Codeinterpreter-Codebox

Welcome to Codeinterpreter-Codebox, a cutting-edge cloud service tailored for executing Code Interpreter code.

📜 Table of Contents

🎯 Key Features
💡 What Can You Do?
📋 Related Projects & Their Differences
💻 Sample Codes
📦 Deployment
📧 Contact

🎯 Key Features

🐍 Independent Jupyter Sandbox Environment
A completely isolated Jupyter sandbox environment that supports state-preserving code execution. Supports file uploads and downloads.
🐳 Dockerized One-click Deployment
🌐 Completely Free and Open Source
🛡️ Information Security
Supports completely private deployments without the need to upload files to external servers.
🚀 More Flexible
Supports the opening of more ports to achieve more custom functionality requirements, such as: connecting to databases, connecting to the internet, connecting to other servers.

💡 What Can You Do?

Execute Python code.
Supports reading of uploaded files (Upload and Analyze Files).

if name=='main':
session=CodeinterpreterSession()
try:
session.upload_files(['./../data/test_data.csv'])
session.chat('How many columns are in the uploaded file?')
finally:
session.close()

Supports downloading files from the sandbox.
Supports dynamic scaling for commercial deployments.
Supports custom feature modifications, open ports, network connections, etc., supports web crawlers.

📋 Related Projects & Their Differences

Code Interpreter api
open-interpreter
E2B

Feature/Project	Codeinterpreter-Codebox★	Code Interpreter api	open-interpreter	E2B
Private Commercial Deployment	✅	❌	❌	❌
Remote Invocation	✅	✅	❌	✅
Fully Open Source	✅	❌	✅	❌
Free of Charge	✅	❌	✅	❌
Customization	✅	❌	❌	❌
Information Security	✅	❌	✅	❌
Multiple Programming Languages	❌	❌	✅	✅
Local Execution	✅	✅	✅	❌
Direct Invocation w/o Deployment	❌	✅	✅	✅

★ Indicates our project

💻 Sample Codes

Server-Client Interaction: jupyter server communicate showcases the details of interaction between the server and the Jupyter container.
Remote Execution: jupyter call demonstrates how to invoke a web service via an HTTP request, execute code, and retrieve results.

test_code="""
import docker
print(docker.version)

"""
execute(test_code)

#output
Execute Result= {"output_type":"text/plain","content":"6.1.3\n","files":null}

Client Session Demonstration: client session invocation shows how to allow LLM to invoke and execute code in a project through session and HTTP request.

session=CodeinterpreterSession()
try:
session.upload_files(['./../data/test_data.csv'])
session.chat('How many columns are in the uploaded file?')
finally:
session.close()

📦 Deployment

Install Docker
Linux: Install Docker by terminal
Mac os: Install Docker desktop for mac
Windows: Install Docker desktop for windows
Modify Docker config file
Head over to docker compose and substitute 'CODEBOX_ROOT_PATH' and 'YOUR_MNT_PATH:/codebox' with your file path to mount 'YOUR_MNT_PATH'.
Launch server

Build custom jupyter image

cd docker
docker build -t scipy-notebook:custom -f Dockerfile .

Build web server image

cd app
docker-compose -f docker_dev.yml build

Launch server

docker-compose -f docker_dev.yml up

📧 Contact

WeChat: zjajzzj1996
Email: zhangzhehian@gmail.com

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!