Super cool; thank you for looking into this!

If it doesn't seem overly complex to implement, I think I'd prefer to use an explicit command-line flag (like beet import --read-log foo.log) as opposed to implicitly allowing logs as arguments alongside files/directories. The main benefit would be that it will be more "debuggable" from the perspective of the user: if the log file isn't want they expect or is malformed, they can get a useful error message instead of letting the text file be treated as a media file. I hope this makes sense!

If it doesn't seem overly complex to implement, I think I'd prefer to use an explicit command-line flag (like beet import --read-log foo.log) as opposed to implicitly allowing logs as arguments alongside files/directories.

Done. I also added a small test for the helper function.

Sorry for the force-pushing, I abused git as a way to deploy the code on my NAS (so I can test it on my library). Also don't have windows so I had to use the CI for trial and error test driven development ;-)

Awesome; thanks for taking this "flag" approach! I think this is basically there—I have one or two very low-level comments for your consideration.

Looking great. Thanks for the fast and clean work.