The warning of iter applies here. This check can take a looooonnnnggg time to finish.

I think this should be a file existence check here.
It will be faster than iterating over a long list of ids... (which is, from the current implementation's standpoint, the result of walking the disk)

I mean the __iter__ function...

I did not get the test's intent at first sight.
Can you please rename the variables to something more explicit?

For example:

• content1 -> original_content_data
• content2 -> corrupted_content_data
• ra -> content2_id
• rr -> content_restored_id

For me, the tests can help in documenting the code so we must strive to make them clear.

Of course, I apologize in advance if some existing are not ^^
It's a goal to achieve which is not always easy to reach!

Don't the in syntax only use the __contains__ method (which is a file existence check) ?

Indeed, nice catch!
Nothing to do then ^^