swarmauri_tool_jupyterexecutenotebook 0.11.0.dev1

Swarmauri notebook-execution tool for running Jupyter notebooks with nbclient and capturing cell outputs.

Swarmauri Jupyter Execute Notebook Tool

swarmauri_tool_jupyterexecutenotebook executes every cell in a Jupyter notebook with nbclient, captures notebook outputs in-place, and returns the executed NotebookNode for downstream processing.

Why

• Run notebooks inside Swarmauri workflows without shelling out to a separate notebook runner.
• Capture generated outputs for reporting, validation, or export steps.
• Keep notebook execution inside a simple Python tool contract.

Features

• Loads notebooks from disk with nbformat.
• Executes all cells with nbclient.NotebookClient.
• Returns the executed notebook object rather than a file path.
• Supports per-cell execution timeout control.
• Uses allow_errors=True, so the notebook can retain partial outputs even when a cell fails.

FAQ

What does this tool return?

It returns the NotebookNode after execution. If execution fails partway through, the tool still returns the notebook object so callers can inspect partial output.

Does it stop on the first cell error?

No. The underlying client is created with allow_errors=True, which preserves notebook state and outputs even when cells fail.

When should I use this instead of the parameterized execution tool?

Use this package for direct notebook execution. Use swarmauri_tool_jupyterexecutenotebookwithparameters when you need papermill-style parameter injection and an output notebook path.

Installation

uv add swarmauri_tool_jupyterexecutenotebook

pip install swarmauri_tool_jupyterexecutenotebook

Usage

from swarmauri_tool_jupyterexecutenotebook import JupyterExecuteNotebookTool

tool = JupyterExecuteNotebookTool()
executed = tool("notebooks/report.ipynb", timeout=60)

for cell in executed.cells:
    if cell.get("outputs"):
        print(cell["outputs"])

Examples

Execute a notebook before export

from swarmauri_tool_jupyterexecutenotebook import JupyterExecuteNotebookTool

tool = JupyterExecuteNotebookTool()
notebook = tool("analytics/daily_metrics.ipynb", timeout=120)

print(len(notebook.cells))

Inspect partial output after a failing cell

executed = JupyterExecuteNotebookTool()("debug/failing_notebook.ipynb")

for index, cell in enumerate(executed.cells):
    print(index, cell.get("outputs", []))

Related Packages

• swarmauri_tool_jupyterreadnotebook
• swarmauri_tool_jupyterwritenotebook
• swarmauri_tool_jupyterexecutenotebookwithparameters
• swarmauri_tool_jupyterexecuteandconvert
• swarmauri_tool_jupyterclearoutput

Foundational Swarmauri Packages

• swarmauri
• swarmauri_core
• swarmauri_base
• swarmauri_standard

More Documentation

• Swarmauri SDK repository
• nbclient documentation
• nbformat documentation