nbformat 3.0.0

Parse Jupyter Notebooks
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52

# Changelog

All notable changes to `nbformat` will be documented in this file.

## [Unreleased]

## [3.0.0] - 2026-04-26

## [2.2.0] - 2026-04-25

## [2.1.0] - 2026-04-25

### Changed

- `serialize_notebook` now sorts object keys alphabetically at every depth, matching Python `nbformat.write`'s `json.dumps(sort_keys=True)` output. This removes workspace-feature-flag sensitivity: before, output order depended on whether downstream crates enabled `serde_json/preserve_order`. Notebooks saved through this serializer will produce smaller and more stable git diffs.

### Added

- `tests/regenerate_expected.py` writes canonical Python `nbformat.write` output to `tests/notebooks/expected/` for a subset of fixtures. The new `test_matches_python_oracle_output` conformance test asserts byte-for-byte parity against those files.

## [2.0.0] - 2026-04-12

### Breaking

- Add `Notebook::V4QuirksMode` variant for v4.5 notebooks with missing cell IDs. (#295)
- Mark `Notebook` enum as `#[non_exhaustive]`.

### Added

- `V4Quirks` wrapper with `repair()` method to surface and fix spec violations via the type system.
- `Quirk` enum (also `#[non_exhaustive]`) with `MissingCellId` variant.

## [1.2.2] - 2026-03-14

## [1.2.1] - 2026-03-06

## [1.2.0] - 2026-02-23

### Added

- Support for v3 notebook parsing and upconversion to v4. (#275)

## [1.1.0] - 2026-02-20

### Fixed

- Accept both string and array for cell source.
- Trailing newline bug in MultilineString and media serialization. (#271)

## [1.0.0] - 2026-01-07

Stable release of the Jupyter notebook format parser. See git history for earlier changes.