import re
import csv
import json
import os
from typing import List, Dict, Any
import arxiv
import pandas as pd
from mcp.server.fastmcp import FastMCP
# Initialize FastMCP server
mcp = FastMCP("ArXiv Research Assistant")
@mcp.tool()
def search_arxiv(query: str, max_results: int = 5) -> str:
"""
Search ArXiv for papers.
Args:
query: The search query.
max_results: Maximum number of results to return (default 5).
Returns:
JSON string containing a list of papers with title, authors, published date, and abstract.
"""
client = arxiv.Client()
search = arxiv.Search(
query=query,
max_results=max_results,
sort_by=arxiv.SortCriterion.Relevance
)
results = []
for result in client.results(search):
paper_info = {
"title": result.title,
"authors": [author.name for author in result.authors],
"published": result.published.strftime("%Y-%m-%d"),
"pdf_url": result.pdf_url,
"abstract": result.summary
}
results.append(paper_info)
return json.dumps(results, indent=2)
@mcp.tool()
def extract_key_findings(abstract: str) -> str:
"""
Extract key findings (Problem, Method, Result) from an abstract using heuristic keyword matching.
Args:
abstract: The text of the paper abstract.
Returns:
JSON string with keys 'problem', 'method', 'result'.
"""
# 1. Clean and Split into sentences
text = abstract.replace("\n", " ")
# Simple regex split by . or ? or ! followed by space or end of string
sentences = re.split(r'(?<=[.?!])\s+', text)
findings = {
"problem": [],
"method": [],
"result": []
}
# 2. Keyword definitions
keywords = {
"problem": ["problem", "challenge", "issue", "limitat", "address", "motivation", "gap"],
"method": ["method", "propose", "approach", "framework", "architecture", "algorithm", "technique", "use", "using"],
"result": ["result", "show", "demonstrate", "find", "achieve", "perform", "improve", "accuracy", "state-of-the-art"]
}
# 3. Classify sentences
for sent in sentences:
sent_lower = sent.lower()
# Simple scoring: count keyword matches for each category
scores = {cat: 0 for cat in keywords}
for cat, kw_list in keywords.items():
for kw in kw_list:
if kw in sent_lower:
scores[cat] += 1
# Assign sentence to category with highest score (if score > 0)
best_cat = None
max_score = 0
for cat, score in scores.items():
if score > max_score:
max_score = score
best_cat = cat
if best_cat:
findings[best_cat].append(sent)
# Join lists into single strings for better readability
final_findings = {k: " ".join(v) if v else "Not explicitly found." for k, v in findings.items()}
return json.dumps(final_findings, indent=2)
@mcp.tool()
def save_to_bibliography(paper_metadata: str) -> str:
"""
Save a paper's metadata and key findings to a local references.csv file.
Args:
paper_metadata: JSON string or dict. Should contain 'title', 'authors', 'published', 'pdf_url'.
Can optionally contain 'problem', 'method', 'result' from extract_key_findings.
Returns:
Success message.
"""
if isinstance(paper_metadata, str):
try:
data = json.loads(paper_metadata)
except json.JSONDecodeError:
return "Error: paper_metadata must be a valid JSON string."
else:
data = paper_metadata
required_keys = ["title", "authors", "published", "pdf_url", "problem", "method", "result"]
# Handle missing keys gracefully
row = {k: str(data.get(k, "")) for k in required_keys}
# CSV file path
file_path = "references.csv"
file_exists = os.path.isfile(file_path)
try:
with open(file_path, mode='a', newline='', encoding='utf-8') as f:
writer = csv.DictWriter(f, fieldnames=required_keys)
if not file_exists:
writer.writeheader()
writer.writerow(row)
return f"Successfully added '{row['title']}' to {file_path}"
except Exception as e:
return f"Error saving to bibliography: {str(e)}"
if __name__ == "__main__":
# Initialize and run the server
mcp.run(transport='stdio')
