Develop Reference

Can I build and host custom HTML pages at Read the Docs?

My program’s documentation is mainly written in Sphinx, but it also includes two custom HTML pages:
an example report produced by the program;
an extended reference on certain features of the program.
These two HTML files are produced by the program itself, not by Sphinx.
I want to host my docs on Read the Docs, and it would be very convenient for me to build and host the two custom pages, versioned, together with the Sphinx docs.
My program is already installed in the RtD build environment as I have the Install Project option enabled. And since the RtD docs mention writing your own builder, I gather it might be possible to invoke my program from there and have it dump the HTML content in a specific place.
So I really have two questions:
Is this an appropriate use of Read the Docs? I guess it’s not designed to host arbitrary Web pages — but then again, those files are not arbitrary, they are an important part of the docs.
How would I implement it? I’m having a hard time making sense of the RtD API: is this “builder” related in any way to Sphinx builders? how do I hook it up to RtD? perhaps there is an example somewhere?

I achieved the desired result using Sphinx’s html_extra_path feature:
A list of paths that contain extra files [...] They are copied to the output directory.
To generate these files, I haven’t found a better place than right in my conf.py, which seems a bit precarious, but works so far. Of course, Install your project inside a virtualenv needs to be enabled in Read the Docs advanced settings.
Now my custom notices.html and showcase.html are treated just like the .html pages produced by Sphinx itself, with versioning and redirects: http://httpolice.readthedocs.io/page/notices.html

Documenting parameters with long descriptions in Sphinx

I'm documenting my Python classes using Sphinx, and sometimes I want to give my parameters quite long descriptions to explain something in details. Unfortunately, Sphinx generates ugly output for me which wastes a lot of space and breaks the whole page appearance:
It can be seen that Sphinx creates a table, then puts "Parameters" header to the left cell, and the actual list of parameters to the right cell. But there should be way to avoid creating this table completely. After playing with the page DOM tree I finally can show that I want to achieve:
Is there a built-in way to do this or I'd have to create a PR to Sphinx theme or Sphinx itself?

After posting an issue to Sphinx bug-tracker and having no response, I've decided to roll-out my own solution (better say, hack). To achieve the look I want, I have written a simple Sphinx extension which post-processes generated HTML code. It can be found on PyPI:
https://pypi.python.org/pypi/sphinxcontrib.divparams
https://pythonhosted.org/sphinxcontrib.divparams/
This doesn't seem to be the best way to solve the issue, but the behaviour I wanted to change is deeply hard-coded in docutils, not Sphinx.

Defining a default URL prefix using markdown / django-wiki

I've got a Django app that I'm working on, with a wiki (powered by django-wiki) sitting under the wiki/ folder.
The problem I'm having is that if I create links using Markdown, they all direct to the root /, whereas I want any links generated from the markdown to go into the same subdirectory (under /wiki). The documentation doesn't appear to be particularly forthcoming on this (mainly directing me to to the source code, which so far has revealed nothing).
The other avenue I'm looking into is how to direct Markdown itself to prefix all links with a specified path fragment. Is there a Markdown extension or trick that might be useful for accomplishing this?

Well, it looks like there is just such an extension for this in MarkDown (WikiLinkExtension - which takes a base_url parameter).
I've had to modify my copy of django-wiki to add a new setting to use it (submitted an upstream pull request for it too, since I suspect this will be useful to others). Kinda surprised django-wiki didn't have this support already built but there you go.
EDIT: Ok, it looks like this approach doesn't play nice with a hierarchical Wiki layout (which Django-wiki is designed for). I've cobbled together a hack that allows me to link to child pages of the current page, which is enough to be workable even if it's kind of limited.

Conditionally compile document sections in Sphinx

We are planning on creating a user manual for our software project. Currently, everything related to the code is documented in Sphinx, and we would like to use Sphinx for the manual.
Since we are writing scientific/engineering software, there will be a lot of topics on things like stress, strain, numerical algorithms, etc. For each topic, we will have a few stages of information:
Basic information: This one or two sentence description can be used anywhere we need a short summary of the topic. Example: simple definition of mechanical stress.
More detailed description: This one paragraph explanation can be used as the opening to a help page, or a summary in a more detailed listing of topics. Example, a paragraph about mechanical stress introducing the equation for axial stress.
Technical information. This can be multiple paragraphs on how the topic applies to the problems encountered by users of our software.
Code information. This will be documentation related to where the topic is encountered in the code. For example, we can point to our implementation of a certain numerical algorithm. We can use sphinx-apidoc like we currently do.
As you can see, the information gradually gets more complex. We would like to manage each topic in their own .rst file, and get the required information as needed. For example, maybe we want to use the basic information section in a tooltip. In the actual help menu, we can use the detailed description in a table of contents on a certain class of topics. In the full article on the topic, like what would be present in a full pdf manual, we can present the technical information along with the basic and more detailed descriptions. Lastly, we would like to keep code information only in our internal documentation.
It would be nice to keep all of the information for a single topic in one file, but to conditionally compile different sections based on the documentation we are generating.
Is there a built-in way to do something like this in Sphinx? If someone is doing something similar, can you tell us about it and give us some highlights of your workflow?

In past I wanted to compile two docs, a public and a private but I didn't want to split my source file (rst).
First step I found the only directive and I thought it was the solution. But when I wanted a entire full rst file in just public or private documentation I can't without indent the whole file.
So I write my own Sphinx plugin (scope) to manage all my case. To succeed I used the meta directive that can be place on the top of the file.
Thus
a_doc_for_basic.rst
.. meta::
:scope: basic
Title
=====
My content
a_doc_for_code.rst
.. meta::
:scope: code
Title
=====
My content
And you can continue to use .. only:: directive on file
a_doc_for_all.rst
Title
=====
My content
.. only:: code
a piece of code
You can find plugin source here
As you can see the plugin is pretty simple and works thanks to regexp. That means (regexp) that there is limitation:
The directive .. meta:: :scope: must be place at the top of the file (no line before)
The directive .. meta:: :scope: must match the regexp ^\.\. meta::\s+:scope: ([a-zA-Z0-9_-]+)
The directive .. meta:: :scope: can manage multiple tag but you can easily update the plugin for your needs
Plugin deviate the original use of meta directive docutils.sourceforge.net/docs/ref/rst/directives.html#meta
After that you can build your doc using the following command
sphinx-build ... -t <tag> ...
sphinx-build ... -t code ...
Other thing, you can use the same toctree for all tag because when compiling a doc per tag the toctree will be edited by the plugin to generate a tree without reference to the no-matching documentation.
PS: my plugin is not perfect but I replied to my needs, you can inspired and update it.

have sphinx report broken links

When building html documentation, how do you force sphinx to report, or create an error, on links that don't exist?
Specifically, I have properties and methods within my Python project that have been removed or renamed, and it is hard to find all the dead links with the sphinx generated html output.
I feel like I'm staring at the answer here:
http://sphinx-doc.org/glossary.html, as descriped in the opening paragraph.
I'm obviously not understanding something.

Set the nitpicky configuration variable to True (you can also use the -n option when running sphinx-build).
In nitpicky mode, a cross-reference to a function (such as :func:`myfunc`), class, or other object that cannot be found will generate a warning message.

I think CheckExternalLinksBuilder is what you're looking for.
It's basically used by calling 'sphinx-build' with -b linkcheck option. Please see sphinx-build for more info. Also, take a look at the list of sphinx-extensions here and here.