Python 2/3 and IPython 3 compatible!

notedown is a simple tool to create IPython notebooks from markdown (and r-markdown).

notedown separates your markdown into code and not code. Code blocks (fenced or indented) go into input cells, everything else goes into markdown cells.

Usage:

notedown input.md > output.ipynb

Installation:

pip install notedown

or the latest on github:

pip install https://github.com/aaren/notedown/tarball/master

Convert a notebook into markdown, stripping all outputs:

notedown input.ipynb --to markdown --strip > output.md

Convert a notebook into markdown, with output JSON intact:

notedown input.ipynb --to markdown > output_with_outputs.md

The outputs are placed as JSON in a code-block immediately after the corresponding input code-block. notedown understands this convention as well, so it is possible to convert this markdown-with-json back into a notebook.

This means it is possible to edit markdown, convert to notebook, play around a bit and convert back to markdown.

NB: currently, notebook and cell metadata is not preserved in the conversion.

Strip the output cells from markdown:

notedown with_output_cells.md --to markdown --strip > no_output_cells.md

notedown notebook.md --run > executed_notebook.ipynb

You can configure IPython to seamlessly use markdown as its storage format. Add the following to ipython_notebook_config.py in your ipython profile (probably ~/.ipython/profile_default):

c.NotebookApp.contents_manager_class = 'notedown.NotedownContentsManager'

Now you can edit your markdown files in the browser, execute code, create plots - all stored in markdown!

You can use notedown to convert r-markdown as well. We just need to tell notedown to use knitr to convert the r-markdown. This requires that you have R installed with knitr.

Convert r-markdown into markdown:

notedown input.Rmd --to markdown --knit > output.md

Convert r-markdown into an IPython notebook:

notedown input.Rmd --knit > output.ipynb

• --rmagic will add %load_ext rpy2.ipython at the start of the notebook, allowing you to execute code cells using the rmagic extension (requires rpy2). notedown does the appropriate %R cell magic automatically.

Fenced code blocks annotated with a language other than python are read into cells using IPython's %% cell magic.

You can disable this with --nomagic.

• --pre lets you add arbitrary code to the start of the notebook. e.g. notedown file.md --pre '%matplotlib inline' 'import numpy as np'

By using the --match argument. notedown defaults to converting all code-blocks into code-cells. This behaviour can be changed by giving a different argument to --match:

• --match=all: convert all code blocks (the default)

• --match=fenced: only convert fenced code blocks

• --match=language: only convert fenced code blocks with 'language' as the syntax specifier.

• --match=strict: only convert code blocks with Pandoc style attributes containing 'python' and 'input' as classes. i.e. code blocks must look like

    ```{.python .input}
    code
    ```
  

Try editing the markdown in the IPython Notebook using the NotedownContentsManager (see above).

You can get an interactive ipython session in vim by using vim-ipython, which allows you to connect to a running ipython kernel. You can send code from vim to ipython and get code completion from the running kernel. Try it!

Try using either vim-markdown or vim-pandoc. Both are clever enough to highlight code in markdown.

This is experimental!

Convert a notebook into markdown, rendering cell outputs as native markdown elements:

notedown input.ipynb --render

This means that e.g. png outputs become ![](data-uri) images and that text is placed in the document.

Of course, you can use this in conjuntion with runipy to produce markdown-with-code-and-figures from markdown-with-code:

notedown input.md --run --render > output.md

Not a notebook in sight!

The --render flag forces the output format to markdown.

• Python 3 support
• unicode support
• IPython 3 support
• Allow kernel specification