plotsettings 1.0.4-2

Easily format figures to match the publication requirements of an academic journal

The difference between publishing figures using just matplotlib (A), and using matplotlib with plotsettings (B):

Highlights

Tearing your hair out trying to make your figures conform to the author guidelines for a particular journal? Got rejected from PNAS and can’t bring yourself to change all the fonts for submission to PLOS ONE? Let plotsettings make it easy for you! plotsettings is a convenient way of making sure your figures fit the requirements for publication. One line is sufficient to choose a target journal, and just one more line to automatically output figures that fit cleanly into 1, 2 or even 1.5 columns! You can even set the aspect ratio of your figure and be warned if your figure gets taller than the height of one page. In fact, plotsettings already knows the appropriate font, text size, and figure dimensions for all of these journals:

• Cell (use argument ‘Cell’)

• Copeia (use argument ‘Copeia’)

• Deep Sea Research II (use argument ‘DSRII’)

• Ecology Letters (use argument ‘EcolLett’)

• Global Change Biology (use argument ‘GlobChangeBio’)

• Global Environmental Change (use argument ‘GlobEnvChange’)

• Integrative and Comparative Biology (use argument ‘IntCompBiol’)

• Journal of Experimental Biology (use argument ‘JEB’)

• Limnology and Oceanography (use argument ‘LimnolOcean’)

• Marine Ecology Progress Series (use argument ‘MEPS’)

• Nature magazine (use argument ‘Nature’)

• Oecologia (use argument ‘Oecologia’)

• Proceedings of the National Academy of Sciences, USA (use argument ‘PNAS’)

• Proceedings of the Royal Society B (use argument ‘ProcRoySocB’)

• Public Library of Science One (use argument ‘PLOSOne’)

• Public Library of Science Biology (use argument ‘PLOSBio’)

• Science magazine (use argument ‘Science’)

• Presentation (okay, this is not a journal but it’s still useful for outputting figures to presentation slides; access with the argument ‘Presentation’)

Don’t see the journal you want on the list (say you want to publish in the Proceedings of the 6th ACM Workshop on Next Generation Mobile Computing for Dynamic Personalised Travel Planning)? Compile your own list of journals by creating a python file containing a single dictionary with settings for every journal you use! You can specify any parameters that are accepted by matplotlib.rcParams as well as the column width, gutter width, page height and the way that multi-panel figures are labeled.

Also, a Bonus!:

• 1-line labeling of all the subplots in a figure (e.g. with ‘(a)’, ‘(b)’, ‘(c)’ etc.) using the standalone function panel_labels or the method Set.panel_labels!

Installation

plotsettings has only been tested in Python 2.7

Install through pip:

$ pip install plotsettings

Requires the following non-standard libraries:

• matplotlib

Because preferred installation of matplotlib can vary depending on the operating system, matplotlib will not automatically be installed as a dependency. Instead, installation will raise an exception if matplotlib cannot be found in the pythonpath. In this case, please install matplotlib via your preferred method, most of which are explained by matplotlib

Usage

First set the journal you want to submit to:

publishable = plotsettings.Set('MEPS') # Lets publish in Marine Ecology Progress Series!

Then set the dimensions of a particular figure with the line:

publishable.set_figsize(n_columns = 1, n_rows = 1)

This will cause the next figure that is drawn to be 1 column wide (81 mm for MEPS) x 1 row high (the concept of ‘rows’ is a little made up, but the default is that one row is equal to one column width multiplied by the golden ratio, so in this case 50.1 mm). Once the first figure is drawn, we can set the next figure to be 2 columns wide and 1 row tall and this time set the row height to be equal to the column width like this:

publishable.set_figsize(2, 1, aspect_ratio = 1)

Importantly, plotsettings doesn’t just calculate the width of a 2-column figure as two times the width of one column, but includes the width of the gutter (the space between columns on a page) as well. Therefore, the figure that follows the above line will end up being 169 mm wide (2 columns of 81 mm each plus a 7 mm gutter) and 81 mm tall (row height = 1*column width).

Once the figure has been created, conveniently add labels to each subplot (if you have a multipart figure):

publishable.panel_labels(fig = fig, position = 'outside', case = 'lower',
                                                 prefix = '', suffix = '.', fontweight = 'bold')

to create the labels (‘a.’, ‘b.’, ‘c.’ …) in bold letters outside the top-left corner of each subplot.

Custom journal settings can be used by specifying a python file on the PYTHONPATH:

publishable = plotsettings.Set('my_journal_name', 'module_name')

The file ‘module_name’ should contain a single dictionary named journals with the following structure:

journals = {'journal1':      {'rcParams':     {'param1': value1,
                                               'param2': value2 ...},
                              'figsize':      {'param1': value1,
                                               'param2': value2 ...},
                              'panel_labels': {'param1': value1,
                                               'param2': value2 ...},
            'journal2':      {'rcParams':     {'param1': value1,
                                               'param2': value2 ...},
                              'figsize':      {'param1': value1,
                                               'param2': value2 ...},
                              'panel_labels': {'param1': value1,
                                               'param2': value2 ...},
            'journal3'...
            }

where ‘journalx’ are the identifying names of academic journals (e.g. ‘Nature’), with the specifications for each journal being divided into 3 dictionaries:

• rcParams: All parameters are optional. Any valid input to pyplot.rcParams (for example font name and sizes, default linewidths) is accepted. Definitions of valid keys to rcParams can be found here.

• figsize: Set figure dimension calculations. Requires the parameters column_width, gutter_width, and units. The parameter max_height is optional. See below for details.

• panel_labels: Set default panel labels (i.e. the text that identifies each subplot in a figure as A, B, C, etc.). All parameters are optional. See below for details.

The possible non-rcParams parameters are:

• figsize:

  • column_width (required) - the maximum width a figure is allowed to be while still fitting withing a single column.

  • gutter_width (required) - the width of the gutter (space between columns). This can usually be found by comparing the maximum width that a journal allows for a single- column figure with the maximum width of a 2-column figure. For example, PLoS One allows a 1-column figure to be 83 mm in width and a 2-column figure to be 173.5 mm, meaning that the gutter width (173.5 - 83*2) must be 7.5 mm wide.

  • max_height (optional) - the maximum height a figure is allowed to be while fitting on a single page (i.e. the page height).

  • units (required) - the units in which the above are reported. Can be one of ‘mm’, ‘cm’, ‘inch’, or ‘pts’

• panel_labels:

  • fontweight (optional) - the font weight of panel annotations (e.g. A, B, C etc.). Default is ‘bold’

  • case (optional) - whether to capitalize (‘upper’) or not capitalize (‘lower’) the panel labels.

  • prefix (optional) - characters to prepend to panel label (e.g. if the desired label style is (A), (B), etc., set label_prefix to ‘)’).

  • suffix (optional) - characters to append to panel label (e.g. if the desired label style is a., b., etc., set label_suffix = ‘.’)

  • fontsize (optional) - font size in pts of the label. Defaults to rcParams[‘font.size’]

Changelog

1.0.4-2 (NOVEMBER/14/2014)

• Updated doctests

1.0.4-1 (NOVEMBER/14/2014)

• Removed deprecated keywords from default journal parameters

1.0.4 (OCTOBER/20/2014)

• Added ability for panel_labels method to automatically detect axes that contain only colorbars and not label them (use detect_colorbars = True). This method relies on the assumption that colorbar axes are not navigable (e.g. cannot be panned or zoomed in the interactive figure). This property was chosen because it seems to work both for colorbars created by pyplot.colorbar() as well as those created explicitly in a new axis such as using the AxesGrid toolkit.

1.0.3 (OCTOBER/15/2014)

• package matplotlib is no longer explicitly required in setup.py. Installation will raise an error if matplotlib is not present - please install in your preferred way.

1.0.2 (OCTOBER/15/2014)

• Changed format of dictionary specifying journal settings. Settings for each journal are now divided between the dictionaries ‘rcParams’, ‘figsize’ and ‘panel_labels’ instead of being amalgamated into a single dictionary.

• Added journals Science, Integrative and Comparative Zoology, Copeia, Cell, Global Change Biology, Global Environmental Change, Limnology and Oceanography, Nature, PLOS Biology and Oecologia to the list of natively supported publications.

• Added method panel_labels to class Set to allow panel labels (e.g. A, B, C etc.) to automatically follow default settings for the journal (e.g. boldface, uppercase, etc.)

1.0.1 (OCTOBER/14/2014)

• Fixed bug that made plotsettings.Set unable to find the default journals module

• Added function panel_labels for convenient, 1-line addition of formatted panel labels (e.g. A, B, C) to every subplot in a figure.

• Added ‘Presentation’ as a journal type for PowerPoint slides

1.0.0 (OCTOBER/13/2014)

• First release