Veriminder Quick Start Guide

This guide provides the essential steps to get the Veriminder application running locally.

Prerequisites

• Python 3.8+ & venv
• MySQL Server
• Git

---

Setup Instructions

1. Clone & Install Dependencies

   # Clone the repository
   git clone <your-repo-url>
   cd Veriminder_Submission
   
   # Create and activate a virtual environment
   python3 -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   
   # Install required packages
   pip install -r requirements.txt

2. Download BIRD Database

   • Download the BIRD development dataset from the BIRD Benchmark Website.
   • Update the references to the SQLITE_DB_PATH to the dataset location in your environment.

3. Setup MySQL Database

   • Ensure your MySQL server is running.
   • Create a new database.
   • Run the database_setup.sql script to create the necessary tables and seed data.

4. Configure Credentials

   • Before running, you must update the DB_CONFIG dictionary in all relevant .py files with your MySQL credentials.
   • You also need to add your Google and Anthropic API keys (needed only if you plan to generate new prompt templates) in the same files.
   • Files to update are located in ui/services/, prompt_builder/, and query_models/.

5. Regenerate Prompts (Optional)

   • If you wish to create new AI prompts, run the candidate and critic modules located in the /prompt_builder directory.

     # Example
     python prompt_builder/12_candidates.py
     python prompt_builder/3_critics.py
     python prompt_builder/run_candidate_baqr_prompt.py
     python prompt_builder/run_critics_on_candidate_data.py
     python prompt_builder/moe_prompt_builder.py

6. Run the Flask Webserver

   • Navigate to the ui directory and run the healthcheck.py script.

     cd ui
     python healthcheck.py

7. Access the Application

   • Open your web browser and go to: http://localhost