Mpc Sintific Search By Ai For Ai

🚀 Scientific Search MCP for Cursor

This is a Model Control Protocol (MCP) implemented for Cursor, offering the ability to conduct scientific searches through multiple academic search engines.

⚠️ Important Note

The vibration is implemented by AI coding.

✨ Features

• Integrated into Cursor using the MCP protocol.
• Cross - engine scientific search:
  • DuckDuckGo (General web search)
  • Google Scholar (Academic paper search)
  • PubMed (Biomedical literature search)
  • arXiv (Pre - print server for fields such as physics, mathematics, and computer science)
  • Semantic Scholar (AI - driven research paper search via the Tor proxy)

📦 Installation

Clone the repository

git clone https://github.com/yourusername/cursor-scientific-search.git
cd cursor-scientific-search

Install dependencies

pip install -r requirements.txt

For specific search engines, you may need to install additional dependencies:

# For arXiv search
pip install arxiv

# For Google Scholar
pip install scholarly

# For PubMed
pip install pymed

# For Semantic Scholar (Requires Tor)
pip install requests[socks]

Tor Setup for Semantic Scholar

Semantic Scholar's search uses Tor to bypass rate limits and IP blocks. To use this feature:

1. Install Tor:

   • Linux: sudo apt install tor (Debian/Ubuntu) or sudo pacman -S tor (Arch)
   • macOS: brew install tor
   • Windows: Download and install Tor Browser, which includes the Tor service.

2. Start the Tor service:

   • Linux: sudo systemctl start tor or sudo service tor start
   • macOS: brew services start tor
   • Windows: The Tor Browser includes the service, or you can use Tor Expert Bundle.

3. Verify that Tor is running on port 9050 (default):

   nc -z localhost 9050 && echo "Tor is running"
   

4. The scientific search MCP will automatically route Semantic Scholar requests through Tor.

💻 Usage Examples

Start the MCP server

python scientific_search_mcp.py

Connect to the MCP server in Cursor

Use the scientific search tool

Basic Usage

Search for scientific papers on different engines:

# Search for quantum computing papers on arXiv
results = await mcp_scientific_search_search_scientific(
    query="quantum computing", 
    engine="arxiv"
)

# Search for COVID - 19 research on PubMed
results = await mcp_scientific_search_search_scientific(
    query="COVID-19 treatment",
    engine="pubmed"
)

# Search for artificial intelligence papers on Google Scholar
results = await mcp_scientific_search_search_scientific(
    query="artificial intelligence",
    engine="gScholar"
)

# Search for natural language processing on Semantic Scholar
results = await mcp_scientific_search_search_scientific(
    query="natural language processing",
    engine="semScholar"
)

Advanced Usage

Extract metadata from search results:

from typing import Dict, Any
def print_result(result: Dict[str, Any]) -> None:
    """Print the metadata of search results."""
    print(f"Title: {result['title']}")
    print(f"Authors: {result['authors']}")
    print(f"Abstract: {result['abstract']}")
    print(f"Publication: {result.get('publication', 'N/A')}")
    print(f"DOI: {result.get('doi', 'N/A')}")
    print(f"URL: {result['url']}\n")

# Example usage
results = await mcp_scientific_search_search_scientific(
    query="machine learning",
    engine="arxiv"
)
for result in results:
    print_result(result)

⚠️ Important Notes

• Google Scholar Rate Limit: Avoid making a large number of requests in a short period to prevent triggering Google's rate limit.
• Tor Stability: Ensure that the Tor service is running; otherwise, Semantic Scholar requests may fail.

🛠️ Troubleshooting

Problem: Unable to connect to Semantic Scholar

Solution:

1. Check if Tor is installed and configured correctly.
2. Verify that Tor is listening on port 9050.
3. Ensure a stable network connection.

Problem: Empty search results

Possible reasons:

• Incorrect request parameters (e.g., invalid query string or incorrect engine name).
• Interrupted network connection.

📄 Example Code Output

Run the following code:

from mcpScientificSearch import MPSSearcher
import asyncio

async def main():
    async with MPSSearcher() as searcher:
        # Search for quantum computing papers on arXiv
        arxiv_results = await searcher.search(
            query="quantum computing",
            engine="arxiv"
        )
        print("arXiv Results:")
        for result in arxiv_results:
            print(f"Title: {result['title']}")
            print(f"Authors: {result['authors']}\n")

        # Search for artificial intelligence papers on Google Scholar
        gscholar_results = await searcher.search(
            query="artificial intelligence",
            engine="gScholar"
        )
        print("Google Scholar Results:")
        for result in gscholar_results:
            print(f"Title: {result['title']}")
            print(f"Authors: {result['authors']}\n")

if __name__ == "__main__":
    asyncio.run(main())

Sample Output:

arXiv Results:
Title: Quantum Computing for Beginners
Authors: John Doe, Jane Smith

Google Scholar Results:
Title: Introduction to Artificial Intelligence
Authors: Alice Johnson, Bob Brown
...