summaryxtract

A Python package for structured, reliable extraction of key insights from cultural texts using LLM-powered pattern matching.

Overview

summaryxtract processes user-provided text (e.g., articles, discussions) to extract and summarize core themes, historical context, and contributions related to cultural topics. It leverages LLM7 (by default) or custom LLMs (via LangChain) with retry mechanisms and pattern validation for consistent, accurate results.

Key Features

✅ Structured Summarization – Extracts key insights from textual descriptions. ✅ Pattern Matching & Validation – Ensures responses adhere to predefined formats. ✅ LLM Flexibility – Works with LLM7 (default) or any LangChain-compatible LLM (OpenAI, Anthropic, Google, etc.). ✅ Retry Mechanism – Handles API failures gracefully. ✅ No Media Processing – Focuses purely on text extraction.

Installation

pip install summaryxtract

Usage

Basic Usage (Default LLM7)

from summaryxtract import summaryxtract

response = summaryxtract(user_input="Your text about cultural history here...")
print(response)  # List of extracted insights

Custom LLM Integration

You can replace the default LLM7 with any LangChain-compatible LLM (e.g., OpenAI, Anthropic, Google).

Example: Using OpenAI

from langchain_openai import ChatOpenAI
from summaryxtract import summaryxtract

llm = ChatOpenAI()
response = summaryxtract(user_input="Your text...", llm=llm)

Example: Using Anthropic

from langchain_anthropic import ChatAnthropic
from summaryxtract import summaryxtract

llm = ChatAnthropic()
response = summaryxtract(user_input="Your text...", llm=llm)

Example: Using Google Vertex AI

from langchain_google_genai import ChatGoogleGenerativeAI
from summaryxtract import summaryxtract

llm = ChatGoogleGenerativeAI()
response = summaryxtract(user_input="Your text...", llm=llm)

Parameters

Default LLM (LLM7)

• Provider: LLM7 (free tier available).
• Rate Limits: Sufficient for most use cases (free tier).
• Custom API Key: Pass via api_key or LLM7_API_KEY env var.

Get a Free API Key: 🔗 Register at LLM7

Error Handling

• If the LLM call fails, a RuntimeError is raised with the error message.
• Retry mechanisms ensure robustness.

Contributing & Issues

🐛 Report bugs/feature requests: 🔗 GitHub Issues

Author

👤 Eugene Evstafev 📧 hi@euegne.plus 🔗 GitHub: chigwell