Don't show recommendations below a certain threshold during import

[arsaboo]

Is it possible to ignore recommendations below a certain threshold (say 60%) during the import? That is, only show recommendations that match at least 60%. I want to do a quick check on my library to see if there are any new matches on MB. SO, I would like to use the --timid option but ignore all low-quality matches.

Basically, I want to skip over if there are no good (over 60%) matches and confirm with me if there are good matches.

I played with medium_rec_thresh but that does not seem to exclude low matches.

[sampsyo]

Hmm; I don't think we have a way to ignore these matches altogether. But you're right to play with medium_rec_thresh, which will at least avoid showing you a "best match" to confirm when all the matches are low-quality and instead show you the list to pick from.

It sounds like what you'd want is to automatically skip when all the matches are bad. This is not quite that, but you could imagine setting default_action to skip to at least make it easy to skip stuff (just hit the return key):
https://beets.readthedocs.io/en/stable/reference/config.html#default-action

[arsaboo]

It would be nice to have a config setting to ignore low-quality matches.

[arsaboo]

@sampsyo I am trying to add this feature, but I am a little unsure how distance gets added to an album. I see that plugins have get_albums (e.g., here)function that return a list of potential albums. However, it is not clear to me how can I get the album distance for each of these albums. I see that plugins have a album_distance function but it seems it is applied automatically. Is it possible to get the distance value for each album? Does get_albums automatically return sorted by distance?

[sampsyo]

Here's where distances come from:

beets/beets/autotag/match.py

Lines 421 to 434 in 5aae612

	# Get the change distance.
	dist = distance(items, info, mapping)
	# Skip matches with ignored penalties.
	penalties = [key for key, _ in dist]
	ignored = cast(Sequence[str], config['match']['ignored'].as_str_seq())
	for penalty in ignored:
	if penalty in penalties:
	log.debug('Ignored. Penalty: {0}', penalty)
	return
	log.debug('Success. Distance: {0}', dist)
	results[info.album_id] = hooks.AlbumMatch(dist, info, mapping,
	extra_items, extra_tracks)

In particular, in _add_candidate, where we ingest a new candidate from MB or a plugin, we call the distance function to calculate its quality and the construct an AlbumMatch object that includes it.