rstdoc 1.8.0

rstdoc - support documentation in restructedText (rst)

See background and documentation.

Many companies use DOCX and thus produce an information barrier. Working with text is more integrated in the (software) development process. A final format can be DOCX, but, at least during development, text is better.

Sphinx is an extension of Docutils used for many (software) projects, but it does not support creation of DOCX files, which certain companies demand. Pandoc does support DOCX, but does not support the Sphinx extensions, hence :ref: and the like cannot be used.

This python package supports working with RST as documentation format without depending on Sphinx.

• link RST documents using substitutions (generated in _links_xxx.r?st)

• create a .tags file to jump around in an editor that support ctags

• RST handling with python: reformat/create RST tables

• post-process Pandoc’s conversion from DOCX to RST

• pre-process Pandoc’s conversion from RST to DOCX

• Support in building with WAF (or Makefile)

  • expand SimpleTemplate template files .stpl

  • graphics files (.tikz, .svg, .dot, .uml, .eps or .stpl thereof, and .pyg) are converted to .png and placed into ./_images or <updir>/_images, if there, else into current directory.

  • a gen file specifies how RST should be generated from source code files (see dcx.py)

The conventions used are shown

• by the example produced via rstdcx --rest smplfldr or rstdcx --stpl smplfldr

• by the documentation sources that can be found at https://github.com/rpuntaie/rstdoc/tree/master/doc

pip install rstdoc installs:

Module	CLI Script	Description
dcx	rstdcx, rstdoc	create .tags, labels and links
fromdocx	rstfromdocx	Convert DOCX to RST using Pandoc
listtable	rstlisttable	Convert RST grid tables to list-tables
untable	rstuntable	Converts certain list-tables to paragraphs
reflow	rstreflow	Reflow paragraphs and tables
reimg	rstreimg	Rename images referenced in the RST file
retable	rstretable	Transforms list tables to grid tables

rstdcx

restructuredText sources are split into two types of files: main files considered by Sphinx, and included files. Which of .rest or .rst is main or included is determined by source_suffix in a <root>/conf.py or by the extension of the _links_xxx files.

rstdoc creates documentation (PDF, HTML, DOCX) from restructuredText (.rst, .rest) using either

• Pandoc

• Sphinx

• Docutils configurable

rstdoc and rstdcx command line tools call dcx.py. which

• creates .tags to jump around with the editor

• handles .stpl files

• processes gen files (see examples produced by –rest)

• creates links files (_links_docx.r?st, _links_sphinx.r?st, …)

• forwards known files to either Pandoc, Sphinx or Docutils

See example at the end of dcx.py. It is supposed to be used with a build tool. make and waf examples are included.

• Initialize example tree (add --rstrest to make .rst main and .rest included files):

  $ ./dcx.py –rest repo #repo/doc/{sy,ra,sr,dd,tp}.rest files OR $ ./dcx.py –stpl repo #repo/doc/{sy,ra,sr,dd,tp}.rest.stpl files $ ./dcx.py –ipdt repo #repo/pdt/AAA/{i,p,d,t}.rest.stpl files $ ./dcx.py –over repo #.rest all over

• Only create .tags and _links_xxx.r?st:

  $ cd repo
  $ rstdoc

• Create the docs (and .tags and _links_xxx.r?st) with make:

  $ make html #OR
  $ make epub #OR
  $ make latex #OR
  $ make docx #OR
  $ make pdf

  The latter two are done by Pandoc, the others by Sphinx.

• Create the docs (and .tags and _links_xxx.r?st) with waf:

  Instead of using make one can load dcx.py (rstdoc.dcx) in waf. waf also considers all recursively included files, such that a change in any of them results in a rebuild. All files can have an additional .stpl extension to use SimpleTemplate.

  $ waf configure #also copies the latest version of waf in here $ waf –docs docx,sphinx_html,rst_odt $ #or you provide –docs during configure to always compile the docs

  • rst_xxx: via rst2xxx.py

  • sphinx_xxx: via Sphinx and

  • just xxx: via Pandoc.

The following image language files should be parallel to the .r?st files. They are automatically converted to .png and placed into ./_images or <updir>/_images or else parallel to the file.

• .tikz or .tikz.stpl. This needs LaTex.

• .svg or .svg.stpl

• .dot or .dot.stpl

  This needs graphviz.

• .uml or .uml.stpl

  This needs plantuml . Provide either

  • plantuml.bat with e.g. java -jar "%~dp0plantuml.jar" %* or

  • plantuml sh script with java -jar `dirname $BASH_SOURCE`/plantuml.jar "$@"

• .eps or .eps.stpl embedded postscript files.

  This needs inkscape.

• .pyg contains python code that produces a graphic. If the python code defines a to_svg or a save_to_png function, then that is used, to create a png. Else the following is tried

  • pyx.canvas.canvas from the pyx library or

  • cairocffi.Surface from cairocffi

  • matplotlib. If matplotlib.pyplot.get_fignums()>1 the figures result in <name><fignum>.png

  The same code or the file names can be used in a .r?st.stpl file with pngembed() or dcx.svgembed() to embed in html output.

  {{!svgembed("egpyx.pyg",outinfo)}}
  <%
  ansvg=svgembed('''
  from svgwrite import cm, mm, drawing
  d=drawing.Drawing(viewBox=('0 0 300 300'))
  d.add(d.circle(center=(2*cm, 2*cm), r='1cm', stroke='blue', stroke_width=9))
  '''.splitlines(),outinfo)
  %>
  {{!ansvg}}

Conventions

Files

• fain files and included files: .rest, .rst or vice versa. .txt are for literally included files (use :literal: option).

• templates separately rendered : *.rest.stpl and *.rst.stpl template included: *.rst.tpl Template lookup is done in . and .. with respect to the current file.

  • with %include('some.rst.tpl', param="test") with optional parameters

  • with %globals().update(include('utility.rst.tpl')) if it contains only definitions

Links

• .. _`id`: are reST targets. reST targets should not be template-generated. The template files should have a higher or equal number of targets than the generated file, in order for tags to jump to the template original. If one wants to generate reST targets, then this should better happen in a previous step, e.g. with gen files mentioned above.

• References use replacement substitutions : |id|.

• If you want an overview of the linking (traceability), add .. include:: _traceability_file.rst to index.rest or another .rest parallel to it. It is there in the example project, to include it in tests. _traceability_file.{svg,png,rst} are all in the same directory.

Link files are created in link roots, which are folders where the first main file (.rest or .rst) is encoutered during depth-first traversal. Non-overlapping link root paths produce separately linked file sets.

.. include:: /_links_sphinx.r?st, with the one initial / instead of a relative or absolute path, will automatically search upward for the _links_xxx.r?st file (_sphinx is replaced by what is needed by the wanted target when the docs are generated).

Sphinx conf.py is augmented by configuration for Pandoc and Docutils. It should be where the input file is, or better at the project root to be usable with waf.

See the example project created with --rest/stpl/ipdt/over and the sources of the documentation of rstdoc.

rstdcx CLI

Without parameters: creates |substitution| links and .tags ctags for reST targets.

With two or three parameters: process file or dir to out file or dir through Pandoc, Sphinx, Docutils (third parameter):

• html, docx, odt, pdf, … uses Pandoc.

• rst_html, rst_odt, rst_pdf, … uses rst2html, …

• sphinx_html, sphinx_pdf, … uses Sphinx. Sphinx provides a nice entry point via the sphinx bootstrap theme.

4th parameter onward become python defines usable in .stpl files.

Pdf output needs latex. Else you can make odt or docx and use

• win: swriter.exe --headless --convert-to pdf Untitled1.odt

• linux: lowriter --headless --convert-to pdf Untitled1.odt

Inkscape (.eps, .svg), Dot (.dot), Planuml (.uml), latex (.tex,.tikz) are converted to .png into ./_images or <updir>/_images or ‘.’. Any of the files can be a SimpleTemplate template (xxx.yyy.stpl).

Configuration is in conf.py or ../conf.py.

rstdoc --stpl|--rest|--ipdt|-over create sample project trees.

--stpl with .rest.stpl template files, --rest with only a doc folder with .rest files, --ipdt with inform-plan-do-test enhancement cycles --over with .rest files all over the project tree including symbolic links

Examples

Example folders (see wscript and Makefile there):

rstdoc --rest <folder> [--rstrest]
rstdoc --stpl <folder> [--rstrest]
rstdoc --ipdt <folder> [--rstrest]
rstdoc --over <folder> [--rstrest]

Examples usages with the files generated by rstdoc --stpl tmp:

cd tmp/doc
rstdcx   #expand .stpl and produce .tag and _links_xxx files

#expand stpl and append substitutions (for simple expansion use ``stpl <file> .``)
rstdcx dd.rest.stpl - rest           # expand to stdout, appending dd.html substitutions, to pipe to Pandoc
rstdcx dd.rest.stpl - html.          # as before
rstdcx dd.rest.stpl - docx.          # expand to stdout, appending dd.docx substitutions, to pipe to Pandoc
rstdcx dd.rest.stpl - newname.docx.  # expand template, appending substitutions for target newname.docx
rstdcx dd.rest.stpl - html           # expand to stdout, already process through Pandoc to produce html on stdout
rstdcx dd.rest.stpl                  # as before
rstdcx sy.rest.stpl - rst_html       # expand template, already process through Docutils to produce html on stdout
stpl sy.rest.stpl | rstdcx - - sy.html. # appending sy.html substitutions, e.g. to pipe to Pandoc
stpl dd.rest.stpl | rstdcx - - dd.html  # appending tp.html substitutions and produce html on stdout via Pandoc
rstdcx dd.rest.stpl dd.rest          # expand into dd.rest, appending substitutions for target dd.html
rstdcx dd.rest.stpl dd.html html     # expand template, process through Pandoc to produce dd.html
rstdcx dd.rest.stpl dd.html          # as before
rstdcx dd.rest.stpl dd.html rst_html # expand template, already process through Docutils to produce dd.html
rstdcx dd.rest.stpl dd.docx          # expand template, process through Pandoc to produce dd.docx
rstdcx dd.rest.stpl dd.odt pandoc    # expand template, process through Pandoc to produce dd.odt
rstdcx dd.rest.stpl dd.odt           # as before
rstdcx dd.rest.stpl dd.odt rst_odt   # expand template, process through Docutils to produce dd.odt
rstdcx dd.rest.stpl dd.odt rst       # as before
rstdcx . build html                  # convert current dir to build output dir using pandoc
rstdcx . build sphinx_html           # ... using sphinx (if no index.rest, every file separately)

#Sphinx is not file-oriented
#but with rstdcx you need to provide the files to give Sphinx ``master_doc`` (normally: index.rest)
#Directly from ``.stpl`` does not work with Sphinx
rstdcx index.rest ../build/index.html sphinx_html   # via Sphinx the output directory must be different

#convert the graphics and place the into _images or <updir>/_images
#if no _images directory exists they will be placed into the same directory
rstdcx egcairo.pyg
rstdcx egdot.dot.stpl
rstdcx egeps.eps
rstdcx egother.pyg
rstdcx egplt.pyg
rstdcx egpygal.pyg
rstdcx egpyx.pyg
rstdcx egsvg.svg.stpl
rstdcx egtikz.tikz
rstdcx egtikz1.tikz
rstdcx eguml.uml

#Convert graphics to a png (even if _images directory exists):
rstdcx eguml.uml eguml.png

#Files to other files:

rstdoc dd.rest.stpl dd.rest
rstdoc dd.rest.stpl dd.html html
rstdoc dd.rest.stpl dd.html
rstdoc sr.rest.stpl sr.html rst_html
rstdoc dd.rest.stpl dd.docx
rstdoc dd.rest.stpl dd.odt pandoc
rstdoc dd.rest.stpl dd.odt
rstdoc sr.rest.stpl sr.odt rst_odt
rstdoc sr.rest.stpl sr.odt rst
rstdoc index.rest build/index.html sphinx_html

#Directories to other directories with out info:

rstdoc . build html
rstdoc . build sphinx_html

Grep with python re in .py, .rst, .rest, .stpl, .tpl:

rstdoc --pygrep inline

Grep for keyword lines containing ‘png’:

rstdoc --kw png

Default keyword lines:

.. {kw1,kw2}
{{_ID3('kw1 kw2')}}
%__ID3('kw1 kw2')
:ID3: kw1 kw2

rstfromdocx

rstfromdocx: shell command

fromdocx: rstdoc module

Convert DOCX to RST in a subfolder of current dir, named after the DOCX file. It also creates conf.py, index.py and Makefile and copies dcx.py into the folder.

See rstdcx for format conventions for the RST.

There are options to post-process through:

--listtable (--join can be provided)
--untable
--reflow (--sentence True,  --join 0)
--reimg

rstfromdocx -lurg combines all of these.

To convert more DOCX documents into the same RST documentation folder, proceed like this:

• rename/copy the original DOCX to the name you want for the .rest file

• run rstfromdocx -lurg doc1.docx; instead of -lurg use your own options

• check the output in the doc1 subfolder

• repeat the previous 2 steps with the next DOCX files

• create a new folder, e.g. doc

• merge all other folders into that new folder

fromdocx.docx_rst_5 creates 5 different rst files with different postprocessing.

See rstreflow for an alternative proceeding.

rstlisttable

rstlisttable: shell command

listable: rstdoc module

Convert RST grid tables to list-tables.

1. Convert grid tables in a file to list-tables. The result is output to stdout:

   $ listtable.py input.rst

2. Convert several files:

   $ listtable.py input1.rst input2.rst
   $ listtable.py *.rst

3. Use pipe (but cat might not keep the encoding):

   $ cat in.rst | listtable.py -  | untable.py - > out.rst

Options

-j, --join

e.g.002. Join method per column: 0=””.join; 1=” “.join; 2=”\n”.join

rstuntable

rstuntable: shell command

untable: rstdoc module

Convert tables of following format to paragraphs with an ID. The ‘-’ in ID is removed and the ID is made lower case. Bold is removed.

ID-XY-00	text goes here
ID-XY-01	text again goes here

If the first entry does contain no word chars or spaces between words, then the table stays. For a different behavior replace paragraph23.

A file produced from a docx using pandoc or fromdocx.py will need a pre-processing via rstlisttable to convert grid tables to list-table tables. This is done in one step with rstfromdocx -lu doc.rst.

rstreflow

rstreflow: shell command

reflow: rstdoc module

Reflow tables and paragraphs in a rst document produced from a docx.

Post-process a docx in this order:

rstfromdocx doc.docx
rstlisttable doc/doc.rst > doc/tmp.rst
rstuntable doc/tmp.rst > doc/tmp1.rst
rstreflow doc/tmp1.rst > doc/tmp2.rst
rstreimg doc/tmp2.rst > doc/tmp3.rst
rm doc/doc.rst
mv doc/tmp3.rst doc/doc.rst
rm doc/tmp*

Check the intermediate results.

Else one can also do inplace:

rstfromdocx doc.docx
rstlisttable -i doc/doc.rst
rstuntable -i doc/doc.rst
rstreflow -i doc/doc.rst
rstreimg -i doc/doc.rst

Note

DOCX to RST using Pandoc

rstfromdocx -lurg doc.rst converts a docx to RST and does all the post-processing in one step.

It is adviced, though, to compare the output with the original and do some manual corrections here and there.

rstreimg

rstreimg: shell command

reimg: rstdoc module

Reimg renames the images in the rst file and the files themselves. It uses part of the document name and a number as new names.

This is useful, if more RST documents converted from DOCX should be combined in one directory and the names of the images have no meaning (image13,…).

rstretable

rstretable: shell command

retable: rstdoc module

Transforms list tables to grid tables.

This file also contains the code from the Vim plugin vim-rst-tables-py3, plus some little fixes. rstdoc is used by the Vim plugin vim_py3_rst , which replaces vim-rst-tables-py3.