Skip to main content

AutoGen Built-Simple Research Tools

PyPI version Python 3.9+ License: MIT

GPU-accelerated research tools for Microsoft AutoGen agents.

Search millions of academic papers on PubMed (4.48M+ articles) and ArXiv (2.77M+ papers) with semantic understanding. Built on Built-Simple's high-performance research APIs.

🚀 Features

  • 🔬 PubMed Search: Semantic + keyword hybrid search across 4.48M biomedical articles
  • 📚 ArXiv Search: Fast search across 2.77M preprints (physics, math, CS, AI/ML)
  • ⚡ GPU-Accelerated: Millisecond response times via Built-Simple APIs
  • 📄 Full Text: Retrieve complete article text from PubMed Central
  • 🤖 AutoGen Native: One-line registration with AutoGen agents

📦 Installation

pip install autogen-builtsimple

Or with AutoGen included:

pip install autogen-builtsimple[autogen]

🏃 Quick Start

import os
from autogen import ConversableAgent
from autogen_builtsimple import register_research_tools

# Create agents
assistant = ConversableAgent(
    name="ResearchAssistant",
    system_message="""You are a research assistant that helps find and analyze 
    scientific papers. Use the search tools to find relevant literature.""",
    llm_config={"config_list": [{"model": "gpt-4", "api_key": os.environ["OPENAI_API_KEY"]}]},
)

user_proxy = ConversableAgent(
    name="User",
    llm_config=False,
    is_termination_msg=lambda msg: "TERMINATE" in msg.get("content", ""),
    human_input_mode="NEVER",
)

# Register all research tools with one line!
register_research_tools(assistant, user_proxy)

# Start researching
chat_result = user_proxy.initiate_chat(
    assistant,
    message="Find 5 recent papers on CRISPR gene therapy for cancer treatment"
)

🔧 Available Tools

PubMed Tools

Tool Description
search_pubmed Hybrid semantic + keyword search across PubMed
get_pubmed_full_text Retrieve full article text (PMC articles)
get_pubmed_metadata Batch fetch metadata for multiple PMIDs

ArXiv Tools

Tool Description
search_arxiv Hybrid/vector/text search across ArXiv
get_arxiv_paper_url Get PDF and abstract URLs for a paper

📖 Usage Examples

Search PubMed

from autogen_builtsimple import search_pubmed

# Semantic search for medical literature
results = search_pubmed(
    query="machine learning early cancer detection biomarkers",
    limit=10,
    min_year=2020
)
print(results)

Search ArXiv

from autogen_builtsimple import search_arxiv

# Search for AI papers
results = search_arxiv(
    query="large language models reasoning capabilities",
    limit=10,
    search_type="hybrid"  # or "vector" or "text"
)
print(results)

Get Full Text

from autogen_builtsimple import get_pubmed_full_text

# Get complete article text
full_text = get_pubmed_full_text(pmid="35421456")
print(full_text)

Selective Registration

from autogen_builtsimple import register_pubmed_tools, register_arxiv_tools

# Register only PubMed tools
register_pubmed_tools(assistant, user_proxy)

# Or only ArXiv tools
register_arxiv_tools(assistant, user_proxy)

🔑 API Keys (Optional)

The tools work without API keys using IP-based rate limiting. For higher limits, get a free API key:

results = search_pubmed(
    query="diabetes treatment",
    limit=20,
    api_key="your-api-key"
)

🏗️ Architecture

┌─────────────────┐     ┌──────────────────┐     ┌─────────────────┐
│  AutoGen Agent  │────▶│  autogen-builtsimple │────▶│  Built-Simple   │
│   (Assistant)   │     │      Tools       │     │     APIs        │
└─────────────────┘     └──────────────────┘     └─────────────────┘
                                                        │
                                ┌───────────────────────┴───────────────────────┐
                                │                                               │
                        ┌───────▼───────┐                               ┌───────▼───────┐
                        │    PubMed     │                               │    ArXiv      │
                        │  GPU Search   │                               │  GPU Search   │
                        │  4.48M docs   │                               │  2.77M docs   │
                        └───────────────┘                               └───────────────┘

📊 Performance

Endpoint Latency Index Size
PubMed Hybrid Search ~100-150ms 4.48M articles
ArXiv Hybrid Search ~70-90ms 2.77M papers
PubMed Full Text ~50ms PMC subset

🤝 Contributing

Contributions welcome! Please feel free to submit issues and pull requests.

📄 License

MIT License - see LICENSE for details.

🔗 Links

Metadata

Release files for autogen-builtsimple 0.1.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for autogen-builtsimple 0.1.0
File Size Uploaded
autogen_builtsimple-0.1.0.tar.gz 10.2 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for autogen-builtsimple 0.1.0
File Interpreter ABI Platform
autogen_builtsimple-0.1.0-py3-none-any.whl Python 3 none any Details

Total release size: 20.6 kB

Release files / autogen_builtsimple-0.1.0.tar.gz

Download URL autogen_builtsimple-0.1.0.tar.gz
Size 10.2 kB
Tags Source
SHA-256 checksum
How to use checksums
5140042603298f845a37a947144b2f196be88aaf033ce6c606a90b605b5787d4
BLAKE2b-256 checksum
How to use checksums
204df6c14855f936ea3e4658f5b9b49d895b170d8765cf670828bb980acf260f
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.11.2

Release files / autogen_builtsimple-0.1.0-py3-none-any.whl

Download URL autogen_builtsimple-0.1.0-py3-none-any.whl
Size 10.4 kB
Tags Python 3
SHA-256 checksum
How to use checksums
9853975274e0b474a4a67d4361c0ad3ad8a0e6b110c5348f724cf6a3f7fed53a
BLAKE2b-256 checksum
How to use checksums
b60494ebdd624d1a9432ffbb86ec3b804e71da79804e6f6ae997d47b1b7a57a8
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.11.2

Release history Release notifications | RSS feed

This release

0.1.0 This release

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page