pyprik 0.2.1

Intelligent data matching library with LLM-powered natural language responses

pyprik

pyprik is an intelligent data matching library that finds the best matching items based on your requirements. The library includes Large Language Model (LLM) integration for natural language responses and conversational search capabilities.

Features

• Intelligent Data Matching: Advanced algorithms for finding relevant items in datasets
• LLM Integration: Natural language responses for search results
• Conversational Search: Search using natural language queries
• Smart Explanations: Detailed explanations of why certain results were matched
• REST API Server: FastAPI server with intelligent endpoints
• Multi-LLM Support: Compatible with OpenAI GPT and Google Gemini models
• Agent Architecture: Modular design with dedicated agent components

Installation

pip install pyprik

Quick Start

Basic Data Matching

from pyprik import find_top_matching
import pandas as pd

# Your data
data = {
    'Product': ['Laptop', 'Smartphone', 'Tablet'],
    'Brand': ['Dell', 'Apple', 'Samsung'],
    'RAM': ['8GB', '4GB', '6GB'],
    'Price': [600, 999, 300]
}
products = pd.DataFrame(data)

# Find matches
requirements = {'Brand': 'Apple', 'RAM': '4GB'}
results = find_top_matching(products, requirements, top_n=2)
print(results)

Agent-Based Server

from pyprik.agent import serveragent

# Start server
serveragent.run_server(host="127.0.0.1", port=8000)

# Or create server instance
server = serveragent.create_server_agent()

LLM-Enhanced Search

from pyprik import smart_product_search, setup_default_llm

# Setup LLM (requires API key)
setup_default_llm('openai', 'gpt-3.5-turbo')

# Get natural language response
response = smart_product_search(
    dataset=products,
    requirements={'Brand': 'Apple'},
    top_n=3,
    natural_response=True
)
print(response)

REST API Server

Start an intelligent search server:

from pyprik.agent import serveragent
serveragent.run_server(host="127.0.0.1", port=8000)

API Endpoints

• POST /search - Intelligent product search
• POST /conversational-search - Natural language search
• POST /explain - Explain search results
• POST /configure-llm - Configure LLM settings
• POST /upload-dataset - Upload your dataset
• GET /dataset-info - Get dataset information

LLM Configuration

1. Get an API Key:

   • OpenAI: https://platform.openai.com/api-keys
   • Google AI Studio: https://makersuite.google.com/app/apikey

2. Set Environment Variable:

   # For OpenAI
   export OPENAI_API_KEY="your-api-key-here"
   
   # For Gemini
   export GEMINI_API_KEY="your-api-key-here"
   

3. Install Optional Dependencies (if not auto-installed):

   pip install openai google-generativeai fastapi uvicorn
   

Core Functions

Data Matching Functions

• find_top_matching(dataset, requirements, top_n) - Find best matches
• find_matching(dataset, requirements) - Get all matches with scores

LLM Functions

• smart_product_search() - Enhanced search with LLM responses
• conversational_search() - Natural language search
• explain_search_results() - Detailed explanations
• setup_default_llm() - Configure LLM settings

Server Functions

• run_server() - Start FastAPI server
• create_server_agent() - Create server instance

Use Cases

• E-commerce: Find products matching specific criteria
• Real Estate: Search properties with natural language
• Job Matching: Match candidates to positions
• Product Catalogs: Intelligent product recommendations
• Data Analysis: Understand matching algorithms and results

License

MIT License - see LICENSE.txt for details.