Hello World.py

This Python guide will show you the basics of E2B:

  • Connect the code interpreter to an LLM
  • Prompt the LLM to generate the Python code
  • Execute the AI-generated Python code in a secure E2B sandbox

Get full code

Check out the full code in our cookbook.

Overview

  1. Install Code Interpreter SDK
  2. Prepare prompt and tools for LLM
  3. Prepare code interpreting
  4. Call LLM and parse response with tools
  5. Create code interpreter and run everything
  6. Save generated chart

1. Install Code Interpreter SDK

The Code Interpreter SDK allows you to run AI-generated code in a secure small VM - E2B sandbox - made for AI code execution. Inside the sandbox is a Jupyter server running that you can control from our SDK through the notebook.execCell() method.

Check out the SDK's repository on GitHub.

pip install e2b_code_interpreter python-dotenv

Get your E2B API key here and save it to .env in your root directory.

.env

E2B_API_KEY="e2b-api-key"

2. Prepare prompt and tools for LLM

We'll be using Anthropic's Claude 3 Opus model but E2B works with any LLM so feel free to pick any you want!

Usually, all you need from the model is just support for tool use. If the LLM doesn't support tool use, you can ask the LLM to respond with Markdown or XML and parse the LLM's output on your own. Then just pass the parsed code from code blocks to the code interpreter.

Create the model.py file and paste the following code.

model.py

MODEL_NAME = "claude-3-opus-20240229"

SYSTEM_PROMPT = """
## your job & context
you are a python data scientist. you are given tasks to complete and you run python code to solve them.
- the python code runs in jupyter notebook.
- every time you call `execute_python` tool, the python code is executed in a separate cell. it's okay to multiple calls to `execute_python`.
- display visualizations using matplotlib or any other visualization library directly in the notebook. don't worry about saving the visualizations to a file.
- you have access to the internet and can make api requests.
- you also have access to the filesystem and can read/write files.
- you can install any pip package (if it exists) if you need to but the usual packages for data analysis are already preinstalled.
- you can run any python code you want, everything is running in a secure sandbox environment.
"""

tools = [
    {
        "name": "execute_python",
        "description": "Execute python code in a Jupyter notebook cell and returns any result, stdout, stderr, display_data, and error.",
        "input_schema": {
            "type": "object",
            "properties": {
                "code": {
                    "type": "string",
                    "description": "The python code to execute in a single cell."
                }
            },
            "required": ["code"]
        }
    }
]

This defines our system prompt and the tools dictionary with available tools for the LLM - namely the "execute_python" tool. A little bit later, we'll connect "execute_python" to the E2B's code interpreter.

3. Prepare code interpreting

We'll create a new function called code_interpret() in a separate file code_interpreter.py.

code_interpreter.py

from e2b_code_interpreter import CodeInterpreter

def code_interpret(code_interpreter: CodeInterpreter, code: str):
  print(f"\n{'='*50}\n> Running following AI-generated code:\n{code}\n{'='*50}")
  exec = code_interpreter.notebook.exec_cell(
    code,
    # You can stream logs from the code interpreter
    # on_stderr=lambda stderr: print("\n[Code Interpreter stdout]", stderr),
    # on_stdout=lambda stdout: print("\n[Code Interpreter stderr]", stdout),
    #
    # You can also stream additional results like charts, images, etc.
    # on_result=...
  )

  if exec.error:
    print("[Code Interpreter error]", exec.error) # Runtime error
  else:
    return exec.results, exec.logs

This function takes the CodeInterpreter object from our SDK and code as paramaters. The code parameter is the code generated by the LLM.

Inside the function, we call the code_interpreter.notebook.exec_cell() method. The exec_cell() takes the code argument and executes this code inside E2B sandbox.

4. Call LLM and parse response with tools

We're using Claude 3 Opus. Get your Anthropic API key, save it to .env file, and install the Anthropic SDK.

.env

ANTHROPIC_API_KEY="anthropic-api-key"

pip install anthropic

Now we'll put everything together. Create the main.py file, import dependencies, and create the chat() function that will do the LLM calling and tool parsing.

main.py

import base64
from dotenv import load_dotenv
from anthropic import Anthropic
from typing import List, Tuple
from e2b_code_interpreter import CodeInterpreter, Result
from e2b_code_interpreter.models import Logs

from e2b_hello_world.model import MODEL_NAME, SYSTEM_PROMPT, tools
from e2b_hello_world.code_interpreter import code_interpret

# Load the .env file
load_dotenv()

client = Anthropic()

def chat(code_interpreter: CodeInterpreter, user_message: str) -> Tuple[List[Result], Logs]:
    print(f"\n{'='*50}\nUser Message: {user_message}\n{'='*50}")

    message = client.beta.tools.messages.create(
        model=MODEL_NAME,
        system=SYSTEM_PROMPT,
        max_tokens=4096,
        messages=[{"role": "user", "content": user_message}],
        tools=tools,
    )

    print(f"\n{'='*50}\nModel response: {message.content}\n{'='*50}")

    if message.stop_reason == "tool_use":
        tool_use = next(block for block in message.content if block.type == "tool_use")
        tool_name = tool_use.name
        tool_input = tool_use.input

        print(f"\n{'='*50}\nUsing tool: {tool_name}\n{'='*50}")

        if tool_name == "execute_python":
            return code_interpret(code_interpreter, tool_input["code"])
        return []

5. Create code interpreter and run everything

Now we are ready to run our program. At the end of main.py add the following code prompting the LLM to visualize a distribution of the height of men and print the median.

main.py

def main():
  user_message = "Visualize a distribution of height of men based on the latest data you know. Also, print the median value."

  # Create the CodeInterpreter object and save it as code_interpreter
  with CodeInterpreter() as code_interpreter:
    code_interpreter_results, code_interpreter_logs = chat(
      code_interpreter,
      user_message,
    )

    print(code_interpreter_logs)

    first_result= code_interpreter_results[0]
    print(first_result)

After running your code with the following command

$ python main.py

you should see results similar to this:

stdout=['The median male height is 175.5 cm\n'] stderr=[]
<Figure size 800x400 with 1 Axes>

We got our median in the logs (stdout, and stderr) but we also something intering in first_result.

<Figure size 800x400 with 1 Axes>

6. Save generated chart

This looks like a plot. Let's save it to a file. Add the following to the end of main.py and run the code again with python main.py in your terminal.

main.py

    # If we received a chart in PNG form, we can visualize it
    if first_result.png:
      # Decode the base64 encoded PNG data
      png_data = base64.b64decode(first_result.png)

      # Generate a unique filename for the PNG
      filename = f"chart.png"

      # Save the decoded PNG data to a file
      with open(filename, "wb") as f:
          f.write(png_data)

      print(f"Saved chart to {filename}")

The chart was saved in the chart.png file and it should look similar to this:

Chart visualizing distribution height of men