researchbriefagent 0.1.1

A Python package for creating a research brief agent

Research Brief Agent

A Python library for an AI agent that generates concise research briefs from web searches using LangChain, Anthropic's Claude, and Tavily search.

Overview

The ResearchBriefAgent automates the process of researching a topic and producing a structured brief, including key points and sources. Built with LangChain's ecosystem, it leverages Anthropic's Claude model for reasoning and Tavily for web search, with SQLite-backed memory for context persistence.

Features

• Topic-Based Research: Generate briefs on any topic with a single function call.
• Structured Output: Receive formatted briefs with key points and cited sources.
• Memory Persistence: Retain conversation context across queries using SQLite checkpointing.
• Extensible: Built on LangChain, allowing integration with additional tools and models.

Installation

Install the library via PyPI:

pip install research_brief_agent

Usage

from research_brief_agent import ResearchBriefAgent
agent = ResearchBriefAgent(anthropic_api_key="your_anthropic_api_key", tavily_api_key="your_tavily_api_key")
brief = agent.generate_brief("Impact of renewable energy on climate change")
print(brief)

Example Output:

Research Brief: Impact of renewable energy on climate change

Key Points

• Reduces greenhouse gas emissions by replacing fossil fuels.
• Solar and wind power adoption has surged globally.
• Critical for limiting warming to 1.5°C.

Sources

• https://www.iea.org/reports/renewables-2023
• https://www.un.org/en/climatechange/renewable-energy

Requirements

Python: 3.8 or higher Dependencies: langchain-community, langgraph, langchain-anthropic, tavily-python, langgraph-checkpoint-sqlite API Keys: Obtain from Anthropic and Tavily Setup Clone the repo:

git clone https://github.com/yourusername/research_brief_agent.git
cd research_brief_agent

Install dependencies:

pip install -r requirements.txt

Or install from PyPI:

pip install research_brief_agent

Contributing

Contributions are welcome! Fork the repo, create a feature branch, commit changes, and open a pull request. See CONTRIBUTING.md for details (to be added).

License

MIT License - see LICENSE for details.

Contact

Email: [e-mail]

Issues: GitHub

Acknowledgments

Built with LangChain and Anthropic. Search powered by Tavily.

---

Verification

This is the exact content from my previous response, presented as a single Markdown file. You can save it directly as README.md in your project directory. The "Usage" section remains a single, copy-paste-friendly block, and all badges, sections, and details are preserved.

Next Steps

• Replace placeholders (yourusername, [e-mail]) with your actual details.
• Save this text as README.md in your project root.
• Push to GitHub and publish to PyPI to activate the badges.

Let me know if you need help with any part of the implementation!