2. Writing Guidelines¶
In general, when creating reST documentation for the QGIS project, please follow the Python documentation style guidelines. For convenience, we provide a set of general rules we rely on for writing QGIS documentation below.
To each webpage of the documentation corresponds a
Sections used to structure the text are identified through their title which is underlined (and overlined for the first level). Same level titles must use same character for underline adornment. In QGIS Documentation, you should use following styles for chapter, section, subsection and minisec.
******** Chapter ******** Section ======= Subsection ---------- Minisec ....... Subminisec ^^^^^^^^^^
Lists are useful for structuring the text. Here are some simple rules common to all lists:
Start all list items with a capital letter
Do not use punctuation after list items that only contain a single simple sentence
Use period (
.) as punctuation for list items that consist of several sentences or a single compound sentence
Indentation in ReStructuredText should be aligned with the list or markup marker. It is also possible to create block quotes with indentation. See the Specification
#. In a numbered list, there should be three spaces when you break lines #. And next items directly follow * Nested lists * Are also possible * And when they also have a line that is too long, the text should be naturally aligned * and be in their own paragraph However, if there is an unindented paragraph, this will reset the numbering: #. This item starts at 1 again
Anchors inside the text can be used to create hyperlinks to sections or pages.
The example below creates the anchor of a section (e.g., Label/reference title)
.. _my_anchor: Label/reference ---------------
To call the reference in the same page, use
see my_anchor_ for more information.
which will return:
see my_anchor for more information.
Notice that it will jump to the line/thing following the ‘anchor’. You do not need to use apostrophes, but you do need to have empty lines after the anchor.
Another way to jump to the same place from anywhere in the
documentation is to use the
see :ref:`my_anchor` for more information.
which will create a link with the caption instead (in this case the title of this section!):
see Labels/references for more information.
So, reference 1 (my_anchor) and reference 2 (Labels/references). Because the reference often displays a full caption, it is not really necessary to use the word section. Note that you can also use a custom caption to describe the reference:
see :ref:`Label and reference <my_anchor>` for more information.
see Label and reference for more information.
You can put an image inside text or add an alias to use everywhere. To use an image
inside a paragraph, first create an alias in the
.. |nice_logo| image:: /static/common/logo.png :width: 1 em
and then call it in your paragraph:
My paragraph begins here with a nice logo |nice_logo|.
This is how the example will be displayed:
To allow preview rendering in GitHub that is as close as possible
to HTML rendering, you will also need to add the image replacement
call at the end of the file you changed.
This can be done by copy-pasting it from
substitutions.txt or by executing the
Currently, to ensure consistency and help in the use of QGIS icons, a list of aliases is built and available in the Substitutions chapter.
.. _figure_logo: .. figure:: /static/common/logo.png :width: 20 em :align: center A caption: A logo I like
The result looks like this:
To avoid conflicts with other references, always begin figure
_figure_ and use terms that easily connect to the
figure caption. While only the centered alignment is mandatory for the image,
feel free to use any other options for figures (such as
scale…) if needed.
The scripts will insert an automatically generated number before the caption of the figure in the generated HTML and PDF versions of the documentation.
To use a caption (see My caption) just insert indented text after a blank line in the figure block.
A figure can be referenced using the reference label like this:
renders like this:
see Fig. 2.20
This is the preferred way of referencing figures.
:numref: to work, the figure must have a caption.
It is possible to use
:ref: instead of
:numref: for reference, but
this returns the full caption of the image.
renders like this:
A simple table can be coded like this
======= ======= ======= x y z ======= ======= ======= 1 2 3 4 5 ======= ======= =======
It will render like this:
\ (backslash) followed by an empty space to leave an empty space.
You can also make more complicated tables and reference them:
.. _my_drawn_table: +---------------+--------------------+ | Windows | macOS | +---------------+--------------------+ | |win| | |osx| | +---------------+--------------------+ | and of course not to forget |nix| | +------------------------------------+ My drawn table, mind you this is unfortunately not regarded as a caption You can reference it like this: my_drawn_table_.
My drawn table, mind you this is unfortunately not regarded as a caption
You can reference to it like this my_drawn_table.
For even more complex tables, it is easier to use
.. list-table:: :header-rows: 1 :widths: 20 20 20 40 * - What - Purpose - Key word - Description * - **Test** - ``Useful test`` - complexity - Geometry. One of: * Point * Line
Geometry. One of:
An index is a handy way to help the reader find information in a document. QGIS documentation provides some essential indices. There are a few rules that help us provide a set of indices that are really useful (coherent, consistent and really connected to each other):
An index should be human readable, understandable and translatable; an index can be made from many words but you should avoid any unneeded
-… characters to link them i.e.,
Loading layersinstead of
Capitalize only the first letter of the index unless the word has a particular spelling. E.g.,
Keep an eye on the existing Index list in order to reuse the most convenient expression with the right spelling and avoid unnecessary duplicates.
Several index tags exist in RST. You can use the inline
within normal text:
QGIS can load several :index:`Vector formats` supported by GDAL/OGR ...
Or you can use the
.. index:: block-level markup which links to the
beginning of the next paragraph. Because of the rules mentioned above, it
is recommended to use the block-level tag:
.. index:: WMS, WFS, Loading layers
It is also recommended to use index parameters such as
see, in order to build a more structured and interconnected index table.
See Index generating
for more information on index creation.
Sometimes, you may want to emphasize some points of the description, either to
warn, remind or give some hints to the user. In QGIS Documentation, we use reST
special directives such as
.. seealso::`, ``.. note:: and
These directives generate frames that highlight your comments. See Paragraph Level markup
for more information.
A clear and appropriate title is required for both warnings and tips.
.. tip:: **Always use a meaningful title for tips** Begin tips with a title that summarizes what it is about. This helps users to quickly overview the message you want to give them, and decide on its relevance.
You may also want to give examples and insert code snippets. In this case,
write the comment below a line with the
:: directive inserted. For
a better rendering, especially to apply color highlighting to code according
to its language, use the code-block directive, e.g.
.. code-block:: xml.
More details at
While texts in note, tip and warning frames are translatable, be aware that code block frames do not allow translation. So avoid comments not related to the code and keep comments as short as possible.
Please note: Footnotes are not recognized by any translation software and it is also not converted to pdf format properly. So, if possible, don’t use footnotes within any documentation.
This is for creating a footnote (showing as example 1)
Which will point to:
Updates of core plugins
Here are some hints to create new, nice looking screenshots.
The images should be placed in an image (
img/) folder that
is located in the same folder as the referencing
You can find some prepared QGIS-projects that are used to create screenshots in the
./qgis-projectsfolder of this repository. This makes it easier to reproduce screenshots for the next version of QGIS. These projects use the QGIS Sample Data (aka Alaska Dataset), which should be placed in the same folder as the QGIS-Documentation Repository.
Reduce the window to the minimal space needed to show the feature (taking the whole screen for a small modal window > overkill)
The less clutter, the better (no need to activate all the toolbars)
Don’t resize them in an image editor; the size will be set into the
.rstfiles if necessary (downscaling the dimensions without properly upping the resolution > ugly)
Cut the background
Make the top corners transparent if the background is not white
Set print size resolution to
135 dpi(e.g. in Gimp set the print resolution and save). This way, images will be at original size in html and at a good print resolution in the PDF. You can also use ImageMagick convert command to do a batch of images:
convert -units PixelsPerInch input.png -density 135 output.png
Save them as
The screenshot should show the content according to what is described in the text
If you are on Ubuntu, you can use the following command to remove the global menu function and create smaller application screens with menus:
sudo apt autoremove appmenu-gtk appmenu-gtk3 appmenu-qt
If you want to write documentation for Processing algorithms, consider these guidelines:
Processing algorithm help files are part of QGIS User Guide, so use the same formatting as User Guide and other documentation.
Each algorithm documentation should be placed in the corresponding provider folder and group file, e.g. the algorithm Voronoi polygon belongs to the QGIS provider and to the group vectorgeometry. So the correct file to add the description is:
Before starting to write the guide, check if the algorithm is already described. In this case, you can enhance the existing description.
It is extremely important that each algorithm has an anchor that corresponds to the provider name + the unique name of the algorithm itself. This allows the Help button to open the Help page of the correct section. The anchor should be placed above the title, e.g. (see also the Labels/references section):
.. _qgisvoronoipolygons: Voronoi polygons ----------------
To find out the algorithm name you can just hover the mouse on the algorithm in the Processing toolbox.
Avoid using “This algorithm does this and that…” as the first sentence in the algorithm description. Try to use more general expressions like:
Takes a point layer and generates a polygon layer containing the...
Avoid describing what the algorithm does by replicating its name and please don’t replicate the name of the parameter in the description of the parameter itself. For example if the algorithm is
Voronoi polygonconsider to describe the
Layer to calculate the polygon from.
Indicate in the description whether the algorithm has a default shortcut in QGIS or supports in-place editing.
Add images! A picture is worth a thousand words! Use
.pngformat and follow the general guidelines for documentation (see the Figures and Images section for more info). Put the image file in the correct folder, i.e. the
imgfolder next to the
.rstfile you are editing.
If necessary, add links in the “See also” section that provide additional information about the algorithm (e.g., publications or web-pages). Only add the “See also” section if there is really something to see. As a good practice, the “See also” section can be filled with links to similar algorithms.
Give clear explanation for algorithm parameters and outputs: take inspiration from existing algorithms.
Avoid duplicating detailed description of algorithm options. Add this information in the parameter description.
Avoid adding information about the vector geometry type in the algorithm or parameter description, as this information is already available in the parameter descriptions.
Add the default value of the parameter, e.g.:
* - **Number of points** - ``NUMBER_OF_POINTS`` - [number] Default: 1 - Number of points to create
Describe the type of input supported the parameters. There are several types available you can pick one from:
Point vector layer
Line vector layer
Polygon vector layer
Generic vector layer
Vector field numeric
Vector field string
Vector field generic
Same output type as input type
same as input
Study an existing and well documented algorithm, and copy all the useful layouts.
When you are finished, just follow the guidelines described in A Step By Step Contribution to commit your changes and make a Pull Request
Here is an example of an existing algorithm to help you with the layout and the description:
.. _qgiscountpointsinpolygon: Count points in polygon ----------------------- Takes a point and a polygon layer and counts the number of points from the point layer in each of the polygons of the polygon layer. A new polygon layer is generated, with the exact same content as the input polygon layer, but containing an additional field with the points count corresponding to each polygon. .. figure:: img/count_points_polygon.png :align: center The labels in the polygons show the point count An optional weight field can be used to assign weights to each point. Alternatively, a unique class field can be specified. If both options are used, the weight field will take precedence and the unique class field will be ignored. ``Default menu``: :menuselection:`Vector --> Analysis Tools` Parameters .......... .. list-table:: :header-rows: 1 :widths: 20 20 20 40 * - Label - Name - Type - Description * - **Polygons** - ``POLYGONS`` - [vector: polygon] - Polygon layer whose features are associated with the count of points they contain * - **Points** - ``POINTS`` - [vector: point] - Point layer with features to count * - **Weight field** Optional - ``WEIGHT`` - [tablefield: numeric] - A field from the point layer. The count generated will be the sum of the weight field of the points contained by the polygon. * - **Class field** Optional - ``CLASSFIELD`` - [tablefield: any] - Points are classified based on the selected attribute and if several points with the same attribute value are within the polygon, only one of them is counted. The final count of the points in a polygon is, therefore, the count of different classes that are found in it. * - **Count field name** - ``FIELD`` - [string] Default: 'NUMPOINTS' - The name of the field to store the count of points * - **Count** - ``OUTPUT`` - [vector: polygon] Default: [Create temporary layer] - Specification of the output layer type (temporary, file, GeoPackage or PostGIS table). Encoding can also be specified. Outputs ....... .. list-table:: :header-rows: 1 :widths: 20 20 20 40 * - Label - Name - Type - Description * - **Count** - ``OUTPUT`` - [vector: polygon] - Resulting layer with the attribute table containing the new column with the points count