Failure to output embedded image to docx

[Analect]

@hugobuddel . Thanks for your efforts with this tool.
I was trying to get two things working. Based on the *.rst file (middle below), I am able to output html and pdf with sections numbered as well as embedded images, however neither work when outputting to a docx. Is this possible, do you know? Perhaps I'm just missing some added configuration?

Thanks.

[hugobuddel]

Thank you for the issue @Analect.

Not all features are supported in the docx plugin yet; we focused on the ones we need ourselves. I'll take a look at it next week. The image support shouldn't be difficult to add (we had it before). I don't know about the section numbering, lets see.

[Analect]

@hugobuddel ... much appreciated. Thanks.

[Analect]

@hugobuddel ... any joy in getting some of those elements working? Thanks

[hugobuddel]

@Analect , does attached screenshot do what you want? Created from sample_5 in the hb/feature/issue18 branch: https://github.com/mherkazandjian/docxsphinx/tree/hb/feature/issue18/examples/sample_5

Main points:

• Image support is improved, you can now specify 'height' and 'width'. 'align' is ignored though.
• The section numbers are defined in the styles the template docx file. 'sample_5' contains an example template. (I'm not sure exactly which style contains this information, probably one of the headers.)

[Analect]

@hugobuddel ... sorry for the delayed response. I tested your docxsphinx-hb-feature-issue18 branch and it appears to work well. Thanks for that. Is it possible for this to get merged into master?

[hugobuddel]

Great that it works! I've assigend #19 to @mherkazandjian to merge it into develop first.

[Analect]

@hugobuddel ... I'm now trying to get this working in the context of my own sphinx docs. I'm working with a conda environment where I had previously installed the master branch. I then checked out the fixed docxsphinx-hb-feature-issue18 branch ... and ran through a reinstall, as per README. Not sure if that is the right approach or not ... or if I needed to uninstall in that conda environment before reinstalling. I updated my conf.py to broadly emulate yours (in terms of referencing template.docx etc.. however, I'm hitting this error. Any thoughts on what might be causing this. I haven't necessarily added explicit widths/heights to my embedded images.

Exception occurred:
  File "C:\Users\TXXXXXX\AppData\Local\Continuum\Miniconda3\envs\client-docs\lib\site-packages\docx\shared.py", line 113, in __new__
    emu = int(points * Length._EMUS_PER_PT)
TypeError: unsupported operand type(s) for *: 'NoneType' and 'int'

[Analect]

@hugobuddel ... actually ... the above was resolved by adding a width parameter to figures .. without this, I was getting the error above, for some reason.

One final question ... should this package be able to handle graphviz and perhaps plantuml

.. graphviz::

    digraph G {
    rankdir=LR
    node [shape=cds, width=1, height=1]
    "Test\n 1" -> "Test\n 2" -> "Test\n 3" -> "Test\n 4" -> "Test\n 5"
    }

resulting in ...

[hugobuddel]

thanks for testing this @Analect. The feature branch now works properly when not specifying the image dimensions.

One thing about the templates is that the plugin currently does not respect the template_path. Created #22 for this. The current situation is workable though.

We will probably not add graphviz support any time soon; would be nice though. If you wish you can create a separate issue to track it. For now you could convert the graphs to png's.

[Analect]

Thanks @hugobuddel ... I'll close this out for now then. Appreciate your support with this.