plainb

Convert plain-text notebook files to Jupyter nbformat 4 JSON — in the browser or in Node.

Supported formats

| Format | Description | | ------------------ | -------------------------------------------------------------------------------------------------------- | | .py percent | VS Code / Spyder / Jupytext percent format — cells delimited by # %% | | .md classic | Markdown file where fenced code blocks with a language tag become code cells | | .md MyST | MyST Notebook format — {code-cell} directives and +++ cell breaks | | Sphinx Gallery | Sphinx-Gallery .py scripts — RST docstring + # %% sections |

Install

npm install plainb

Usage

import { parse, parsePy, parseMd, parseClassicMd, parseMystMd, parseSphinxGallery } from "plainb";

// Auto-dispatch by format
const notebook = parse(source, "py"); // or 'md' | 'sphinx-gallery'

// Or call parsers directly
const nb1 = parsePy(source); // .py percent format
const nb2 = parseMd(source); // .md classic or MyST (auto-detected)
const nb3 = parseClassicMd(source); // .md classic (explicit)
const nb4 = parseMystMd(source); // .md MyST (explicit)
const nb5 = parseSphinxGallery(source); // Sphinx Gallery

The returned object is a valid nbformat.INotebookContent (nbformat 4.5) — ready to save as .ipynb or pass to any Jupyter rendering library.

.py percent format

# %% [markdown]
# # Introduction
# Some **markdown** text.

# %%
import numpy as np
x = np.linspace(0, 1, 100)

# %% My cell name tags='["hide-input"]'
print(x)

• Cell type: [markdown], [md], [raw], or omitted (defaults to code)
• Cell name: free text after the type tag → stored in metadata.name
• Tags: tags='["tag1","tag2"]' → stored in metadata.tags
• Markdown content: line-comment prefix (# ) or triple-quoted string (""")

.md classic

# My Notebook

Some prose text becomes a **markdown cell**.

```python
x = 1 + 1
```

Only fenced blocks with a language identifier become code cells.
Untagged fences stay as markdown.

.md MyST

---
kernelspec: python3
---

# Title

+++ {"tags": ["hide-input"]}

Markdown cell content.

```{code-cell} ipython3
:tags: ["parameters"]
x = 42
```

Supports {code-cell}, {raw-cell}, {markdown-cell} directives and +++ cell breaks.

Sphinx Gallery

"""
================
My Gallery Script
================

Module docstring becomes the first markdown cell.
RST headings are converted to Markdown.
"""

# Authors: ...  ← stripped
# SPDX-...      ← stripped

# %%
# Section heading
# ---------------
# Comment block becomes a markdown cell.

import numpy as np   # code follows

API

parse(text: string, format: 'py' | 'md' | 'sphinx-gallery'): Notebook

parsePy(text: string): Notebook
parseMd(text: string): Notebook          // auto-detects classic vs MyST
parseClassicMd(text: string): Notebook  // classic Markdown (explicit)
parseMystMd(text: string): Notebook     // MyST Notebook (explicit)
parseSphinxGallery(text: string): Notebook

// Types (re-exported from the package)
interface Notebook   { nbformat: 4; nbformat_minor: 5; metadata: …; cells: Cell[] }
type Cell            = CodeCell | MarkdownCell | RawCell

Demo

git clone https://github.com/trungleduc/plainb
cd plainb
npm install
npm run demo

License

MIT