1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
|
INTRODUCTION
The Sphinx documentation system uses the fully installed Python tree
to build a set of documents (generally in HTML). In GNU Radio, the
documentation system is done through Doxygen in the public header
(/include/foo.h) files. Doxygen first builds its documentation files,
then the swig_docs program uses Doxygen's XML output and smashed the
documentation from each header file into the SWIG'd Python
block. Basically, using a single documentation markup, Doxygen, we
expose the documentation strings in both the Doxygen-built manual and
within the Python blocks themselves.
Sphinx takes this process one step farther by reading the docstrings
of all Python blocks and creating its own manual. This has two
benefits. First, the Sphinx documentation looks nice and is formatted
in such a way that Python users of GNU Radio can easy see the module
structure and hierarchy. It also not only takes the Doxygen
documentation from C++, but it also allows us to take any Python files
and include their documentation.
The end result is two manuals: one for Python and one for C++ users
without having to duplicate comments, markup, or documentation.
BUILDING THE SPHINX MANUAL
Building the Sphinx docs takes some manual intervention as it
requires GNU Radio to already be installed. So first follow the steps
to build and install GNU Radio.
In the build directory, a helper file is created called
run_sphinx_build.sh. This is a Linux shell script that runs the
sphinx-build command with all of the normal settings and important
directories preloaded. For non Linux systems, it should be easy to
pull out the executable and options to run it by hand.
The run_sphinx_build.sh outputs the manual into
$builddir/docs/sphinx/sphinx_out. Open up the index.html file in a
browser to view it.
ADDING NEW CONTENT TO THE SPHINX MANUAL
Although the content of the sphinx manual is automatically generated,
new blocks are not automatically added to the generated documentation.
The procedure for adding new content is best illustrated with two
examples.
1) Adding a new C++ signal processing block gnuradio.gr.myslicer
Edit file gnuradio/docs/sphinx/source/gr/index.rst and add the line
> gnuradio.gr.myslicer
under the "Slicing and Dicing Streams" subheading.
Edit file gnuradio/docs/sphinx/source/gr/slicedice_blk.rst and add
the line
>.. autooldblock:: gnuradio.gr.myslicer
2) Adding a new python hierarchical block gnuradio.digital.mymod
Edit file gnruadio/docs/sphinx/source/digital/index.rst and add the
line
> gnuradio.digital.mymod
under the "Signal Processing Blocks" subheading.
Edit file gnuradio/docs/sphinx/source/digital/blocks.rst and add
the line
>.. autopyblock:: gnuradio.digital.mymod
Notice that the 'autopyblock' directive is used rather than the
'autoblock' directive. This lets sphinx know that it is displaying
a python hierarchical signal processing block so that it can format
it appropriately.
The process for documenting objects that are not signal processing
blocks is similar but rather than using the 'autooldblock', and
'autopyblock' directives the standard sphinx directives such as
'autofunction' and 'autoclass' can be used.
Finally for signal processing blocks using the 3.7 style the directive
'autoblock' rather than 'autooldblock' can be used.
|
