read_paper.py•2.26 kB
"""Read functionality for the arXiv MCP server."""
import json
from pathlib import Path
from typing import Dict, Any, List
import mcp.types as types
from ..config import Settings
settings = Settings()
read_tool = types.Tool(
name="read_paper",
description="Read the full content of a stored paper in markdown format",
inputSchema={
"type": "object",
"properties": {
"paper_id": {
"type": "string",
"description": "The arXiv ID of the paper to read",
}
},
"required": ["paper_id"],
},
)
def list_papers() -> list[str]:
"""List all stored paper IDs."""
return [p.stem for p in Path(settings.STORAGE_PATH).glob("*.md")]
async def handle_read_paper(arguments: Dict[str, Any]) -> List[types.TextContent]:
"""Handle requests to read a paper's content."""
try:
paper_ids = list_papers()
paper_id = arguments["paper_id"]
# Check if paper exists
if paper_id not in paper_ids:
return [
types.TextContent(
type="text",
text=json.dumps(
{
"status": "error",
"message": f"Paper {paper_id} not found in storage. You may need to download it first using download_paper.",
}
),
)
]
# Get paper content
content = Path(settings.STORAGE_PATH, f"{paper_id}.md").read_text(
encoding="utf-8"
)
return [
types.TextContent(
type="text",
text=json.dumps(
{
"status": "success",
"paper_id": paper_id,
"content": content,
}
),
)
]
except Exception as e:
return [
types.TextContent(
type="text",
text=json.dumps(
{
"status": "error",
"message": f"Error reading paper: {str(e)}",
}
),
)
]