AIRGAP StudioAIRGAP Studio

Jupyter Notebook Support

Editing and running Jupyter notebooks in AIRGAP Assistant

Overview

AIRGAP Assistant fully supports Jupyter notebook (.ipynb) files. It recognizes cell-level context, enabling AI-powered code generation, analysis, and editing.

Key Features

AI-Powered Cell Editing

Open a notebook file and ask AIRGAP Assistant to create or modify specific cells.

"Write code to read a CSV file with pandas in the 3rd cell"
"Change the visualization code in the last cell from matplotlib to seaborn"

Cell-Level Context Awareness

The AI understands the entire notebook structure:

  • It recognizes variables and functions defined in previous cells.
  • It refers to markdown cell descriptions to understand code intent.
  • It can examine output results and diagnose errors.

Data Analysis Code Generation

You can request code for data analysis workflows in natural language.

"Check for missing values in this dataframe and show summary statistics"
"Visualize yearly sales trends as a line chart"
"Perform a correlation analysis and display it as a heatmap"

Prerequisites

Python Environment

A Python environment is required to run Jupyter notebooks.

# Check if Python is installed
python --version

# Install Jupyter packages (use offline packages in air-gapped environments)
pip install jupyter notebook

Kernel Setup

Select a Python kernel at the top of the notebook. If no kernel is available, run Jupyter: Select Interpreter from the Command Palette (Ctrl+Shift+P).

Use Cases

Exploratory Data Analysis (EDA)

  1. Add a CSV/Excel file to the workspace.
  2. Create a new notebook.
  3. Ask AIRGAP Assistant to "load this data and perform basic EDA."
  4. The AI automatically generates cells for data loading, summary statistics, and visualization.

Report Writing

You can create analysis reports by combining markdown cells and code cells. Ask the AI to "summarize the analysis results in markdown."

Limitations

  • Running notebooks (executing cells) requires a local Python kernel.
  • In air-gapped environments, offline packages must be prepared in advance for pip install.
  • Output from large datasets may not fully fit within the AI context window.