# Usage Examples
This document provides examples of using the XRootD MCP Server with various MCP clients.
## Tool Reference
### list_directory
List contents of a directory on the XRootD server.
**Input:**
```json
{
"path": "/store/data/2024"
}
```
**Output:**
```json
[
{
"name": "file1.root",
"isDirectory": false,
"size": 1048576,
"modificationTime": "2024-01-15T10:30:00.000Z"
},
{
"name": "subdir",
"isDirectory": true,
"size": 4096,
"modificationTime": "2024-01-14T08:20:00.000Z"
}
]
```
### get_file_info
Get detailed metadata about a file or directory.
**Input:**
```json
{
"path": "/store/data/2024/file1.root"
}
```
**Output:**
```json
{
"path": "/store/data/2024/file1.root",
"size": 1048576,
"modificationTime": "2024-01-15T10:30:00.000Z",
"isDirectory": false
}
```
### read_file
Read contents of a file. Supports byte range reading.
**Full file read:**
```json
{
"path": "/store/data/config.txt"
}
```
**Partial read (bytes 0-1023):**
```json
{
"path": "/store/data/large_file.root",
"start": 0,
"end": 1023
}
```
### check_file_exists
Check if a file or directory exists.
**Input:**
```json
{
"path": "/store/data/2024/file1.root"
}
```
**Output:**
```json
{
"path": "/store/data/2024/file1.root",
"exists": true
}
```
### get_directory_size
Calculate total size of a directory recursively.
**Input:**
```json
{
"path": "/store/data/2024"
}
```
**Output:**
```json
{
"path": "/store/data/2024",
"size": 104857600,
"sizeHuman": "100.00 MB"
}
```
## Common Use Cases
### Browse Directory Tree
1. Start at root: `list_directory` with path `/`
2. Explore subdirectories: `list_directory` with each subdirectory path
3. Get file details: `get_file_info` for files of interest
### Check File Before Reading
1. Check existence: `check_file_exists`
2. Get metadata: `get_file_info` to check size
3. Read file: `read_file` to access content
### Analyze Storage Usage
1. List directory: `list_directory` to see contents
2. Calculate size: `get_directory_size` for each subdirectory
3. Compare sizes to identify large directories
### Read Large Files Efficiently
1. Get file info: `get_file_info` to check total size
2. Read in chunks: Multiple `read_file` calls with different byte ranges
3. Process incrementally without loading entire file
## Integration with Claude Desktop
Add to `~/Library/Application Support/Claude/claude_desktop_config.json`:
```json
{
"mcpServers": {
"xrootd": {
"command": "node",
"args": [
"/absolute/path/to/xrootd-mcp-server/build/index.js"
],
"env": {
"XROOTD_SERVER": "root://dtn-eic.jlab.org"
}
}
}
}
```
Then restart Claude Desktop. You can now ask Claude to:
- "List files in the root directory on the XRootD server"
- "What files are available on the XRootD server?"
- "Check if /path/to/file.root exists on the XRootD server"
- "What's the size of a specific directory?"
## Example Conversation with Claude
**You:** "Can you check what's available on the XRootD server?"
**Claude:** *Uses list_directory tool to browse the directory and presents the results*
**You:** "What's the total size of a specific directory?"
**Claude:** *Uses get_directory_size to calculate and report the size*
**You:** "Can you read a specific file if it exists?"
**Claude:** *Uses read_file to access and display the content*
## Tips
- Always use absolute paths starting with `/`
- For large files, use byte range reading to avoid memory issues
- Check file existence before attempting to read
- Use `get_directory_size` carefully on large directory trees (it can take time)
