how to extract docx image using sha1 as name

[badiku]

hi, doc says that ir default to sha1 but I always get names like media/image1.jpg , which is not original path and not sha1

I think sha1 name is better for finding out difference in image updating. is the documentation wrong? is there any parameter to force extract media to use sha1 based file name?

--extract-media=DIR|FILE.zip            
Extract images and other media contained in or linked from the      source document to the path DIR, creating it if      necessary, and adjust the images references in the document so      they point to the extracted files. Media are downloaded, read from      the file system, or extracted from a binary container (e.g. docx),      as needed. The original file paths are used if they are relative      paths not containing ... Otherwise filenames are      constructed from the SHA1 hash of the contents

[jgm]

I believe the docs are correct. Relevant code is at https://github.com/jgm/pandoc/blob/main/src/Text/Pandoc/MediaBag.hs#L97-L104
I don't know why you would have gotten image1.jpg if that is not the original filename, but you'd have to give a reproducible case for us to investigate further (and say what version of pandoc you're using).
There is no way to force SHA1 names for all resources.

[badiku]

every image file embedded in Office Word is automatically renamed to image1.jpg image2.jpg etc.
of I extract a docx file with 7zip I get files with name image1.jpg.
this is the so-called original file parh?

[jgm]

Yes, it is how it is stored internally in the docx container. Pandoc has no way of knowing what the name of the file you inserted is; once it's in the docx, it has this new name, and that is Word's doing, not pandoc's.

[badiku]

since Pandoc has no way of knowing what the name of the file you inserted is, that is to say, pandoc doesn't know the original path name, so should it ise sha1 name, as doc says ?###

[jgm]

No - as the doc says, it's the original name (but that means the name of the file in the container).
"Media are downloaded, read from the file system, or extracted from a binary container (e.g. docx), as needed." In this case, they are extracted from a binary container, and what is relevant is the names they have in that container.

[badiku]

I see. so for docx it is never named after sha1. only way mayb write lua filter to get sha1 and rename those extracted image files. and change them in text

[badiku]

so this is a good time for me to learn to write a lua filter.

please tell me if there's any problem or better idea

-- rename image file name in media to sha1 hash
--  pandoc --lua-filter image2hash.lua t.docx --extract-media=. -o t.html
Pandoc = function (doc)
  local images = {}
  for filepath, mimetype, contents in pandoc.mediabag.items() do
    table.insert(images, {filepath, mimetype, contents})
    pandoc.mediabag.delete(filepath)
  end
  local filepaths = {}
  for item in pandoc.List(images):iter() do
    local filepath, mimetype, contents = table.unpack(item)
    local sha1 = pandoc.utils.sha1(contents)
    local ext = select(2, pandoc.path.split_extension(filepath) ) or '.png'
    filepath = 'media/' ..sha1 ..ext
    filepaths[#filepaths+1] = filepath
    pandoc.mediabag.insert(filepath, mimetype, contents)
  end
  local n = 0
  return doc:walk{Image = function(el) n=n+1 el.src = filepaths[n] return el end}
end