extracting images from a document

[peczony]

Hi,

my task is to parse some docx files, extracting plain text and images from them.

am i right in thinking (after reading the docs and playing around with some code) that there is no way so far to extract images from a document? I.e. I know that the document has an inline_shapes property, but:

• the run doesn’t have one, so I can’t establish the link between an image and a run
• InlineShape doesn’t have a method for saving an image to disk?

I conjecture that the best way I have is to unzip the docx and parse xml?

[scanny]

You can take a couple approaches.

All the images are in the .docx (zip archive) as themselves. The filenames are generic, like image1.jpg, but they're all there and you don't need any XML parsing to get at them. Just open as a zip and extract all the members with a member name ("path") that starts with 'word/media/'. You might want to check to make sure the extension is sensible, like .jpg, .png, etc.; there might be other things in there like videos or sound files, I'm not sure.

Of course there's no correlation between the image file and its placement in the document. If you want that you would need to parse some XML.

If you need to get into positions, let me know more about what you're trying to accomplish and I can suggest an approach. You can probably use the existing API to get pretty close and then just use a little bit of lxml API code to get the rest.

You might also want to take a look at Package.image_parts. You can get all the same things as you can get from the zip approach by starting there and using the properties of docx.parts.image.ImagePart like .blob and .content_type.

[peczony]

Thanks for the hint at Package.image_parts.

However I ended up by using the package pydocx, which can translate docx to html preserving images at their positions in base64-encoded form, which allows extracting them and the only thing which is lost is original filename (which is not very important)

[scanny]

Glad you got it working @peczony :)

[yashugupta786]

@scanny I want to extract the text and images with their location and need to write in new docx file . I have extracted all the text from the docx file and write it to a new file but when i am not able to extract the images and write it to new file please help me . As i am stuck . I have already raised issue in stackoverflow
'''
https://stackoverflow.com/questions/65843126/how-to-extract-text-and-images-from-a-docx-file-and-write-it-to-new-docx-file-wi
'''
Thanks in advance