Load articles from other file formats

[Jenselme]

We could load data from:

• epub: should be easily doable, it’s HTML in a ZIP. What would be nice would be to allow import of all the HTML pages in one article, each one in an independent article, each one in a group of related articles. The only issue is to load the ZIP while being protected from ZIP bombs.
• odt/docx: it’s XML in a ZIP. I don’t think converting to HTML can be part of this project. We could have a service based on LibreOffice that does the conversion to HTML for us.
• pdf: probably the hardest one.

[Jenselme]

Let’s try to support:

• Direct HTML upload.
• Direct MHTML upload.
• epub upload (if I can manage to prevent ZIP bomb).
• Link that responds with an epub.

Other formats are too weird. Let’s just check we can save the link to keep it around with tags. And users will be able to convert it to HTML themselves to save it.

[Jenselme]

Notes:

• Currently, all articles must have a link. If we allow direct upload, this won’t be the case.
• We would also need a way to store the images associated with the epub and update the HTML to points to the images we saved. This would required extra work and some changes to the deploy configuration to configure where to store the images.