Introduction

Jupyter (né IPython) notebook files are simple JSON documents, containing text, source code, rich media output, and metadata. each segment of the document is stored in a cell.

Some general points about the notebook format:

Top-level structure

At the highest level, a Jupyter notebook is a dictionary with a few keys:

• metadata (dict)
• nbformat (int)
• nbformat_minor (int)
• cells (list)

{
  "metadata" : {
    "signature": "hex-digest", # used for authenticating unsafe outputs on load
    "kernel_info": {
        # if kernel_info is defined, its name field is required.
        "name" : "the name of the kernel"
    },
    "language_info": {
        # if language_info is defined, its name field is required.
        "name" : "the programming language of the kernel",
        "version": "the version of the language",
        "codemirror_mode": "The name of the codemirror mode to use [optional]"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0,
  "cells" : [
      # list of cell dictionaries, see below
  ],
}

Some fields, such as code input and text output, are characteristically multi-line strings. When these fields are written to disk, they may be written as a list of strings, which should be joined with '' when reading back into memory. In programmatic APIs for working with notebooks (Python, Javascript), these are always re-joined into the original multi-line string. If you intend to work with notebook files directly, you must allow multi-line string fields to be either a string or list of strings.

Cell Types

There are a few basic cell types for encapsulating code and text. All cells have the following basic structure:

{
  "cell_type" : "name",
  "metadata" : {},
  "source" : "single string or [list, of, strings]",
}

{
  "cell_type" : "markdown",
  "metadata" : {},
  "source" : ["some *markdown*"],
}

{
  "cell_type" : "code",
  "execution_count": 1, # integer or null
  "metadata" : {
      "collapsed" : True, # whether the output of the cell is collapsed
      "autoscroll": False, # any of true, false or "auto"
  },
  "source" : ["some code"],
  "outputs": [{
      # list of output dicts (described below)
      "output_type": "stream",
      ...
  }],
}

{
  "output_type" : "stream",
  "name" : "stdout", # or stderr
  "text" : ["multiline stream text"],
}

{
  "output_type" : "display_data",
  "data" : {
    "text/plain" : ["multiline text data"],
    "image/png": ["base64-encoded-png-data"],
    "application/json": {
      # JSON data is included as-is
      "json": "data",
    },
  },
  "metadata" : {
    "image/png": {
      "width": 640,
      "height": 480,
    },
  },
}

{
  "output_type" : "execute_result",
  "execution_count": 42,
  "data" : {
    "text/plain" : ["multiline text data"],
    "image/png": ["base64-encoded-png-data"],
    "application/json": {
      # JSON data is included as-is
      "json": "data",
    },
  },
  "metadata" : {
    "image/png": {
      "width": 640,
      "height": 480,
    },
  },
}

{
  'ename' : str,   # Exception name, as a string
  'evalue' : str,  # Exception value, as a string

  # The traceback will contain a list of frames,
  # represented each as a string.
  'traceback' : list,
}

{
  "cell_type" : "raw",
  "metadata" : {
    # the mime-type of the target nbconvert format.
    # nbconvert to formats other than this will exclude this cell.
    "format" : "mime/type"
  },
  "source" : ["some nbformat mime-type data"]
}

Backward-compatible changes

The notebook format is an evolving format. When backward-compatible changes are made, the notebook format minor version is incremented. When backward-incompatible changes are made, the major version is incremented.

As of nbformat 4.x, backward-compatible changes include:

• new fields in any dictionary (notebook, cell, output, metadata, etc.)
• new cell types
• new output types

New cell or output types will not be rendered in versions that do not recognize them, but they will be preserved.

Metadata

Metadata is a place that you can put arbitrary JSONable information about your notebook, cell, or output. Because it is a shared namespace, any custom metadata should use a sufficiently unique namespace, such as metadata.kaylees_md.foo = "bar".

Metadata fields officially defined for Jupyter notebooks are listed here: