An open library of matplotlib styles

## Project description

# mpls
mpls is a small lightweight pure Python package that manages the access to a repository of matplotlib styles and makes
it easy to use and switch between different styles. To specify the style of the figures in your Python script or
notebook just write
python
import matplotlib.pyplot as plt
import mpls

mpls.use('gplot', context='a5')
...

and depending on the configuration of your style library you have access to styles from a remote (public) library or the
personal style library on your computer.

I work as a scientist and one of the major tasks of my daily work is to create reports, publications, and presentations.
Personally, I prefer figures over text, but, I know, sometimes it is inevitable to add some paragraphs of text between the
figures. However, being faced with the situation to produce many figures for _multiple contexts_ (report, journal, slides,
etc.) with _consistent_ _style_ and _colors_, I started to use matplotlib stylesheets to be able to store and load my
custom settings. To be honest, I didn't enjoy working with stylesheet files very much due to a couple of reasons (e.g. ).

Similar to [seaborn][1], mpls styles are organized in three types: **context**, **style**, and
**palette**.

The **context** defines size relevant properties such as figure size, size of marker
points, etc.
More details on which rcParams belong to the **context** can be found [here](doc/context.md).

The **style** defines the general style properties such as basic figure _colors_ and _fonts_.
More details on which rcParams belong to the **style** can be found [here](doc/style.md).

The **palette** defines the color properties such as the colormap of images or the
color of markers and lines.
More details on which rcParams belong to the **palette** can be found [here](doc/palette.md).

### Style files
Technically, each of the three style types represents a subset of the rcParams dict in [matplotlib][2]. The
parameters dict instances are stored in a in so-called _style files_. These style files are regular _json_ files, with
the exception that it is allowed to have C-style comments in the file. A small example of a valid **context** style file
is
javascript
{
// this is a line comment
"figure.figsize": [3, 4],

/* this is a multi-
line comment */
"font.size": 10
}

For more examples of style files have a look at the files in the [stylelib](stylelib/) of this repository.

_Note_: In the current version, parameters defined in style files are not restricted to a certain subset of rcParams.
Generally, you can define any parameter in any style file! This is useful in some cases. However, this may change in
the future and in order to make working with style files easy and intuitive, please restrict yourself to the
parameters specified in the respective style file documentation.

## Installation
The easiest way to install mpls is to use pip, i.e.

pip install mpls


However, if you want to work with the most recent version of mpls, you can just clone the repository and run
setuptools in the source folder like

python setup.py install


## Requirements
There is only one (quite obvious) requirement
- matplotlib

## Examples
An example where the plotting context is modified temporarily.
python
import matplotlib.pyplot as plt
import mpls

mpls.use(context='a4-gr', style='thesis', palette='grayscale')

# create some plot
...

with mpls.temp(context='a4-gr-ls'):
# temporarily switch to A4 landscape format
...

# continue with regular A4 format
...


An example of mixing matplotlib and mpls styles
python
import matplotlib.pyplot as plt
import mpls

# mix a matplotlib style with an mpls palette
mpls.use('dark_background', palette='grayscale')

# create some plots
...


## Contributing
Contributions to the mpls code or the stylelib in this repository are very welcome. Just issue a pull request at the
github page.

## Using a custom style library
As default mpls fetches styles from the stylelib folder in this repository. But it is also possible to fetch files
from any other _remote_ or _local_ repository. The easiest way to fetch mpls styles from a custom style library is to
provide the style_url parameter when calling use or temp, e.g.
python
import mpls
mpls.use(context='a4', style='thesis', style_url='http://some.other.repository.com/stylelib/{type}_{name}.json')


If you want to switch the style library for a longer session, it is more convenient to change the default stylelib_url and stylelib_format in your mpls configuration, i.e.
python
import mpls
...
# switch to another remote stylelib temporarily
mpls.configure(stylelib_url="http://some.other.repository.com/stylelib/" stylelib_format="{type}_{name}.json")

# or switch to a local stylelib (and keep current the stylelib_format)
mpls.configure(stylelib_url="~/stylelib/{type}/{name}.json")

Two placeholders {type} and {name} may be placed in the url. Internally, mpls will
substitute these placeholders with the parameters provided in the frontend methods. For example, the call
python
mpls.use(context='a4')

boils down to
python
mpls.get(name='a4', type='context')

which eventually calls
python
style_url = stylelib_url + stylelib_format.format(name=name, type=type)

to replace the placeholders in the stylelib_format to retrieve the actual style file url.

To make your changes permanent, just provide the save=True parameter when switching the stylelib_url in the
configuration or call configure later, i.e.

# save changes immediately
mpls.configure(stylelib_url='~/path/to/stylelib/', save=True)
# or configure some parameter first
mpls.configure(stylelib_url='~/path/to/stylelib/')
[...]
# and save changes some later time
mpls.configure(save=True)

This will save your changes to the mpls configuration file and which is loaded every time mpls is initialized.

If you are not convinced or just want to know a bit more about how to modify the style of your matplotlib plots, please
refer to the respective [matplotlib page][3] for more information on how to customize the style of your plots directly
with matplotlib, or visit the [seaborn][1] website to have a look at another popular matplotlib style package (and much
more).

[1]: http://seaborn.pydata.org
[2]: http://matplotlib.org
[3]: http://matplotlib.org/users/customizing.html
[4]: https://docs.python.org/3.7/library/stdtypes.html?highlight=str.format#str.format

## Project details

Uploaded source