Jupytext

------------------

**Added**

- The `language_info` section is not part of the default header any more. Language information is now taken from metadata `kernelspec.language`. ([105](https://github.com/mwouts/jupytext/issues/105)).
- When opening a paired notebook, the active file is now the file that was originally opened ([118](https://github.com/mwouts/jupytext/issues/118)). When saving a notebook, timestamps of all the alternative representations are tested to ensure that Jupyter's autosave does not override manual modifications.
- Jupyter magic commands are now commented per default in the `percent` format ([126](https://github.com/mwouts/jupytext/issues/126), [#132](https://github.com/mwouts/jupytext/issues/132)). Version for the `percent` format increases from '1.1' to '1.2'. Set an option `comment_magics` to `false` either per notebook, or globally on Jupytext's contents manager, or on `jupytext`'s command line, if you prefer not to comment Jupyter magics.
- Jupytext command line has a pre-commit mode ([121](https://github.com/mwouts/jupytext/issues/121)).

------------------

**Added**

- `bash` scripts as notebooks ([127](https://github.com/mwouts/jupytext/issues/127))
- R scripts with `.r` extension are supported ([122](https://github.com/mwouts/jupytext/issues/122))
- Jupytext selects the first kernel that matches the language ([120](https://github.com/mwouts/jupytext/issues/120))

------------------

**Added**

- Notebook metadata is filtered - only the most common metadata are stored in the text representation ([105](https://github.com/mwouts/jupytext/issues/105))
- New config option `freeze_metadata` on the content manager and on the command line interface (defaults to `False`). Use this option to avoid creating a YAML header or cell metadata if there was none initially. ([110](https://github.com/mwouts/jupytext/issues/110))
- Language magic arguments are preserved in R Markdown, and also supported in `light` and `percent` scripts ([111](https://github.com/mwouts/jupytext/issues/111), [#114](https://github.com/mwouts/jupytext/issues/114), [#115](https://github.com/mwouts/jupytext/issues/115))
- First markdown cell exported as a docstring when using the Sphinx format ([107](https://github.com/mwouts/jupytext/issues/107))

------------------

**Added**

- Frozen cells are supported in R Markdown, light and percent scripts ([101](https://github.com/mwouts/jupytext/issues/101))
- Inactive cells extended to percent scripts ([108](https://github.com/mwouts/jupytext/issues/108))
- `jupytext` gains a `--version` argument ([103](https://github.com/mwouts/jupytext/issues/103))
- "ExecuteTime" cell metadata is not included in the text representation anymore ([106](https://github.com/mwouts/jupytext/issues/106))

------------------

**Added**

- Round trip conversion testing with `jupytext --test` was improved ([99](https://github.com/mwouts/jupytext/issues/99))
- Round trip conversion tested on Jake Vanderplas' Python for Data Science Handbook.

**Fixed**

- Nested lists and dictionaries are now supported in notebook metadata
- Final empty code cell supported in Sphinx representation

------------------

**Fixed**

- Sphinx format tested on `World population` notebook ([97](https://github.com/mwouts/jupytext/issues/97))
- Mirror test made stronger on this occasion!
- Markdown representation recognize Julia, Scheme and C++ code cells as such
- Light representation of Scheme and C++ notebooks fixed ([61](https://github.com/mwouts/jupytext/issues/61))