jpg, 1.jpg, etc.).
• The JSON and .pt files come directly from the Reveal Preprocessing output.

---

⚙️ Installation

1. Prerequisites

Ensure you have the following installed:

• Node.js (v22.13.0 recommended)
• yarn (npm install -g yarn)
• Angular CLI (v14.1.0)
• Python (3.9+)
• conda (recommended for managing the backend environment)

---

2. Frontend Setup

# Navigate to the frontend folder
cd front

# Install dependencies with yarn (not npm)
yarn install

# Run the frontend
ng serve

Once it compiles successfully, open your browser and visit: 👉 http://localhost:4200

3. Backend Setup

# Open a new terminal
cd back

# Create and activate a conda environment
conda create -n reveal python=3.9 -y
conda activate reveal

# Install dependencies
pip install -r requirements.txt

# Run the backend server
python app.py

The backend will start running.

🧠 Troubleshooting

Angular not recognized?

Install it globally:

npm install -g @angular/cli@14.1.0

Missing Node version?

Install Node using nvm:

nvm install 22.13.0
nvm use 22.13.0

Backend cannot find dataset files?

Double-check that your directory structure exactly matches the layout shown above and that filenames are consistent with the preprocessing output.