QuerySUTRA 0.4.1

SUTRA: AI-powered data analysis with automatic MySQL export

QuerySUTRA

SUTRA: Structured-Unstructured-Text-Retrieval-Architecture

AI-powered data analysis library. Upload PDFs, query with natural language, export to MySQL automatically.

Installation

pip install QuerySUTRA
pip install QuerySUTRA[mysql]  # For MySQL export

Quick Start

from sutra import SUTRA

# Upload PDF and auto-export to MySQL in ONE step
sutra = SUTRA(api_key="your-openai-key")

sutra.upload("data.pdf", auto_export_mysql={
    'host': 'localhost',
    'user': 'root', 
    'password': '123456',
    'database': 'my_database'  # Auto-creates if not exists
})

# Query immediately
result = sutra.ask("Show me all people")
print(result.data)

Features

1. Automatic MySQL Export

Database auto-created if not exists. No errors.

# Upload and export to MySQL automatically
sutra.upload("data.pdf", auto_export_mysql={
    'host': 'localhost',
    'user': 'root',
    'password': 'your_password',
    'database': 'my_new_database'  # Creates automatically
})

2. Complete Data Extraction

Processes entire PDF in chunks. Extracts ALL employees (not just first 10).

sutra.upload("large_document.pdf")  # Extracts all 50+ employees
sutra.tables()  # Shows all extracted tables

3. Natural Language Queries

result = sutra.ask("Show all people from California")
result = sutra.ask("Who has Python skills?", table="skills")
result = sutra.ask("Count employees by state", viz="pie")

4. Custom Visualizations

result = sutra.ask("Sales by region", viz="pie")
result = sutra.ask("Trends", viz="line")
result = sutra.ask("Compare", viz="bar")
result = sutra.ask("Data", viz="scatter")

5. Load Existing Databases

# Load SQLite
sutra = SUTRA.load_from_db("data.db", api_key="key")

# Connect to MySQL
sutra = SUTRA.connect_mysql("localhost", "root", "pass", "database")

# Connect to PostgreSQL  
sutra = SUTRA.connect_postgres("localhost", "postgres", "pass", "database")

6. Smart Features (Optional)

sutra = SUTRA(
    api_key="your-key",
    use_embeddings=True,    # Cache similar queries (saves API calls)
    fuzzy_match=True,       # "New York City" matches "New York"
    check_relevance=True,   # Detect irrelevant queries
    cache_queries=True      # Cache exact queries
)

7. Direct SQL (Free)

result = sutra.sql("SELECT * FROM people WHERE state='CA'")
print(result.data)

Complete Workflow

In Colab:

from sutra import SUTRA

sutra = SUTRA(api_key="your-key")
sutra.upload("employee_data.pdf")
sutra.tables()  # See extracted tables

# Export and download
sutra.export_db("data.db", format="sqlite")
from google.colab import files
files.download("data.db")

On Windows:

from sutra import SUTRA

# Load downloaded database
sutra = SUTRA.load_from_db("data.db", api_key="your-key")

# Export to MySQL (auto-creates database)
sutra.save_to_mysql("localhost", "root", "password", "my_database")

# Verify in MySQL
sutra_mysql = SUTRA.connect_mysql("localhost", "root", "password", "my_database")
sutra_mysql.tables()

Export Options

# SQLite
sutra.export_db("backup.db", format="sqlite")

# SQL dump
sutra.export_db("schema.sql", format="sql")

# JSON
sutra.export_db("data.json", format="json")

# Excel
sutra.export_db("data.xlsx", format="excel")

# MySQL (auto-creates database)
sutra.save_to_mysql("localhost", "root", "pass", "new_db")

# PostgreSQL
sutra.save_to_postgres("localhost", "postgres", "pass", "new_db")

API Reference

Initialize

SUTRA(api_key, db, use_embeddings, check_relevance, fuzzy_match, cache_queries)

Class Methods

• load_from_db(path, api_key) - Load SQLite
• connect_mysql(host, user, password, database) - Connect MySQL
• connect_postgres(host, user, password, database) - Connect PostgreSQL

Instance Methods

• upload(data, name, auto_export_mysql) - Upload with optional auto-export
• ask(question, viz, table) - Natural language query
• sql(query, viz) - Direct SQL
• tables() - List tables
• schema(table) - Show schema
• peek(table, n) - Preview data
• export_db(path, format) - Export database
• save_to_mysql(host, user, password, database) - Export to MySQL (auto-creates DB)
• save_to_postgres(...) - Export to PostgreSQL
• backup(path) - Backup
• close() - Close

Troubleshooting

MySQL database doesn't exist

• Fixed in v0.4.0 - auto-creates database automatically
• No need to manually create database

Only 10 employees extracted from 50-employee PDF

• Fixed in v0.4.0 - processes entire PDF in chunks
• Upgrade: pip install --upgrade QuerySUTRA

connect_mysql() not found

• Update: pip install --upgrade QuerySUTRA
• Install MySQL support: pip install QuerySUTRA[mysql]

Supported Formats

CSV, Excel, JSON, SQL, PDF, Word, Text, Pandas DataFrame

Requirements

• Python 3.8+
• OpenAI API key
• MySQL/PostgreSQL (optional)

License

MIT License

Changelog

v0.4.0

• AUTO-CREATES MySQL database (no more errors)
• Complete PDF extraction (all pages, all employees)
• Chunk processing for large documents
• One-line auto-export to MySQL
• Simplified everything

v0.3.x

• MySQL/PostgreSQL connectivity
• Embeddings caching
• Fuzzy matching
• Custom visualizations

---

Made by Aditya Batta