> ## Documentation Index > Fetch the complete documentation index at: https://nusaai-edit.mintlify.site/llms.txt > Use this file to discover all available pages before exploring further. # Code Interpreter > > Learn how using Neosantara And E2B Code Interpreter [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1r5I11nX5_W6xBKUX-0uPF4S6yjRImVu8?usp=sharing) ## Overview This guide explains how to use the **Neosantara** and **E2B Code Interpreter SDK** to create a secure cloud sandbox powered by [Firecracker](https://github.com/firecracker-microvm/firecracker). The sandbox includes a running Jupyter server, enabling large language models (LLMs) to execute Python code for tasks such as data analysis and visualization. *** ## Prerequisites To get started, ensure you have the following: Sign up at [Neosantara](https://app.neosantara.xyz/api-keys) to receive **10,000 free Token Credits** monthly. No credit card required. Obtain an API key from [E2B](https://e2b.dev/docs) to use the Code Interpreter. * **Node.js**: Version 16 or higher. * **Python**: Version 3.6 or higher. * **Required Python Packages**: * [`openai`](https://pypi.org/project/openai/): For interacting with Neosantara's API. * [`e2b_code_interpreter`](https://pypi.org/project/e2b-code-interpreter/): For running code in the E2B sandbox. * [`python-dotenv`](https://pypi.org/project/python-dotenv/): For managing environment variables. *** ## Installation Install the required Python packages using pip: ```bash icon="terminal" theme={null} pip install e2b_code_interpreter==1.0.0 python-dotenv openai -q ``` For additional details on available methods, refer to the [E2B documentation](https://e2b.dev/docs/quickstart). *** ## Setup Instructions Create a new file, e.g., `index.py`, to hold your code. Ensure the file has a `.py` extension. The script name can be customized (e.g., `index.py` or `my_script.py`), but it must be a valid `.py` file. Create a `.env` file in your project directory to store your API keys: ```env theme={null} NAI_API_KEY=your_neosantara_api_key E2B_API_KEY=your_e2b_api_key ``` Replace `your_neosantara_api_key` and `your_e2b_api_key` with the keys obtained from [Neosantara](https://app.neosantara.xyz/api-keys) and [E2B](https://e2b.dev/docs), respectively. Choose a code generation model from Neosantara. Available options include: * `nusantara-base` (default) * `archipelago-70b` * `Llama-3.3-Nemotron-Super-49B` See the full list of models at [Neosantara Models](https://www.neosantara.xyz/models). The code interpreter uses a dataset located at `/home/user/data.csv` in the sandbox. The CSV file uses a comma (`,`) as the delimiter and contains the following columns: | Column Name | Example Value | Description | | --------------------------------------------------- | --------------------- | ------------------------------------------------------- | | `country` | Argentina, Australia | Country name | | `Region` | SouthAmerica, Oceania | Geographic region | | `Surface area (km2)` | 2780400 | Land area in square kilometers | | `Population in thousands (2017)` | 44271 | Population in thousands | | `Population density (per km2, 2017)` | 16.2 | People per square kilometer | | `Sex ratio (m per 100 f, 2017)` | 95.9 | Male-to-female ratio | | `GDP: Gross domestic product (million current US$)` | 632343 | GDP in million USD | | `GDP per capita (current US$)` | 14564.5 | GDP per person | | `Life expectancy at birth, total (years)` | 76.4 | Average life expectancy | | ... (and more) | | See the full list in the [example code](#example-code). | *** ## Example Code Below is a complete example demonstrating how to set up the Code Interpreter, upload a dataset, and create a visualization (e.g., a linear regression chart of GDP per capita vs. life expectancy). ````python icon="python" theme={null} import os from dotenv import load_dotenv from openai import OpenAI from e2b_code_interpreter import Sandbox import re import json # Load environment variables load_dotenv() # API keys NAI_API_KEY = os.getenv("NAI_API_KEY") E2B_API_KEY = os.getenv("E2B_API_KEY") # Model selection MODEL_NAME = 'nusantara-base' # Alternatives: 'archipelago-70b', 'Llama-3.3-Nemotron-Super-49B' # System prompt for the LLM SYSTEM_PROMPT = """You're a Python data scientist. You are given tasks to complete and you run Python code to solve them. Information about the csv dataset: - It's in the `/home/user/data.csv` file - The CSV file is using , as the delimiter - It has the following columns (examples included): - country: "Argentina", "Australia" - Region: "SouthAmerica", "Oceania" - Surface area (km2): for example, 2780400 - Population in thousands (2017): for example, 44271 - Population density (per km2, 2017): for example, 16.2 - Sex ratio (m per 100 f, 2017): for example, 95.9 - GDP: Gross domestic product (million current US$): for example, 632343 - GDP growth rate (annual %, const. 2005 prices): for example, 2.4 - GDP per capita (current US$): for example, 14564.5 - Economy: Agriculture (% of GVA): for example, 10.0 - Economy: Industry (% of GVA): for example, 28.1 - Economy: Services and other activity (% of GVA): for example, 61.9 - Employment: Agriculture (% of employed): for example, 4.8 - Employment: Industry (% of employed): for example, 20.6 - Employment: Services (% of employed): for example, 74.7 - Unemployment (% of labour force): for example, 8.5 - Employment: Female (% of employed): for example, 43.7 - Employment: Male (% of employed): for example, 56.3 - Labour force participation (female %): for example, 48.5 - Labour force participation (male %): for example, 71.1 - International trade: Imports (million US$): for example, 59253 - International trade: Exports (million US$): for example, 57802 - International trade: Balance (million US$): for example, -1451 - Education: Government expenditure (% of GDP): for example, 5.3 - Health: Total expenditure (% of GDP): for example, 8.1 - Health: Government expenditure (% of total health expenditure): for example, 69.2 - Health: Private expenditure (% of total health expenditure): for example, 30.8 - Health: Out-of-pocket expenditure (% of total health expenditure): for example, 20.2 - Health: External health expenditure (% of total health expenditure): for example, 0.2 - Education: Primary gross enrollment ratio (f/m per 100 pop): for example, 111.5/107.6 - Education: Secondary gross enrollment ratio (f/m per 100 pop): for example, 104.7/98.9 - Education: Tertiary gross enrollment ratio (f/m per 100 pop): for example, 90.5/72.3 - Education: Mean years of schooling (female): for example, 10.4 - Education: Mean years of schooling (male): for example, 9.7 - Urban population (% of total population): for example, 91.7 - Population growth rate (annual %): for example, 0.9 - Fertility rate (births per woman): for example, 2.3 - Infant mortality rate (per 1,000 live births): for example, 8.9 - Life expectancy at birth, female (years): for example, 79.7 - Life expectancy at birth, male (years): for example, 72.9 - Life expectancy at birth, total (years): for example, 76.4 - Military expenditure (% of GDP): for example, 0.9 - Population, female: for example, 22572521 - Population, male: for example, 21472290 - Tax revenue (% of GDP): for example, 11.0 - Taxes on income, profits and capital gains (% of revenue): for example, 12.9 - Urban population (% of total population): for example, 91.7 Generally, you follow these rules: - ALWAYS FORMAT YOUR RESPONSE IN MARKDOWN - ALWAYS RESPOND ONLY WITH CODE IN CODE BLOCK LIKE THIS: `\`\`python {code} `\`\` - The Python code runs in a Jupyter notebook. - Each Python code block is executed in a separate cell. - Display visualizations using matplotlib or other libraries directly in the notebook. - You have access to the internet and can make API requests. - You can read/write files in the sandbox filesystem. - Install any pip package using `!pip install {package}` if needed (common data analysis packages are preinstalled). - All code runs in a secure sandbox environment. """ # Function to execute code in the sandbox def code_interpret(e2b_code_interpreter, code): print("Running code interpreter...") exec = e2b_code_interpreter.run_code( code, on_stderr=lambda stderr: print("[Code Interpreter] Error:", stderr), on_stdout=lambda stdout: print("[Code Interpreter] Output:", stdout), ) if exec.error: print("[Code Interpreter ERROR]", exec.error) return None return exec.results # Function to extract Python code from LLM response def match_code_blocks(llm_response): pattern = re.compile(r"```python match = pattern.search(llm_response) return match.group(1) if match else "" # Function to interact with the LLM and execute code def chat_with_llm(e2b_code_interpreter, user_message): print(f"\n{'='*50}\nUser message: {user_message}\n{'='*50}") messages = [ {"role": "system", "content": SYSTEM_PROMPT}, {"role": "user", "content": user_message}, ] client = OpenAI(api_key=NAI_API_KEY, base_url="https://api.neosantara.xyz/v1") response = client.chat.completions.create(model=MODEL_NAME, messages=messages) response_message = response.choices[0].message python_code = match_code_blocks(response_message.content) if python_code: return code_interpret(e2b_code_interpreter, python_code) else: print(f"No Python code found in response: {response_message}") return [] # Function to upload dataset to the sandbox def upload_dataset(code_interpreter): print("Uploading dataset to Code Interpreter sandbox...") dataset_path = "./data.csv" if not os.path.exists(dataset_path): raise FileNotFoundError("Dataset file not found at ./data.csv") try: with open(dataset_path, "rb") as f: file_buffer = f.read() remote_path = code_interpreter.files.write('data.csv', file_buffer) if not remote_path: raise ValueError("Failed to upload dataset") print("Dataset uploaded to:", remote_path) return remote_path except Exception as error: print("Error during file upload:", error) raise error # Main execution with Sandbox(api_key=E2B_API_KEY) as code_interpreter: # Upload dataset upload_dataset(code_interpreter) # Example task: Create a linear regression chart code_results = chat_with_llm( code_interpreter, "Create a chart showing the linear regression of GDP per capita vs. life expectancy from the dataset. Filter out missing or invalid values.", ) if code_results: first_result = code_results[0] print("Visualization generated successfully!") else: raise Exception("No results from code interpreter") # Display the visualization (e.g., PNG output) first_result ```` *** ## Running the Code 1. **Ensure the dataset** (`data.csv`) is in the same directory as your script. 2. **Run the script** using: ```bash theme={null} python index.py ``` 3. **View the output**: The script uploads the dataset, sends a task to the LLM, executes the generated Python code in the sandbox, and displays the result (e.g., a chart). ### Example Output The example task generates a scatter plot with a linear regression line showing the relationship between GDP per capita and life expectancy, filtered for valid data. Result Image *** ## Key Features * **Secure Sandbox**: Code runs in an isolated environment powered by Firecracker. * **Jupyter Integration**: Execute Python code in a Jupyter notebook within the sandbox. * **Data Visualization**: Use libraries like `matplotlib` or `seaborn` to create charts directly in the notebook. * **File Access**: Read/write files in the sandbox filesystem. * **API Support**: Make API requests from within the sandbox. * **Extensibility**: Install additional Python packages using `!pip install`. *** ## Troubleshooting * **API Key Issues**: Ensure your Neosantara and E2B API keys are valid and correctly set in the `.env` file. * **Dataset Not Found**: Verify that `data.csv` exists in the script's directory. * **Code Execution Errors**: Check the console for error messages from the sandbox (`[Code Interpreter ERROR]`). * **Missing Python Code**: If the LLM response lacks a Python code block, ensure your prompt is clear and specific. For further assistance, refer to the [E2B documentation](https://e2b.dev/docs) or [Neosantara support](https://www.neosantara.xyz/contact).