IQuerio: AI-Powered SQL Optimizer & Vector Playground

IQuerio is a next-gen database assistant (Phase 1 MVP) for devs and AI builders. It optimizes SQL queries, performs vector-based similarity searches, and supports natural language queries using Postgres with pgvector and a FastAPI backend.

Features

Discover IQuerio’s AI-powered tools in the Features Overview.

Setup

1. Clone repo: git clone https://github.com/pranavkp71/IQuerio
2. Install PostgreSQL dev libraries: sudo apt install libpq-dev python3-dev
3. Create a virtual env: python3 -m venv .venv and activate it using source .venv/bin/activate
4. Install dependencies: pip install -r requirements.txt
5. Start Postgres with pgvector: docker-compose up -d
6. Set up DB: python backend/setup_db.py
7. Install CLI: pip install .
8. Run server: uvicorn backend.main:app --reload

Usage

• SQL Optimizer:

  • CLI: db-toolkit optimize --query "SELECT * FROM users WHERE age + 1 > 30"
  • API: curl -X POST "http://127.0.0.1:8000/optimize" -H "Content-Type: application/json" -d '{"query": "SELECT * FROM users WHERE age + 1 > 30"}'
  • Output: {"query": "...", "optimized_query": "SELECT id, name FROM users WHERE age > 29", "issues": [...], "suggestions": [...], "explain_plan": [...]}

• Vector Playground:

  • Upload: db-toolkit upload --user-id 1 --description "AI researcher"

  • API: curl -X POST "http://127.0.0.1:8000/upload-embedding" -H "Content-Type: application/json" -d '{"user_id": 1, "description": "AI researcher"}'

  • Output: {"status": "Embedding uploaded successfully", "user_id": 1}

  • Search: db-toolkit search --description "Tech enthusiast into AI" --limit 2

  • API: curl -X POST "http://127.0.0.1:8000/search-similar" -H "Content-Type: application/json" -d '{"description": "Tech enthusiast into AI", "limit": 2}'

  • Output: {"results": [{"id": 1, "name": "Soman", "description": "...", "distance": 0.66}, ...]}

  • NL Query: db-toolkit nl-query --query "Show users over 30 similar to Tech enthusiast into AI"

  • API: curl -X POST "http://127.0.0.1:8000/nl-query" -H "Content-Type: application/json" -d '{"query": "Show users over 30 similar to Tech enthusiast into AI"}'

  • Output: {"results": [{"id": 3, "name": "Chandran", "description": "...", "distance": 0.98}, ...], "sql": "...", "params": [...]}

• Test DB: curl http://127.0.0.1:8000/db-test

  • Output: {"status": "Connected", "version": "..."} or {"status": "Failed", "error": "..."}

Testing

• Run tests: pytest backend/
• CI/CD: GitHub Actions runs tests and linting on push/PR.

API Specs for Frontend

• GET /: Returns {"message": "IQuerio MVP is alive."}
• GET /db-test: Returns {"status": "Connected", "version": "..."} or {"status": "Failed", "error": "..."}
• POST /optimize:
  • Input: {"query": "<SQL>"}
  • Output: {"query": "...", "optimized_query": "...", "issues": [...], "suggestions": [...], "explain_plan": [...]}
• POST /upload-embedding:
  • Input: {"user_id": int, "description": string}
  • Output: {"status": "...", "user_id": int}
• POST /search-similar:
  • Input: {"description": string, "limit": int}
  • Output: {"results": [{"id": int, "name": string, "description": string, "distance": float}, ...]}
• POST /nl-query:
  • Input: {"query": string}
  • Output: {"results": [{"id": int, "name": string, "description": string, "distance": float}, ...], "sql": string, "params": [...]}

Development

• Backend: FastAPI, psycopg2, sentence-transformers
• Database: Postgres with pgvector
• CLI: db-toolkit (installed via pip install .)
• Testing: pytest with GitHub Actions CI