https://github.com/michaelliao/remote-code-runner
A remote code running service.
https://github.com/michaelliao/remote-code-runner
docker remote-code-execution
Last synced: about 2 months ago
JSON representation
A remote code running service.
- Host: GitHub
- URL: https://github.com/michaelliao/remote-code-runner
- Owner: michaelliao
- License: apache-2.0
- Archived: true
- Created: 2020-04-05T07:00:32.000Z (about 5 years ago)
- Default Branch: master
- Last Pushed: 2021-05-19T11:56:02.000Z (about 4 years ago)
- Last Synced: 2025-03-21T20:18:17.464Z (3 months ago)
- Topics: docker, remote-code-execution
- Language: Python
- Homepage:
- Size: 19.5 KB
- Stars: 134
- Watchers: 4
- Forks: 29
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Remote Code Runner
Remote Code Runner is a simple code service for running code on remote server side. Docker is used to execute untrusted code in a sandbox environment.
# Install
Environment:
- Ubuntu Linux 18.04
- Docker 19.x
- Python 3.8
MAKE SURE the login user has `sudo` privileges and does not require entering a password.
Install all required software:
```
$ sudo apt install git docker.io python3.8
```
Get source:
```
$ cd /srv
$ sudo git clone https://github.com/michaelliao/remote-code-runner.git
```
Generate all from source:
```
$ cd /srv/remote-code-runner
$ sudo python3.8 generate.py
```
Download required docker images by warm up script (this may take a long time):
```
$ cd /srv/remote-code-runner/bin
$ sudo sh warm-up-docker.sh
```
Start server:
```
$ cd /srv/remote-code-runner/bin
$ sudo start-runner.sh
```
# Usage
Using simple HTTP JSON API:
```
$ curl http://server-ip:8080/run -H 'Content-Type: application/json' -d '{"language":"python","code":"import math\nprint(math.pi)"}'
{"error": false, "timeout": false, "truncated": false, "output": "3.141592653589793\n"}
```
API input:
- language: language name, lowercase: `java`, `python`, `ruby`.
- code: language code as JSON string: `import math\nprint(math.pi)`
API output:
- timeout: boolean, is execution timeout.
- error: boolean, is error output. e.g. compile failed or syntax error.
- truncated: boolean, is output was truncated for too many characters.
- output: string, the output of execution.
# Execution
How code are executed on the remote server side:
1. Http server `runner.py` got language name and code from API;
2. Write code into a temperary directory like `/tmp/remote-code-runner/1`;
3. Execute command like `sudo docker run -t --rm -w /app -v /tmp/dir:/app python3 main.py`;
4. Write output into API response;
5. Clean up temperary directory.
# Limitation
- Multiple files are not supported.
- There is no way to read input from console. Any user input code will cause timeout.
# Security
Remote code runner should only be served in private network. User identity check, rate limit must be done in application server or reverse proxy like Nginx.
# Extension
How to add a new language:
1. Add configuration in `config.json`:
```
{
...
"languages": {
...
"node": {
"file": "main.js",
"image": "node:14.17-slim",
"command": "node main.js"
}
}
}
```
The key `node` is the language name.
2. Make sure image is downloaded on local:
```
$ sudo docker run -t --rm node:13.12-slim ls
```
3. Restart `start-runner.sh`.