Queryshield Implementation

Table of Contents

• Queryshield Implementation
  • Table of Contents
  • Description
  • Tech Stack
  • Functionalities
  • Get Started
    • 1) Configure environment variables
    • 2-1) Deploy on Docker
    • 2-2) Deploy on local
  • File Structure
  • Database Schema
    • User
      • user_role Enum
      • user table
    • Analysis
      • analysis_status Enum
      • analysis table
      • analysis_owners table

Description

The repo is the frontend interface of Queryshield, a secure multiparty computation (MPC) cloud service. It leverages the power of cryptographic techniques to enable collaborative computation among multiple parties without compromising the privacy of their data.

Tech Stack

• Frontend: Streamlit, Python
• Database: PostgreSQL
• Authentication: JWT library

Functionalities

• User Registration & Authentication
  • Securely manage user registration and authentication with JWT-based token handling for enhanced security and session management.
• Analyst Features
  • Create Analysis Page: allows analysts to create new analyses with custom schemas, thread modes, and SQL queries.
  • Analysis History Page: allow analyst to trace his/her analysis status.
• Data Owner Features
  • Analysis Catalog Page: allows data owners to view all available analyses, providing an entry point for data sharing.
  • Share Data Page: allow data owner to do secret data sharing using MPC.
• Analysis Detail Page: wiew detailed metadata about specific analyses.

Get Started

1) Configure environment variables

1. Create .env at the root:

   touch .env

2. Edit .env file with these variables:

   JWT_SECRET_KEY=secret
   DATABASE_URL=postgresql+psycopg2://user1:12345678!@host.docker.internal:5432/storage
   DATABASE_URL_VERIFICATION=postgresql+psycopg2://user1:12345678!@host.docker.internal:5432/verification
   POSTGRES_USER=user1
   POSTGRES_PASSWORD=12345678!
   POSTGRES_DB=storage
   POSTGRES_DB_VERIFICATION=verification

2-1) Deploy on Docker

The following instruction is for running the app on Docker, if you want to deploy on your local machine, you can go to 2-2) Deploy on Local

1. Start Docker.

2. Open a terminal and navigate to the scripts folder":

   cd scripts

3. Run the following commands:

   chmod +x start.sh
   ./start.sh

2-2) Deploy on local

You can also run the app on your local machine, rather than using Docker. The following is the instruction.

1. Create and activate a Python virtual environment:

   python -m venv venv
   source venv/bin/activate

2. Install all dependencies

   pip install -r requirements.txt

3. Compile the custom streamlit React component

   cd src/secret_share_component
   npm install
   npm run build

4. Run python ./src/Create_Analysis.py

   cd ../..
   python ./src/Create_Analysis.py

File Structure

.
├── Dockerfile                   # Docker configuration file
├── LICENSE                      # License information
├── README.md                    # Project documentation
├── docker-compose.yml           # Docker Compose configuration
├── requirements.txt             # Python dependencies
├── setup.py                     # Python app setup file
├── scripts                      # Shell scripts
│   ├── cleanup.sh               # Script to remove the entire app from Docker
│   └── start.sh                 # Script to start/restart the app
├── sql                          # SQL scripts
│   ├── db_setup.sql             # Database setup script
│   └── db_storage_setup.sql     # Database storage setup script
├── secret_share_component/      # React.ts secret data sharing component
└── src                          # Source code
    ├── Create_Analysis.py       # Main app (entry point)
    ├── components/              # UI components
    ├── configs/                 # Configuration files
    ├── db/                      # Database connection and operations
    ├── models/                  # Data models
    ├── pages/                   # Application pages
    └── utils/                   # Utility functions

Database Schema

User

user_role Enum

• analyst
• data_owner

user table

Column Name	Column Type
uid	SERIAL ->PK
first_name	VARCHAR(50)
last_name	VARCHAR(50)
email	VARCHAR(50)
pin	VARCHAR(50)
role	user_role

Analysis

analysis_status Enum

• Created
• Ready
• Running
• Failed
• Completed

analysis table

Column Name	Column Type
aid	SERIAL ->PK
analysis_name	TEXT NOT NULL
analyst_id	INTEGER REFERENCES user(uid)
time_created	datetime DEFAULT NOW()
details	JSONB NOT NULL
owners_registered	SERIAL[]
status	analysis_status NOT NULL

analysis_owners table

Column Name	Column Type
analysis_id	INTEGER REFERENCES analysis(user_id)
user_id	INTEGER REFERENCES user(aid)