Interactive import session should offer a "Retry" choice in case of MusicBrainz server error

[bartkl]

When importing using the autotagger, sometimes the MusicBrainz server is not reachable, which gives an error like:
Error: MusicBrainz not reachable in get release ...
In case the import session is interactive, the user gets prompted with the usual choice dialog. It would be very nice if an extra option [R]etry is added in this situation, since quite often it's not server downtime but just hickups (also possibly on your own end). Now I have to skip the album in case I'd like to retry the autotagging process, which is especially tedious if it concerns a mass import.

Would love to hear your thoughts.

[peace899]

I thought musicbrainzngs already has max 8 retries with each API call. As per python-musicbrainzngs issue #174

I'm not sure but I think on each exception, it retries 8 times to do your request? Maybe someone who knows better can explain.

[sampsyo]

It does indeed retry several times before giving up. Of course, there is a chance that waiting a while (for the user to respond) would be enough to clear things up. For example, if the network connection is down, the user can reconnect the cable and then retry.

This seems like a good addition. It's not 100% clear how the layering problems will be solved here, but it's worth a try.

[arcresu]

Just to note that this is related to #2219. Both could potentially be solved by a solution along the lines of #266 where beets keeps track of when a metadata plugin has acted on a given Item. If we knew which Items in the library hadn't been processed (e.g. due to transient communication issues over an API) then it would be possible to re-run the plugin over just those instead of re-scanning the whole library.