Support PDF/UA

[grewn0uille]

Fix #1682

[flintflump]

As already mentioned in #1682 I would like to offer help with this. I already looked at the your changes, but I still have some questions:

• It seems that this PR depends on some unpublished changes in pydyf. I made my own version of pydyf to make this branch work, but I had to disable some things as the resulting PDF was broken. I am not sure if its due to the changes here, or if I forgot some adjustments in pydyf. Would you mind publishing your current progress with pydyf?
• The tagging seems to work mostly fine but the structure tree and parent tree references are apparently not completely valid. Are you working on this right now?
• What are the next steps in your opinion? Do you have a specific roadmap what to tackle next? I don't want to do things in parallel again. ;)

[liZe]

@flintflump Thanks a lot for your help.

• It seems that this PR depends on some unpublished changes in pydyf. I made my own version of pydyf to make this branch work, but I had to disable some things as the resulting PDF was broken. I am not sure if its due to the changes here, or if I forgot some adjustments in pydyf. Would you mind publishing your current progress with pydyf?

Sorry, we forgot to push the changes. It’s done (on master).

• The tagging seems to work mostly fine but the structure tree and parent tree references are apparently not completely valid. Are you working on this right now?

We’ll be working for a client on this topic during August. We’ll first try to make everything work for his needs, so things will probably change during the next days/weeks (including tests).

• What are the next steps in your opinion? Do you have a specific roadmap what to tackle next? I don't want to do things in parallel again. ;)

Adding unit tests, refactoring and cleaning (we have dirty code that may lead to broken structures, we know that). Including #1695 could also be interesting. Then we’ll test our client’s documents and be sure that everything works fine. So it’ll be probably easier to work in parallel when everything is OK for our client, in a few weeks.