feature: add `<a:snd>` (sound) support to click-action

[NathanTech7713]

Hi, me again.

>>> prs = Presentation("narrated_powerpoint_presentation.pptx")
>>> shape = prs.slides[1].shapes[2]
>>> shape.name
'Audio 5'
>>> shape.click_action.action
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "C:\python332\lib\site-packages\pptx\action.py", line 53, in action
    return {
KeyError: 'media'

And if you examine the shape it acts like it is an image file rather than audio.
Preferred solutions/workarounds:

1. To be able to recognise that it is audio.
2. 2. To be able to retrieve the name of the media file which I can then extract from the pptx as a zip.
      Uploading narrated_powerpoint_presentation.pptx…

[scanny]

Okay, this one looks like another implementation boundary, although it's unfortunate that it can produce an exception. Minimally it would be better if it returned PP_ACTION.NONE rather than raising KeyError. I'll see if I can get that fixed up in the release that's currently open and should go out by the end of the month.

The image-like behavior is expected I believe. You haven't mentioned specifically in what ways it acts like one, but there is no "sound" shape and what you would typically expect is a picture of an "audio speaker" icon, which is of course an image. So I'm not sure there is anything out of the ordinary there. The shape.shape_type property will report MSO.PICTURE, but that's because there is no distinct Audio object yet.

As far as being able to detect that it has "linked" sound, that's a legitimate feature request so I've changed the title to reflect that and we'll leave this open as a feature request. I expect audio would be detected something like this:

has_audio = shape.shape_type == MSO_SHAPE_TYPE.MEDIA and shape.media_type == PP_MEDIA_TYPE.SOUND

[NathanTech7713]

@scanny thanks for that. do you know is there a way (current or otherwise), to draw out which file the audio shape is pointing to in the pptx, i.E, which media.mp3 file is associated?

[scanny]

Do a print(shape._sp.xml) on the shape you know has one and let's have a look. Pretty sure that would be doable.

[NathanTech7713]

Heya.
So shape._sp.xml said not found but shape.element.xml worked, assuming that's basically the same thing.
I've included the output below, I assume the either the rId="2" or id="6" points to the media in the format of file="media"+str(id) but I am probably wrong here.

>>> shape.element.xml
<p:pic xmlns:p="http://schemas.openxmlformats.org/presentationml/2006/main" 
       xmlns:a="http://schemas.openxmlformats.org/drawingml/2006/main"
       xmlns:r="http://schemas.openxmlformats.org/officeDocument/2006/relationships">
  <p:nvPicPr>
    <p:cNvPr id="6" name="Audio 5">
      <a:hlinkClick r:id="" action="ppaction://media"/>
    </p:cNvPr>
    <p:cNvPicPr>
      <a:picLocks noChangeAspect="1"/>
    </p:cNvPicPr>
    <p:nvPr>
      <a:audioFile r:link="rId2"/>
      <p:extLst>
        <p:ext uri="{DAA4B4D4-6D71-4841-9C94-3DE7FCFB9230}">
          <p14:media xmlns:p14="http://schemas.microsoft.com/office/powerpoint/2010/main" r:embed="rId1"/>
        </p:ext>
      </p:extLst>
    </p:nvPr>
  </p:nvPicPr>
  <p:blipFill>
    <a:blip r:embed="rId4"/>
    <a:stretch>
      <a:fillRect/>
    </a:stretch>
  </p:blipFill>
  <p:spPr>
    <a:xfrm>
      <a:off x="8382000" y="6096000"/>
      <a:ext cx="609600" cy="609600"/>
    </a:xfrm>
    <a:prstGeom prst="rect">
      <a:avLst/>
    </a:prstGeom>
  </p:spPr>
</p:pic>

[NathanTech7713]

Morning @scanny was just doing a little research and stumbled across a stackoverflow question regarding image file names and how power point seems to embed the original file name as alternative text/image descriptions. Is that what we see through shape.image.filename? If not is there a way to pull this out as if that points to media1.wav for instance, then problem solved for me

[NathanTech7713]

Forgot to include the link, it's in the second answer, not the first.
https://graphicdesign.stackexchange.com/questions/48906/how-to-find-name-of-original-image-in-powerpoint-presentation

[scanny]

@NathanTech7713 I reformatted the XML you provided so it was readable, please take care of that yourself next time.

You'll need to do a little investigation yourself on this, but the audio is either going to be on "rId1" or "rId2". The id="6" is the shape-id and does not correspond to a relationship to another package part.

I'm not sure why there would be 2 relationships but the next step would be to inspect the slide2.xml.rels file and see what these relationships point to. Then you can have a look at those files in the package if necessary. It could be they are two "links" to the same resource or it could be that one is a legacy format and the other a new format of the same audio.

As far as detecting these shapes is concerned you could narrow it to shapes that reported picture type, like shape.shape_type == MSO.PICTURE. From there you could use XPath, maybe to look for the p:nvPr/a:audioFile element:

def has_audio(shape):
    """Return True if `shape` has embedded audio, False otherwise."""
    if shape.shape_type != MSO.PICTURE:
        return False
    audioFiles = shape._pic.xpath(".//p:nvPr/a:audioFile")
    return True if audioFiles else False

Actually, after writing this I notice that the a:audioFile element has an r:link attribute and the p14:media element has an r:embed attribute. This might mean that the latter is a "link" to an embedded sound file and the former is a URL to a sound resource on the web or something.

[NathanTech7713]

@scanny Thanks for that. My sincere apologies. A byproduct of rarely working with others leaves me a bit ignorant of how best to format for other's reading sometimes :) we live and learn!

Anyways, I took a gander at the .rels file you mentioned and sure enough, plain as brass, there it is. Studying this file it looks like one rId points to a 2006 link, while the rId3 points to a 2007 link.

thanks for the code above, as well. I need to sit down si down and play with it and this a bit to see where I get, but if I get somewhere is it worth doing a PR?

[scanny]

It might be worth developing a PR, especially if you will be patching your own fork, but be aware it would be very unlikely to get merged into the library. Short version of that is it's unlikely to have all the analysis documentation, tests, and consistency with implementation conventions and standards. It just generally works out that it's faster to reimplement something from scratch than incorporate a contributed PR. Generally we don't implement anything new unless it's sponsored and we have the nights and weekend time to get to it (which hasn't been the case lately).

But that doesn't mean it wouldn't be used to inform such a feature if it was added or for someone else to pick up and merge into their own fork. You just have to decide whether it's worth it to you if it won't get merged.

[MartinPacker]

Just a thought:

We could use a repository of companion samples. I don't think StackOverflow is it - nor is the GitHub Issues capability.

While I can't own such a thing I'd (for what it's worth) contribute to it.