My clients heavly use MS Word .docx and I need to do a conversion to ProseMirror. I know some tools like Panda etc. but it’s not exctactly the same. Clients demand almost 1:1 similarity.
Did someone try with AI or ML to produce as close as possible result? Any advice? Will Tensorflow work for that?
Why not just convert to HTML and then parse it with ProseMirror? If you are using ML I don’t really know how you are going to score the similarities between the PM and original Word documents. Export them back to Word? Seems quite convoluted.