scrapbook

What it is and what it does

Scrapbook is a library for recording and retrieving data from Jupyter and nteract notebooks. It lets you save data values, dataframes, charts, and images as named scraps during notebook execution, then read them back later or pass them to other notebooks in a pipeline. The library works by encoding scraps as special cell outputs that persist in the notebook file, using encoders to handle different data types (JSON, Arrow, display-only formats). It integrates with papermill for notebook parameterization workflows and provides methods to read single notebooks or collections of notebooks and query their scraps.

The package depends on pandas, papermill, jsonschema, ipython, and pyarrow for its core functionality. It's designed for data scientists and analysts who want to build reproducible notebook-based workflows where intermediate results need to be captured and reused. The API includes functions like `glue()` to save scraps, `read_notebook()` to load them, and `scraps` to access them as a dictionary.

Use it for:

• Save model metrics or validation results from one notebook and load them in a reporting notebook for comparison
• Capture intermediate dataframes or processed data from a notebook to pass as input to another notebook in a pipeline
• Extract and summarize outputs from a batch of executed notebooks to generate a single report or dashboard
• Persist visualization objects or images generated during exploration for later display or documentation

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Scrapbook records data values and visual content from Jupyter notebook execution as named scraps, then retrieves and summarizes them later for use in downstream workflows or analysis.

No. The package is abandoned since 2022-04-13 with no maintenance activity for over two years. While it has low install friction and a permissive license, the lack of updates means compatibility issues with modern versions of its dependencies (pandas, papermill, ipython, pyarrow) are likely unaddressed. For new projects, consider actively maintained alternatives for notebook data persistence and workflow orchestration.

Install

scrapbook on PyPI

pip install scrapbook

uv add scrapbook

poetry add scrapbook

Installing scrapbook

Before you install

Installation is straightforward with low friction. However, the package is abandoned as of 2022-04-13 with no recent maintenance, so expect no bug fixes or updates for new Python or dependency versions.

License in practice

BSD license is permissive, allowing commercial and private use with minimal restrictions—suitable for most projects.

Quickstart

pip install scrapbook

import scrapbook as sb

# In a notebook cell:
sb.glue("result", value)
sb.glue("dataframe", df, 'arrow')

# Later, read the notebook:
nb = sb.read_notebook('notebook.ipynb')
print(nb.scraps)

Requires Python 3.5+; the package is abandoned and may have compatibility issues with current versions of pandas, papermill, ipython, and pyarrow.

Verify before relying

• Whether the package remains compatible with current versions of pandas, papermill, and ipython given its abandoned status since 2022
• Whether scraps persist correctly across different notebook execution environments or kernels
• Performance characteristics when working with large dataframes or many scraps in a single notebook

Package facts

Evidence: scrapbook-0.5.0-py3-none-any.whl

Keywords: jupyter, mapreduce, nteract, pipeline, notebook

Tags

More Software Development packages