@pirate This first version is functional but it is lacking a lot of flexibility. It uses the CHROME_BINARY. I need to add new options to be able to use other browsers.
Also, the installation is not possible via package managers, like it is for wget and chrome. I had to clone it, install npm dependencies and create a symlink to my bin folder. Should we leave this method enabled by default?

It currently works, if you have chrome installed. It creates a new file in the archive named single-file.html. Currently, it is created directly by the single-file process.

A good start, things left:

• add an iframe to the link_details.html template for singlefile
• add a files column icon for SingleFile in the core.admin snapshot list
• add it to archviebox.index.latest_outputs
• add it to archviebox.index.canonoical_outputs
• anywhere else you can fine (search for dom_path and add it next to it everywhere)

• add an iframe to the link_details.html template for singlefile
• add a files column icon for SingleFile in the core.admin snapshot list
• add it to archviebox.index.latest_outputs
• add it to archviebox.index.canonoical_outputs
• anywhere else you can fine (search for dom_path and add it next to it everywhere)

@pirate please review. I think we can close this and add the new config options to the cli in another PR. Let me know if you prefer adding that in here.

The last thing remaining is to add it to the docker image so it works out-of-the-box.

@pirate Added a fixture to the tests to avoid using methods we were not actually testing. By default it will bypass all the extractors. The python tests should be running faster now because of this change.
I also added the singlefile installation as part of the process (it was that or remove the singlefile test).
Additionally, I added a config variable that was missing (SAVE_SINGLEFILE) and that was causing erratic behaviour.